ICML2025 论文笔记 TODO¶

总计: 1225 篇 | 已完成: 1225 | 待更新: 0

a bayesian model selection criterion for selecting pretraining checkpoints | arXiv: 2410.05612
a certified unlearning approach without access to source data | arXiv: 2506.06486
a cognac shot to forget bad memories corrective unlearning for graph neural netw | arXiv: 2412.00789
a cross modal knowledge distillation data augmentation recipe for improving tran | arXiv: 2505.21317
a general graph spectral wavelet convolution via chebyshev order decomposition | arXiv: 2405.13806
a generalizable physics-enhanced state space model for long-term dynamics foreca | arXiv: 2507.10792
a generalization result for convergence in learning-to-optimize | arXiv: 2410.07704
a mathematical framework for ai-human integration in work | arXiv: 2505.23432
a near-optimal single-loop stochastic algorithm for convex finite-sum coupled co | arXiv: 2312.02277
a reasoning-based approach to cryptic crossword clue solving | arXiv: 2506.04824
a recipe for causal graph regression confounding effects revisited | arXiv: 2507.00440
a square peg in a square hole meta-expert for long-tailed semi-supervised learni | arXiv: 2505.16341
a theoretical study of hyper self-attention through the lens of interactions rep | arXiv: 2506.06179
a unified view on learning unnormalized distributions via noise-contrastive esti | arXiv: 2409.18209
aaar-10 assessing ais potential to assist research | arXiv: 2410.22394
ab initio nonparametric variable selection for scalable symbolic regression with | arXiv: 2410.13681
abkd pursuing a proper allocation of the probability mass in knowledge distillat | arXiv: 2505.04560
accelerating spectral clustering under fairness constraints | arXiv: 2506.08143
access controls will solve the dual-use dilemma | arXiv: 2505.09341
action-constrained imitation learning | arXiv: 2508.14379
action-dependent optimality-preserving reward shaping | arXiv: 2505.12611
action-minimization meets generative modeling efficient transition path sampling | arXiv: 2504.18506
actionpiece contextually tokenizing action sequences for generative recommendati | arXiv: 2502.13581
activation space interventions can be transferred between large language models | arXiv: 2503.04429
actor-critics can achieve optimal sample efficiency | arXiv: 2505.03710
ad-hoc human-ai coordination challenge | arXiv: 2506.21490
adadecode accelerating llm decoding with adaptive layer parallelism | arXiv: 2506.03700
adapter naturally serves as decoupler for cross-domain few-shot semantic segment | arXiv: 2506.07376
adaptive elicitation of latent information using natural language | arXiv: 2504.04204
adaptive estimation and learning under temporal distribution shift | arXiv: 2505.15803
adaptive multi-prompt contrastive network for few-shot out-of-distribution detec | arXiv: 2506.17633
adaptivestep automatically dividing reasoning step through model confidence | arXiv: 2502.13943
adaworld learning adaptable world models with latent actions | arXiv: 2503.18938
addressing imbalanced domain-incremental learning through dual-balance collabora | arXiv: 2507.07100
adhmr aligning diffusion-based human mesh recovery via direct preference optimiz | arXiv: 2505.10250
adios antibody development via opponent shaping | arXiv: 2409.10588
adjustment for confounding using pre-trained representations | arXiv: 2506.14329
advagent controllable blackbox red-teaming on web agents | arXiv: 2410.17401
adversarial combinatorial semi-bandits with graph feedback | arXiv: 2502.18826
adversarial cooperative rationalization the risk of spurious correlations in eve | arXiv: 2505.02118
adversarial inception backdoor attacks against reinforcement learning | arXiv: 2410.13995
adversarial manipulation of reasoning models using internal representations | arXiv: 2507.03167
advprompter fast adaptive adversarial prompting for llms | arXiv: 2404.16873
agacci affiliated grading agents for criteria-centric interface in educational c | arXiv: 2507.05321
agent warpp workflow adherence via runtime parallel personalization | arXiv: 2507.19543
aguvis unified pure vision agents for autonomous gui interaction | arXiv: 2412.04454
alberta wells dataset pinpointing oil and gas wells from satellite imagery | arXiv: 2410.09032
algebra unveils deep learning -- an invitation to neuroalgebraic geometry | arXiv: 2501.18915
align-then-unlearn embedding alignment for llm unlearning | arXiv: 2506.13181
aligning llms by predicting preferences from user writing samples | arXiv: 2505.23815
aligning protein conformation ensemble generation with physical feedback | arXiv: 2505.24203
aligning spoken dialogue models from user interactions | arXiv: 2506.21463
all-atom diffusion transformers unified generative modelling of molecules and ma | arXiv: 2503.03965
alpha-sql zero-shot text-to-sql using monte carlo tree search | arXiv: 2502.17248
alphapo reward shape matters for llm alignment | arXiv: 2501.03884
ampo active multi-preference optimization for self-play preference selection | arXiv: 2502.18293
an attack to break permutation-based private third-party inference schemes for l | arXiv: 2505.18332
an efficient matrix multiplication algorithm for accelerating inference in binar | arXiv: 2411.06360
an efficient private gpt never autoregressively decodes | arXiv: 2505.15252
angle domain guidance latent diffusion requires rotation rather than extrapolati | arXiv: 2506.11039
annealing flow generative models towards sampling high-dimensional and multi-mod | arXiv: 2409.20547
any4 learned 4-bit numeric representation for llms | arXiv: 2507.04610
are llm belief updates consistent with bayes theorem | arXiv: 2507.17951
are llms prescient a continuous evaluation using daily news as the oracle | arXiv: 2411.08324
assistancezero scalably solving assistance games | arXiv: 2504.07091
asymrnr video diffusion transformers acceleration with asymmetric reduction and | arXiv: 2412.11706
autoal automated active learning with differentiable query strategy search | arXiv: 2410.13853
autoencoder-based hybrid replay for class-incremental learning | arXiv: 2505.05926
autoformulation of mathematical optimization models using llms | arXiv: 2411.01679
automatic reward shaping from confounded offline data | arXiv: 2505.11478
automl-agent a multi-agent llm framework for full-pipeline automl | arXiv: 2410.02958
autonomy-of-experts models | arXiv: 2501.13074
avoiding catastrophe in online learning by asking for help | arXiv: 2402.08062
avoiding leakage poisoning concept interventions under distribution shifts | arXiv: 2504.17921
b-score detecting biases in large language models using response history | arXiv: 2505.18545
balanced learning for domain adaptive semantic segmentation | arXiv: 2512.06886
balancing efficiency and expressiveness subgraph gnns with walk-based centrality | arXiv: 2501.03113
banyan improved representation learning with explicit structure | arXiv: 2407.17771
bayesian inference for correlated human experts and classifiers | arXiv: 2506.05636
bayesian neural scaling law extrapolation with prior-data fitted networks | arXiv: 2505.23032
beaver building environments with assessable variation for evaluating multi-obje | arXiv: 2507.07769
became bayesian continual learning with adaptive model merging | arXiv: 2504.02666
benchmarking quantum reinforcement learning | arXiv: 2501.15893
benefits of early stopping in gradient descent for overparameterized logistic re | arXiv: 2502.13283
benign overfitting in token selection of attention mechanism | arXiv: 2409.17625
best subset selection optimal pursuit for feature selection and elimination | arXiv: 2501.16815
best-route adaptive llm routing with test-time optimal compute | arXiv: 2506.22716
beyond bradley-terry models a general preference model for language model alignm | arXiv: 2410.02197
beyond communication overhead a multilevel monte carlo approach for mitigating c | arXiv: 2507.05508
beyond cvar leveraging static spectral risk measures for enhanced decision-makin | arXiv: 2501.02087
beyond entropy region confidence proxy for wild test-time adaptation | arXiv: 2505.20704
beyond induction heads in-context meta learning induces multi-phase circuit emer | arXiv: 2505.16694
beyond message passing neural graph pattern machine | arXiv: 2501.18739
beyond one-hot labels semantic mixing for model calibration | arXiv: 2504.13548
beyond self-repellent kernels history-driven target towards efficient nonlinear | arXiv: 2505.18300
beyond sensor data foundation models of behavioral data from wearables improve h | arXiv: 2507.00191
beyond the rainbow high performance deep reinforcement learning on a desktop pc | arXiv: 2411.03820
beyond zero initialization investigating the impact of non-zero initialization o | arXiv: 2505.23194
biassemble learning collaborative affordance for bimanual geometric assembly | arXiv: 2506.06221
binary hypothesis testing for softmax models and leverage score models | arXiv: 2405.06003
binauralflow a causal and streamable approach for high-quality binaural speech s | arXiv: 2505.22865
bipartite ranking from multiple labels on loss versus label aggregation | arXiv: 2504.11284
blockdialect block-wise fine-grained mixed format quantization for energy-effici | arXiv: 2501.01144
blueglass a framework for composite ai safety | arXiv: 2507.10106
boa attention-aware post-training quantization without backpropagation | arXiv: 2406.13474
boosting masked ecg-text auto-encoders as discriminative learners | arXiv: 2410.02131
bopo neural combinatorial optimization via best-anchored and objective-guided pr | arXiv: 2503.07580
bounded rationality for llms satisficing alignment at inference-time | arXiv: 2505.23729
breaking silos adaptive model fusion unlocks better time series forecasting | arXiv: 2505.18442
breaking the n15 additive error barrier for private and efficient graph sparsifi | arXiv: 2507.01873
bridge bootstrapping text to control time-series generation via multi-agent iter | arXiv: 2503.02445
bridging the language gap synthetic voice diversity via latent mixup for equitab | arXiv: 2511.20534
bring reason to vision understanding perception and reasoning through model merg | arXiv: 2505.05464
brite bootstrapping reinforced thinking process to enhance language model reason | arXiv: 2501.18858
broadband ground motion synthesis by diffusion model with minimal condition | arXiv: 2412.17333
build agent advocates not platform agents | arXiv: 2505.04345
ca2-vdm efficient autoregressive video diffusion model with causal generation an | arXiv: 2411.16375
can one safety loop guard them all agentic guard rails for federated computing | arXiv: 2506.20000
can rlhf be more efficient with imperfect reward models a policy coverage perspe | arXiv: 2502.19255
can transformers learn full bayesian inference in context | arXiv: 2501.16825
cape context-aware prompt perturbation mechanism with differential privacy | arXiv: 2505.05922
cascade token-sharded private llm inference | arXiv: 2507.05228
causal abstraction inference under lossy representations | arXiv: 2509.21607
causal discovery from conditionally stationary time series | arXiv: 2110.06257
causal discovery of latent variables in galactic archaeology | arXiv: 2507.00134
causal effect identification in lvlingam from higher-order cumulants | arXiv: 2506.05202
causal evidence for the primordiality of colors in trans-neptunian objects | arXiv: 2507.03760
causal foundation models disentangling physics from instrument properties | arXiv: 2507.05333
causal-pik causality-based physical reasoning with a physics-informed kernel | arXiv: 2505.22861
causality-aware contrastive learning for robust multivariate time-series anomaly | arXiv: 2506.03964
certification for differentially private prediction in gradient-based training | arXiv: 2406.13433
cfp-gen combinatorial functional protein generation via diffusion language model | arXiv: 2505.22869
challenges and future directions of data-centric ai alignment | arXiv: 2410.01957
chameleon a flexible data-mixing framework for language model pretraining and fi | arXiv: 2505.24844
channel normalization for time series channel identification | arXiv: 2506.00432
clarify contrastive preference reinforcement learning for untangling ambiguous q | arXiv: 2506.00388
classifier reconstruction through counterfactual-aware wasserstein prototypes | arXiv: 2512.10878
clients collaborate flexible differentially private federated learning with guar | arXiv: 2402.07002
clipping improves adam-norm and adagrad-norm when the noise is heavy-tailed | arXiv: 2406.04443
closed-form solutions a new perspective on solving differential equations | arXiv: 2405.14620
closed-loop long-horizon robotic planning via equilibrium sequence modeling | arXiv: 2410.01440
clustering properties of self-supervised learning | arXiv: 2501.18452
cocoa-mix confusion-and-confidence-aware mixture model for context optimization | arXiv: 2506.07484
cody counterfactual explainers for dynamic graphs | arXiv: 2403.16846
collaborative mean estimation among heterogeneous strategic agents individual ra | arXiv: 2407.15881
collapse-proof non-contrastive self-supervised learning | arXiv: 2410.04959
come together but not right now a progressive strategy to boost low-rank adaptat | arXiv: 2506.05713
comemo lvlms need image context with image memory | arXiv: 2506.06279
communicating activations between language model agents | arXiv: 2501.14082
commvq commutative vector quantization for kv cache compression | arXiv: 2506.18879
compact matrix quantum group equivariant neural networks | arXiv: 2311.06358
compelling relu networks to exhibit exponentially many linear regions at initial | arXiv: 2311.18022
compositional flows for 3d molecule and synthesis pathway co-design | arXiv: 2504.08051
compositional scene understanding through inverse generative modeling | arXiv: 2505.21780
comrecgc global graph counterfactual explainer through common recourse | arXiv: 2505.07081
concept-based unsupervised domain adaptation | arXiv: 2505.05195
conceptual belief-informed reinforcement learning | arXiv: 2410.01739
configurable preference tuning with rubric-guided synthetic data | arXiv: 2506.11702
conformal prediction as bayesian quadrature | arXiv: 2502.13228
confpo exploiting policy model confidence for critical token selection in prefer | arXiv: 2506.08712
connecting thompson sampling and ucb towards more efficient trade-offs between p | arXiv: 2505.02383
consistency in language models current landscape challenges and future direction | arXiv: 2505.00268
constant stepsize local gd for logistic regression acceleration by instability | arXiv: 2506.13974
constrained hamiltonian systems on observation-induced fiber bundles theory of s | arXiv: 2505.22824
context driving in-context learning for text removal and segmentation | arXiv: 2506.03799
context is key a benchmark for forecasting with essential textual information | arXiv: 2410.18959
context matters query-aware dynamic long sequence modeling of gigapixel images | arXiv: 2501.18984
context tuning for in-context optimization | arXiv: 2507.04221
contextures representations from contexts | arXiv: 2505.01557
continual reinforcement learning by planning with online world models | arXiv: 2507.09177
continualflow learning and unlearning with neural flow matching | arXiv: 2506.18747
continuous semi-implicit models | arXiv: 2506.06778
continuous visual autoregressive generation via score maximization | arXiv: 2505.07812
continuous-time analysis of heavy ball momentum in min-max games | arXiv: 2505.19537
controlling underestimation bias in constrained reinforcement learning for safe | arXiv: 2601.11953
convex markov games a new frontier for multi-agent reinforcement learning | arXiv: 2410.16600
cooperation of experts fusing heterogeneous information with large margin | arXiv: 2505.20853
core context aware transformers for long context language modeling | arXiv: 2412.12465
core knowledge deficits in multi-modal language models | arXiv: 2410.10855
corematching a co-adaptive sparse inference framework with token and neuron prun | arXiv: 2505.19235
correlated errors in large language models | arXiv: 2506.07962
costfilter-ad enhancing anomaly detection through matching cost filtering | arXiv: 2505.01476
counterfactual effect decomposition in multi-agent sequential decision making | arXiv: 2410.12539
counting in small transformers the delicate interplay between attention and feed | arXiv: 2407.11542
cover learning for large-scale topology representation | arXiv: 2503.09767
craftium bridging flexibility and efficiency for rich 3d single- and multi-agent | arXiv: 2407.03969
cross-environment cooperation enables zero-shot multi-agent coordination | arXiv: 2504.12714
cross-regularization adaptive model complexity through validation gradients | arXiv: 2506.19755
crow eliminating backdoors from large language models via internal consistency r | arXiv: 2411.12768
curse of high dimensionality issue in transformer for long-context modeling | arXiv: 2505.22107
curvature enhanced data augmentation for regression | arXiv: 2506.06853
customizing the inductive biases of softmax attention using structured matrices | arXiv: 2509.07963
cut out and replay a simple yet versatile strategy for multi-label online contin | arXiv: 2505.19680
d-fusion direct preference optimization for aligning diffusion models with visua | arXiv: 2505.22002
data-juicer sandbox a feedback-driven suite for multimodal data-model co-develop | arXiv: 2407.11784
datadecide how to predict best pretraining data with small experiments | arXiv: 2504.11393
dctdiff intriguing properties of image generative modeling in the dct space | arXiv: 2412.15032
de-antifake rethinking the protective perturbations against voice cloning attack | arXiv: 2507.02606
de-mark watermark removal in large language models | arXiv: 2410.13808
decision making under the exponential family distributionally robust optimisatio | arXiv: 2411.16829
decoding rewards in competitive games inverse game theory with entropy regulariz | arXiv: 2601.12707
deep electromagnetic structure design under limited evaluation budgets | arXiv: 2506.19384
deep learning is not so mysterious or different | arXiv: 2503.02113
deepseq high-throughput single-cell rna sequencing data labeling via web search- | arXiv: 2506.13817
defame dynamic evidence-based fact-checking with multimodal experts | arXiv: 2412.10510
defending lvlms against vision attacks through partial-perception supervision | arXiv: 2412.12722
deltashap explaining prediction evolutions in online patient monitoring with sha | arXiv: 2507.02342
democratic ai is possible the democracy levels framework shows how it might work | arXiv: 2411.09222
demystifying the paradox of importance sampling with an estimated history-depend | arXiv: 2505.22492
density ratio estimation-based bayesian optimization with semi-supervised learni | arXiv: 2305.15612
deprecating benchmarks criteria and framework | arXiv: 2507.06434
designing cyclic peptides via harmonic sde with atom-bond modeling | arXiv: 2505.21452
differentiable stellar atmospheres with physics-informed neural networks | arXiv: 2507.06357
diffuse everything multimodal diffusion models on arbitrary state spaces | arXiv: 2506.07903
diffusion adversarial post-training for one-step video generation | arXiv: 2501.08316
diffusion sampling correction via approximately 10 parameters | arXiv: 2411.06503
diffusion-vla generalizable and interpretable robot foundation model via self-ge | arXiv: 2412.03293
dilqr differentiable iterative linear quadratic regulator via implicit different | arXiv: 2506.17473
dipllm fine-tuning llm for strategic decision-making in diplomacy | arXiv: 2506.09655
direct discriminative optimization your likelihood-based visual generative model | arXiv: 2503.01103
directed graph grammars for sequence-based learning | arXiv: 2505.22949
discovering global false negatives on the fly for self-supervised contrastive le | arXiv: 2502.20612
discrepancy minimization in input-sparsity time | arXiv: 2210.12468
discrete neural algorithmic reasoning | arXiv: 2402.11628
discriminative policy optimization for token-level reward models | arXiv: 2505.23363
disentangling and integrating relational and sensory information in transformer | arXiv: 2405.16727
disparate conditional prediction in multiclass classifiers | arXiv: 2206.03234
diss-l-ect dissecting graph data with local euler characteristic transforms | arXiv: 2410.02622
distillation of discrete diffusion through dimensional correlations | arXiv: 2410.08709
distilling tool knowledge into language models via back-translated traces | arXiv: 2506.19171
distributed and decentralised training technical governance challenges in a shif | arXiv: 2507.07765
diverging preferences when do annotators disagree and do models know | arXiv: 2410.14632
diverse prototypical ensembles improve robustness to subpopulation shift | arXiv: 2505.23027
diversity by design leveraging distribution matching for offline model-based opt | arXiv: 2501.18768
divide and conquer grounding llms as efficient decision-making agents via offlin | arXiv: 2505.19761
diving into self-evolving training for multimodal reasoning | arXiv: 2412.17451
dlp dynamic layerwise pruning in large language models | arXiv: 2505.23807
do multiple instance learning models transfer | arXiv: 2506.09022
do not mimic my voice speaker identity unlearning for zero-shot text-to-speech | arXiv: 2507.20140
do sparse autoencoders generalize a case study of answerability | arXiv: 2502.19964
do vision-language models really understand visual language | arXiv: 2410.00193
does data scaling lead to visual compositional generalization | arXiv: 2507.07102
does graph prompt work a data operation perspective with theoretical analysis | arXiv: 2410.01635
dont be so negative score-based generative modeling with oracle-assisted guidanc | arXiv: 2307.16463
dont lag rag training-free adversarial detection using rag | arXiv: 2504.04858
doubly protected estimation for survival outcomes utilizing external controls fo | arXiv: 2410.18409
doubly robust fusion of many treatments for policy learning | arXiv: 2505.08092
dpo meets ppo reinforced token optimization for rlhf | arXiv: 2404.18922
drag data reconstruction attack using guided diffusion | arXiv: 2509.11724
dragon guard llm unlearning in context via negative detection and reasoning | arXiv: 2511.05784
drivegpt scaling autoregressive behavior models for driving | arXiv: 2412.14415
dsp dynamic sequence parallelism for multi-dimensional transformers | arXiv: 2403.10266
dssd efficient edge-device llm deployment and collaborative inference via distri | arXiv: 2507.12000
dual form complementary masking for domain-adaptive image segmentation | arXiv: 2507.12008
dynamic benchmarking of reasoning capabilities in code large language models und | arXiv: 2503.04149
dynamic mixture of curriculum lora experts for continual multimodal instruction | arXiv: 2506.11672
dynamical phases of short-term memory mechanisms in rnns | arXiv: 2502.17433
e-lda toward interpretable lda topic models with strong guarantees in logarithmi | arXiv: 2506.07747
easyinv toward fast and better ddim inversion | arXiv: 2408.05159
eccdnamamba a pre-trained model for ultra-long eccdna sequence analysis | arXiv: 2506.18940
editable noise map inversion encoding target-image into noise for high-fidelity | arXiv: 2509.25776
eeg-language pretraining for highly label-efficient clinical phenotyping | arXiv: 2409.07480
efficient and robust semantic image communication via stable cascade | arXiv: 2507.17416
efficient curvature-aware hypergradient approximation for bilevel optimization | arXiv: 2505.02101
efficient diffusion models for symmetric manifolds | arXiv: 2505.21640
efficient generative modeling with residual vector quantization-based tokens | arXiv: 2412.10208
efficient length-generalizable attention via causal retrieval for long-context l | arXiv: 2410.01651
efficient logit-based knowledge distillation of deep spiking neural networks for | arXiv: 2501.15925
efficient molecular conformer generation with so3-averaged flow matching and ref | arXiv: 2507.09785
efficient network automatic relevance determination | arXiv: 2506.12352
efficient noise calculation in deep learning-based mri reconstructions | arXiv: 2505.02007
efficient optimization with orthogonality constraint a randomized riemannian sub | arXiv: 2505.12378
efficient quantification of multimodal interaction at sample level | arXiv: 2506.17248
efficient robotic policy learning via latent space backward planning | arXiv: 2505.06861
efficoder enhancing code generation in large language models through efficiency- | arXiv: 2410.10209
egoprivacy what your first-person camera says about you | arXiv: 2506.12258
eigenspectrum analysis of neural networks without aspect ratio bias | arXiv: 2506.06280
elemental interactive learning from demonstrations and vision-language models fo | arXiv: 2411.18825
elmo efficiency via low-precision and peak memory optimization in large output s | arXiv: 2510.11168
elucidating flow matching ode dynamics with respect to data geometries and denoi | arXiv: 2412.18730
elucidating the design space of multimodal protein language models | arXiv: 2504.11454
embedding safety into rl a new take on trust region methods | arXiv: 2411.02957
emergence in non-neural models grokking modular arithmetic via average gradient | arXiv: 2407.20199
emergent misalignment narrow finetuning can produce broadly misaligned llms | arXiv: 2502.17424
emergent symbolic mechanisms support abstract reasoning in large language models | arXiv: 2502.20332
empirical privacy variance | arXiv: 2503.12314
empower structure-based molecule optimization with gradient guided bayesian flow | arXiv: 2411.13280
enhancing certified robustness via block reflector orthogonal layers and logit a | arXiv: 2505.15174
enhancing cooperative multi-agent reinforcement learning with state modelling an | arXiv: 2505.05262
enhancing decision-making of large language models via actor-critic | arXiv: 2506.06376
enhancing parallelism in decentralized stochastic convex optimization | arXiv: 2506.00961
enhancing rating-based reinforcement learning to effectively leverage feedback f | arXiv: 2506.12822
enhancing statistical validity and power in hybrid controlled trials a randomiza | arXiv: 2410.11713
enhancing target-unspecific tasks through a features matrix | arXiv: 2505.03414
enigma interactive tools substantially assist lm agents in finding security vuln | arXiv: 2409.16165
epicoder encompassing diversity and complexity in code generation | arXiv: 2501.04694
epsilon-vae denoising as visual decoding | arXiv: 2410.04081
ergodic generative flows | arXiv: 2505.03561
erwin a tree-based hierarchical transformer for large-scale physical systems | arXiv: 2502.17019
estimating causal effects in gaussian linear scms with finite data | arXiv: 2601.04673
etta elucidating the design space of text-to-audio models | arXiv: 2412.19351
evaluating deepfake detectors in the wild | arXiv: 2507.21905
evaluating judges as evaluators the jetts benchmark of llm-as-judges as test-tim | arXiv: 2504.15253
evaluating llms across multi-cognitive levels from medical knowledge mastery to | arXiv: 2506.08349
evaluating morphological alignment of tokenizers in 70 languages | arXiv: 2507.06378
evaluating neuron explanations a unified framework with sanity checks | arXiv: 2506.05774
evaluating retrieval-augmented generation agents for autonomous scientific disco | arXiv: 2507.07155
event-aware sentiment factors from llm-augmented financial tweets a transparent | arXiv: 2508.07408
evolve evaluating and optimizing llms for in-context exploration | arXiv: 2410.06238
evolving prompts in-context an open-ended self-replicating perspective | arXiv: 2506.17930
evomesh adaptive physical simulation with hierarchical graph evolutions | arXiv: 2410.03779
exlm rethinking the impact of mask tokens in masked language models | arXiv: 2501.13397
exogenous isomorphism for counterfactual identifiability | arXiv: 2505.02212
expert evaluation of llm world models a high-t c superconductivity case study | arXiv: 2511.03782
explaining fast and slow abstraction and refinement of provable explanations | arXiv: 2506.08505
exploiting similarity for computation and communication-efficient decentralized | arXiv: 2506.05791
explora parameter-efficient extended pre-training to adapt vision transformers u | arXiv: 2406.10973
exploring large action sets with hyperspherical embeddings using von mises-fishe | arXiv: 2507.00518
exploring position encoding in diffusion u-net for training-free high-resolution | arXiv: 2503.09830
expressive score-based priors for distribution matching with geometry-preserving | arXiv: 2506.14607
extreme value policy optimization for safe reinforcement learning | arXiv: 2601.12008
fast and robust task sampling with posterior and diversity synergies for adaptiv | arXiv: 2504.19139
fastcav efficient computation of concept activation vectors for explaining deep | arXiv: 2505.17883
faster and stronger when ann-snn conversion meets parallel spiking calculation | arXiv: 2412.13610
faster rates for private adversarial bandits | arXiv: 2505.21790
featsharp your vision model features sharper | arXiv: 2502.16025
feature learning beyond the lazy-rich dichotomy insights from representational g | arXiv: 2503.18114
federated in-context learning iterative refinement for improved answer quality | arXiv: 2506.07440
fedrag a framework for fine-tuning retrieval-augmented generation systems | arXiv: 2506.09200
fedswa improving generalization in federated learning with highly heterogeneous | arXiv: 2507.20016
fedtail federated long-tailed domain generalization with sharpness-guided gradie | arXiv: 2506.08518
feedforward few-shot species range estimation | arXiv: 2502.14977
ferret federated full-parameter tuning at scale for large language models | arXiv: 2409.06277
few-shot learner generalizes across ai-generated image detection | arXiv: 2501.08763
fg-clip fine-grained visual and textual alignment | arXiv: 2505.05071
fgfp a fractional gaussian filter and pruning for deep neural networks compressi | arXiv: 2507.22527
ficgcn unveiling the homomorphic encryption efficiency from irregular graph conv | arXiv: 2506.10399
fine-grained captioning of long videos through scene graph consolidation | arXiv: 2502.16427
finetuning stellar spectra foundation models with lora | arXiv: 2507.20972
fishers for free approximating the fisher information matrix by recycling the sq | arXiv: 2507.18807
fixed-confidence multiple change point identification under bandit feedback | arXiv: 2507.08994
fixing the loose brake exponential-tailed stopping time in best arm identificati | arXiv: 2411.01808
flam frame-wise language-audio modeling | arXiv: 2505.05335
flat-lora low-rank adaptation over a flat loss landscape | arXiv: 2409.14396
flatquant flatness matters for llm quantization | arXiv: 2410.09426
fleet of agents coordinated problem solving with large language models | arXiv: 2405.06691
flexibility-conditioned protein structure design with flow matching | arXiv: 2508.18211
flexible tails for normalizing flows | arXiv: 2406.16971
flexiclip locality-preserving free-form character animation | arXiv: 2501.08676
flextok resampling images into 1d token sequences of flexible length | arXiv: 2502.13967
floe on-the-fly moe inference on memory-constrained gpu | arXiv: 2505.05950
flow of reasoning training llms for divergent reasoning with minimal examples | arXiv: 2406.05673
flowdrag 3d-aware drag-based image editing with mesh-guided deformation vector f | arXiv: 2507.08285
fmc formalization of natural language mathematical competition problems | arXiv: 2507.11275
foundation model insights and a multi-model approach for superior fine-grained o | arXiv: 2506.14473
foundation models for clinical records at health system scale | arXiv: 2507.00574
foundation molecular grammar multi-modal foundation models induce interpretable | arXiv: 2505.22948
founder grounding foundation models in world models for open-ended embodied deci | arXiv: 2507.12496
fourier position embedding enhancing attentions periodic extension for length ge | arXiv: 2412.17739
freemesh boosting mesh generation with coordinates merging | arXiv: 2505.13573
from black boxes to transparent minds evaluating and enhancing the theory of min | arXiv: 2506.14224
from debate to equilibrium belief-driven multi-agent llm reasoning via bayesian | arXiv: 2506.08292
from language models over tokens to language models over characters | arXiv: 2412.03719
from logits to hierarchies hierarchical clustering made simple | arXiv: 2410.07858
from low rank gradient subspace stabilization to low-rank weights observations t | arXiv: 2407.11239
from passive to active reasoning can large language models ask the right questio | arXiv: 2506.08295
from rag to memory non-parametric continual learning for large language models | arXiv: 2502.14802
from token to rhythm a multi-scale approach for ecg-language pretraining | arXiv: 2506.21803
fsl-sage accelerating federated split learning via smashed activation gradient e | arXiv: 2505.23182
fully dynamic euclidean bi-chromatic matching in sublinear update time | arXiv: 2505.09010
fully heteroscedastic count regression with deep double poisson networks | arXiv: 2406.09262
function encoders a principled approach to transfer learning in hilbert spaces | arXiv: 2501.18373
function-space learning rates | arXiv: 2502.17405
function-to-style guidance of llms for code translation | arXiv: 2507.11083
g-sim generative simulations with large language models and gradient-free calibr | arXiv: 2506.09272
gaprompt geometry-aware point cloud prompt for 3d vision model | arXiv: 2505.04119
gaussian mixture flow matching models | arXiv: 2504.05304
gaussmarker robust dual-domain watermark for diffusion models | arXiv: 2506.11444
gcal adapting graph models to evolving domain shifts | arXiv: 2505.16860
general agents contain world models | arXiv: 2506.01622
generalization analysis for supervised contrastive representation learning under | arXiv: 2505.04937
generalization and robustness of the tilted empirical risk | arXiv: 2409.19431
generalization bounds via meta-learned model representations pac-bayes and sampl | arXiv: 2410.13577
generalization in federated learning a conditional mutual information framework | arXiv: 2503.04091
generalized interpolating discrete diffusion | arXiv: 2503.04482
generation from noisy examples | arXiv: 2501.04179
generative audio language modeling with continuous-valued tokens and masked next | arXiv: 2507.09834
generative social choice the next generation | arXiv: 2505.22939
genmol a drug discovery generalist with discrete diffusion | arXiv: 2501.06158
geometric contact flows contactomorphisms for dynamics and control | arXiv: 2506.17868
geometric generative modeling with noise-conditioned graph networks | arXiv: 2507.09391
geometric representation condition improves equivariant molecule generation | arXiv: 2410.03655
geometry-to-image synthesis-driven generative point cloud registration | arXiv: 2512.09407
glgenn a novel parameter-light equivariant neural networks architecture based on | arXiv: 2506.09625
global context-aware representation learning for spatially resolved transcriptom | arXiv: 2506.15698
global convergence and rich feature learning in l-layer infinite-width neural ne | arXiv: 2503.09565
goirl graph-oriented inverse reinforcement learning for multimodal trajectory pr | arXiv: 2506.21121
gptaq efficient finetuning-free quantization for asymmetric calibration | arXiv: 2504.02692
gpu-friendly and linearly convergent first-order methods for certifying optimal | arXiv: 2603.01306
gradient aligned regression via pairwise losses | arXiv: 2402.06104
gradual transition from bellman optimality operator to bellman operator in onlin | arXiv: 2506.05968
gram a generative foundation reward model for reward generalization | arXiv: 2506.14175
graph attention is not always beneficial a theoretical analysis of graph attenti | arXiv: 2412.15496
graph generative pre-trained transformer | arXiv: 2501.01073
graph-assisted stitching for offline hierarchical reinforcement learning | arXiv: 2506.07744
graph-constrained reasoning faithful reasoning on knowledge graphs with large la | arXiv: 2410.13080
graph-supported dynamic algorithm configuration for multi-objective combinatoria | arXiv: 2505.16471
graph4mm weaving multimodal learning with structural information | arXiv: 2510.16990
gravity-bench-v1 a benchmark on gravitational physics discovery for agents | arXiv: 2501.18411
griffin towards a graph-centric relational database foundation model | arXiv: 2505.05568
grokformer graph fourier kolmogorov-arnold transformers | arXiv: 2411.17296
grokking at the edge of linear separability | arXiv: 2410.04489
guardagent safeguard llm agents by a guard agent via knowledge-enabled reasoning | arXiv: 2406.09187
guidedquant large language model quantization via exploiting end loss guidance | arXiv: 2505.07004
gumiho a hybrid architecture to prioritize early tokens in speculative decoding | arXiv: 2503.10135
handling imbalanced pseudolabels for vision-language models with concept alignme | arXiv: 2505.02056
harmonica harmonizing training and inference for better feature caching in diffu | arXiv: 2410.01723
heavy-tailed linear bandits huber regression with one-pass update | arXiv: 2503.00419
hessian geometry of latent space in generative models | arXiv: 2506.10632
heterogeneous data game characterizing the model competition across multiple dat | arXiv: 2505.07688
hgot self-supervised heterogeneous graph neural network with optimal transport | arXiv: 2506.02619
hi robot open-ended instruction following with hierarchical vision-language-acti | arXiv: 2502.19417
hierarchical and collaborative llm-based control for multi-uav motion and commun | arXiv: 2506.06532
hierarchical masked autoregressive models with low-resolution token pivots | arXiv: 2505.20288
hierarchical refinement optimal transport to infinity and beyond | arXiv: 2503.03025
hierarchical reinforcement learning with targeted causal interventions | arXiv: 2507.04373
hierarchical reinforcement learning with uncertainty-guided diffusional subgoals | arXiv: 2505.21750
high dynamic range novel view synthesis with single exposure | arXiv: 2505.01212
high-resolution live fuel moisture content lfmc maps for wildfire risk from mult | arXiv: 2506.20132
how do transformers learn variable binding in symbolic programs | arXiv: 2505.20896
how far is video generation from world model a physical law perspective | arXiv: 2411.02385
how much can we forget about data contamination | arXiv: 2410.03249
how to move your dragon text-to-motion synthesis for large-vocabulary objects | arXiv: 2503.04257
how to set adamws weight decay as you scale model and dataset size | arXiv: 2405.13698
how to synthesize text data without model collapse | arXiv: 2412.14689
how transformers learn regular language recognition a theoretical study on train | arXiv: 2505.00926
hybrid quantum-classical multi-agent pathfinding | arXiv: 2501.14568
hyperband-based bayesian optimization for black-box prompt selection | arXiv: 2412.07820
hyperbolic-pde gnn spectral graph neural networks in the perspective of a system | arXiv: 2505.23014
hyperimts hypergraph neural network for irregular multivariate time series forec | arXiv: 2505.17431
i2moe interpretable multimodal interaction-aware mixture-of-experts | arXiv: 2505.19190
iclshield exploring and mitigating in-context learning backdoor attacks | arXiv: 2507.01321
identifying and understanding cross-class features in adversarial training | arXiv: 2506.05032
idpa instance decoupled prompt attention for incremental medical object detectio | arXiv: 2506.00406
if open source is to win it must go public | arXiv: 2507.09296
impact iterative mask-based parallel decoding for text-to-audio generation with | arXiv: 2506.00736
implementing adaptations for vision autoregressive model | arXiv: 2507.11441
importance corrected neural jko sampling | arXiv: 2407.20444
importance sampling for nonlinear models | arXiv: 2505.12353
improved and oracle-efficient online ell 1-multicalibration | arXiv: 2505.17365
improved exploration in gflownets via enhanced epistemic neural networks | arXiv: 2506.16313
improved generalization bounds for transductive learning by transductive local c | arXiv: 2309.16858
improved last-iterate convergence of shuffling gradient methods for nonsmooth co | arXiv: 2505.23056
improved learning via k-dtw a novel dissimilarity measure for curves | arXiv: 2505.23431
improved off-policy reinforcement learning in biological sequence design | arXiv: 2410.04461
improved sample complexity for private nonsmooth nonconvex optimization | arXiv: 2410.05880
improving continual learning performance and efficiency with auxiliary classifie | arXiv: 2403.07404
improving flow matching by aligning flow divergence | arXiv: 2602.00869
improving generalization with flat hilbert bayesian inference | arXiv: 2410.04196
improving llm agent planning with in-context learning via atomic fact augmentati | arXiv: 2506.09171
improving llm safety alignment with dual-objective optimization | arXiv: 2503.03710
improving memory efficiency for training kans via meta learning | arXiv: 2506.07549
improving model alignment through collective intelligence of open-source llms | arXiv: 2505.03059
improving rationality in the reasoning process of language models through self-p | arXiv: 2506.22920
improving the diffusability of autoencoders | arXiv: 2502.14831
improving the effective receptive field of message-passing neural networks | arXiv: 2505.23185
improving the variance of differentially private randomized experiments through | arXiv: 2308.00957
improving your model ranking on chatbot arena by vote rigging | arXiv: 2501.17858
imts is worth time times channel patches visual masked autoencoders for irregula | arXiv: 2505.22815
in-context adaptation to concept drift for learned database operations | arXiv: 2505.04404
in-context linear regression demystified training dynamics and mechanistic inter | arXiv: 2503.12734
incremental gradient descent with small epoch counts is surprisingly slow on ill | arXiv: 2506.04126
inductive gradient adjustment for spectral bias in implicit neural representatio | arXiv: 2410.13271
inference-time decomposition of activations itda a scalable approach to interpre | arXiv: 2505.17769
infocons identifying interpretable critical concepts in point clouds via informa | arXiv: 2505.19820
infosam fine-tuning the segment anything model from an information-theoretic per | arXiv: 2505.21920
infosem a deep generative model with informative priors for gene regulatory netw | arXiv: 2503.04483
instruction tuning of large language models for tabular data generation-in one d | arXiv: 2511.23220
instruction-following pruning for large language models | arXiv: 2501.02086
integer programming for generalized causal bootstrap designs | arXiv: 2410.21464
integrating intermediate layer optimization and projected gradient descent for s | arXiv: 2505.20789
interchangeable token embeddings for extendable vocabulary and alpha-equivalence | arXiv: 2410.17161
interior-point vanishing problem in semidefinite relaxations for neural network | arXiv: 2506.10269
internal causal mechanisms robustly predict language model out-of-distribution b | arXiv: 2505.11770
intlora integral low-rank adaptation of quantized diffusion models | arXiv: 2410.21759
invariance makes llm unlearning resilient even to unanticipated downstream fine- | arXiv: 2506.01339
investigating non-transitivity in llm-as-a-judge | arXiv: 2502.14074
is complex query answering really complex | arXiv: 2410.12537
is your llm-based multi-agent a reliable real-world planner exploring fraud dete | arXiv: 2505.16557
is your model fairly certain uncertainty-aware fairness evaluation for llms | arXiv: 2505.23996
isolated causal effects of natural language | arXiv: 2410.14812
it3 idempotent test-time training | arXiv: 2410.04201
joker joint optimization framework for lightweight kernel machines | arXiv: 2505.17765
k2ie kernel method-based kernel intensity estimators for inhomogeneous poisson p | arXiv: 2505.24704
kan-ad time series anomaly detection with kolmogorov-arnold networks | arXiv: 2411.00278
kbqa-o1 agentic knowledge base question answering with monte carlo tree search | arXiv: 2501.18922
kea keeping exploration alive by proactively coordinating exploration strategies | arXiv: 2503.18234
kelps a framework for verified multi-language autoformalization via semantic-syn | arXiv: 2507.08665
kernel-based unsupervised embedding alignment for enhanced visual representation | arXiv: 2506.02557
kinetic langevin diffusion for crystalline materials generation | arXiv: 2507.03602
la rosa enhancing llm efficiency via layerwise rotated sparse activation | arXiv: 2507.01299
label-efficient hyperspectral image classification via spectral film modulation | arXiv: 2512.03430
lacache ladder-shaped kv caching for efficient long-context modeling of large la | arXiv: 2507.14204
lada scalable label-specific clip adapter for continual learning | arXiv: 2505.23271
ladder-residual parallelism-aware architecture for accelerating large model infe | arXiv: 2501.06589
laion-c an out-of-distribution benchmark for web-scale vision models | arXiv: 2506.16950
langdaug langevin data augmentation for multi-source domain generalization in me | arXiv: 2505.19659
language model developers should report train-test overlap | arXiv: 2410.08385
language models over canonical byte-pair encodings | arXiv: 2506.07956
lapsum -- one method to differentiate them all ranking sorting and top-k selecti | arXiv: 2503.06242
large language model llm-enabled in-context learning for wireless network optimi | arXiv: 2408.00214
large language models are demonstration pre-selectors for themselves | arXiv: 2506.06033
large language models to diffusion finetuning | arXiv: 2501.15781
laser attention with exponential transformation | arXiv: 2411.03493
latent imputation before prediction a new computational paradigm for de novo pep | arXiv: 2505.17524
latent variable causal discovery under selection bias | arXiv: 2512.11219
latent variable estimation in bayesian black-litterman models | arXiv: 2505.02185
layer-wise alignment examining safety alignment across image encoder layers in v | arXiv: 2411.04291
layer-wise quantization for quantized optimistic dual averaging | arXiv: 2505.14371
ldmol a text-to-molecule diffusion model with structurally informative latent sp | arXiv: 2405.17829
learnable spatial-temporal positional encoding for link prediction | arXiv: 2506.08309
learning cascade ranking as one network | arXiv: 2503.09492
learning distances from data with normalizing flows and score matching | arXiv: 2407.09297
learning distribution-wise control in representation space for language models | arXiv: 2506.06686
learning dynamics under environmental constraints via measurement-induced bundle | arXiv: 2505.19521
learning invariant causal mechanism from vision-language models | arXiv: 2405.15289
learning mean field control on sparse graphs | arXiv: 2501.17079
learning mixtures of experts with em a mirror descent perspective | arXiv: 2411.06056
learning optimal multimodal information bottleneck representations | arXiv: 2505.19996
learning progress driven multi-agent curriculum | arXiv: 2205.10016
learning safe strategies for value maximizing buyers in uniform price auctions | arXiv: 2406.03674
learning safety constraints for large language models | arXiv: 2505.24445
learning single index models with diffusion priors | arXiv: 2505.21135
learning soft sparse shapes for efficient time-series classification | arXiv: 2505.06892
learning survival distributions with the asymmetric laplace distribution | arXiv: 2505.03712
learning time-aware causal representation for model generalization in evolving d | arXiv: 2506.17718
learning to incentivize in repeated principal-agent problems with adversarial ag | arXiv: 2505.23124
learning to plan reason for evaluation with thinking-llm-as-a-judge | arXiv: 2501.18099
learning to stop deep learning for mean field optimal stopping | arXiv: 2410.08850
learning to trust bellman updates selective state-adaptive regularization for of | arXiv: 2505.19923
learning utilities from demonstrations in markov decision processes | arXiv: 2409.17355
learning-augmented algorithms for mts with bandit access to multiple predictors | arXiv: 2506.05479
learning-augmented hierarchical clustering | arXiv: 2506.05495
lego sketch a scalable memory-augmented neural network for sketching data stream | arXiv: 2505.19561
lemon label error detection using multimodal neighbors | arXiv: 2407.18941
leveraging online olympiad-level math problems for llms training and contaminati | arXiv: 2501.14275
leveraging partial smiles validation scheme for enhanced drug design in reinforc | arXiv: 2505.00530
leveraging predictive equivalence in decision trees | arXiv: 2506.14143
leveraging skills from unlabeled prior data for efficient online exploration | arXiv: 2410.18076
lift the veil for the truth principal weights emerge after rank reduction for re | arXiv: 2506.00772
liger linearizing large language models to gated recurrent structures | arXiv: 2503.01496
lightgts a lightweight general time series forecasting model | arXiv: 2506.06005
lighthouse fast and precise distance to shoreline calculations from anywhere on | arXiv: 2506.18842
lightspeed geometric dataset distance via sliced optimal transport | arXiv: 2501.18901
lineflow a framework to learn active control of production lines | arXiv: 2505.06744
livs a pluralistic alignment dataset for inclusive public spaces | arXiv: 2503.01894
llava-reid selective multi-image questioner for interactive person re-identifica | arXiv: 2504.10174
llavaguard an open vlm-based framework for safeguarding vision datasets and mode | arXiv: 2406.05113
llm data selection and utilization via dynamic bi-level optimization | arXiv: 2507.16178
llm enhancers for gnns an analysis from the perspective of causal mechanism iden | arXiv: 2505.08265
llm social simulations are a promising research method | arXiv: 2504.02234
llm-srbench a new benchmark for scientific equation discovery with large languag | arXiv: 2504.10415
local manifold approximation and projection for manifold-aware diffusion plannin | arXiv: 2506.00867
localizing and mitigating memorization in image autoregressive models | arXiv: 2509.00488
log-sum-exponential estimator for off-policy evaluation and learning | arXiv: 2506.06873
long-form speech generation with spoken language models | arXiv: 2412.18603
long-short alignment for effective long-context modeling in llms | arXiv: 2506.11769
look twice before you answer memory-space visual retracing for hallucination mit | arXiv: 2410.03577
lora fine-tuning without gpus a cpu-efficient meta-generation framework for llms | arXiv: 2507.01806
lscd lomb-scargle conditioned diffusion for time series imputation | arXiv: 2506.17039
lyapunov learning at the onset of chaos | arXiv: 2506.12810
m3-jepa multimodal alignment via multi-gate moe based on the joint-embedding pre | arXiv: 2409.05929
m3hf multi-agent reinforcement learning from multi-phase human feedback of mixed | arXiv: 2503.02077
machine learning from explanations | arXiv: 2507.04788
machines and mathematical mutations using gnns to characterize quiver mutation c | arXiv: 2411.07467
make lora great again boosting lora with adaptive singular values and mixture-of | arXiv: 2502.16894
mapeval a map-based evaluation of geo-spatial reasoning in foundation models | arXiv: 2501.00316
marge improving math reasoning for llms with guided exploration | arXiv: 2505.12500
mastering massive multi-task reinforcement learning via mixture-of-expert decisi | arXiv: 2505.24378
mastering multiple-expert routing realizable h-consistency and strong guarantees | arXiv: 2506.20650
maximal update parametrization and zero-shot hyperparameter transfer for fourier | arXiv: 2506.19396
maximum coverage in turnstile streams with applications to fingerprinting measur | arXiv: 2504.18394
maximum total correlation reinforcement learning | arXiv: 2505.16734
medxpertqa benchmarking expert-level medical reasoning and understanding | arXiv: 2501.18362
meek models shall inherit the earth | arXiv: 2507.07931
merge-friendly post-training quantization for multi-target domain adaptation | arXiv: 2505.23651
merit maximum-normalized element-wise ratio for language model large-batch train | arXiv: 2508.20577
meta-black-box-optimization through offline q-function learning | arXiv: 2505.02010
metaagent automatically constructing multi-agent systems based on finite state m | arXiv: 2507.22606
metadata conditioning accelerates language model pre-training | arXiv: 2501.01956
mf-lal drug compound generation using multi-fidelity latent space active learnin | arXiv: 2410.11226
mib a mechanistic interpretability benchmark | arXiv: 2504.13151
mimicmotion high-quality human motion video generation with confidence-aware pos | arXiv: 2406.19680
mind the gap a practical attack on gguf quantization | arXiv: 2505.23786
mitigating over-squashing in graph neural networks by spectrum-preserving sparsi | arXiv: 2506.16110
mitigating plasticity loss in continual reinforcement learning by reducing churn | arXiv: 2506.00592
mixed-curvature decision trees and random forests | arXiv: 2410.13879
mixture of lookup experts | arXiv: 2503.15798
mixture-of-expert variational autoencoders for cross-modality embedding of type | arXiv: 2507.16817
mka memory-keyed attention for efficient long-context reasoning | arXiv: 2603.20586
mmedpo aligning medical vision-language models with clinical-aware multimodal pr | arXiv: 2412.06141
mminference accelerating pre-filling for long-context vlms via modality-aware pe | arXiv: 2504.16083
moda modular duplex attention for multimodal perception cognition and emotion un | arXiv: 2507.04635
model immunization from a condition number perspective | arXiv: 2505.23760
model swarms collaborative search to adapt llm experts via swarm intelligence | arXiv: 2410.11163
modeling all-atom glycan structures via hierarchical message passing and multi-s | arXiv: 2506.01376
modeling user behavior from adaptive surveys with supplemental context | arXiv: 2507.20919
modern methods in associative memory | arXiv: 2507.06211
modified k-means algorithm with local optimality guarantees | arXiv: 2506.06990
modulated diffusion accelerating generative modeling with modulated quantization | arXiv: 2506.22463
moh multi-head attention as mixture-of-head attention | arXiv: 2410.11842
moma modulating mamba for adapting image foundation models to video recognition | arXiv: 2506.23283
moragent parameter efficient agent tuning with mixture-of-roles | arXiv: 2512.21708
morphtok morphologically grounded tokenization for indian languages | arXiv: 2504.10335
morse dual-sampling for lossless acceleration of diffusion models | arXiv: 2506.18251
mpf aligning and debiasing language models post deployment via multi perspective | arXiv: 2507.02595
mpo an efficient post-processing framework for mixing diverse preference alignme | arXiv: 2502.18699
mtl-ue learning to learn nothing for multi-task learning | arXiv: 2505.05279
multidimensional adaptive coefficient for inference trajectory optimization in f | arXiv: 2404.14161
multiple-policy evaluation via density estimation | arXiv: 2404.00195
multivariate conformal selection | arXiv: 2505.00917
musecontrollite multifunctional music generation with lightweight conditioners | arXiv: 2506.18729
near optimal best arm identification for clustered bandits | arXiv: 2505.10147
near optimal decision trees in a split second | arXiv: 2502.15988
near-optimal consistency-robustness trade-offs for learning-augmented online kna | arXiv: 2406.18752
nearly optimal sample complexity for learning with label proportions | arXiv: 2505.05355
negmerge sign-consensual weight merging for machine unlearning | arXiv: 2410.05583
neighbour-driven gaussian process variational autoencoders for scalable structur | arXiv: 2505.16481
network sparsity unlocks the scaling potential of deep reinforcement learning | arXiv: 2506.17204
neural augmented kalman filters for road network assisted gnss positioning | arXiv: 2507.00654
neural graph matching improves retrieval augmented generation in molecular machi | arXiv: 2502.17874
neural stochastic differential equations on compact state spaces theory methods | arXiv: 2508.17090
neurontune towards self-guided spurious bias mitigation | arXiv: 2505.24048
neutral residues revisiting adapters for model extension | arXiv: 2410.02744
new interaction paradigm for complex eda software leveraging gpt | arXiv: 2307.14740
nextlong toward effective long-context training without long documents | arXiv: 2501.12766
no soundness in the real world on the challenges of the verification of deployed | arXiv: 2506.01054
non-stationary online learning for curved losses improved dynamic regret via mix | arXiv: 2506.10616
nonparametric identification of latent concepts | arXiv: 2510.00136
nonparametric modern hopfield models | arXiv: 2404.03900
nonparametric teaching for graph property learners | arXiv: 2505.14170
normalizing flows are capable generative models | arXiv: 2412.06329
not all explanations for deep learning phenomena are equally valuable | arXiv: 2506.23286
ntpp generative speech language modeling for dual-channel spoken dialogue via ne | arXiv: 2506.00975
of mice and machines a comparison of learning between real world mice and rl age | arXiv: 2505.12204
olica efficient structured pruning of large language models without retraining | arXiv: 2506.08436
omniarch building foundation model for scientific computing | arXiv: 2402.16014
omniaudio generating spatial audio from 360-degree video | arXiv: 2504.14906
omnibal towards fast instruction-tuning for vision-language models via omniverse | arXiv: 2407.20761
on differential privacy for adaptively solving search problems via sketching | arXiv: 2506.05503
on expressive power of looped transformers theoretical analysis and enhancement | arXiv: 2410.01405
on fine-grained distinct element estimation | arXiv: 2506.22608
on measuring long-range interactions in graph neural networks | arXiv: 2506.05971
on temperature scaling and conformal prediction of deep classifiers | arXiv: 2402.05806
on the clean generalization and robust overfitting in adversarial training from | arXiv: 2306.01271
on the dynamic regret of following the regularized leader optimism with history | arXiv: 2505.22899
on the effect of uncertainty on layer-wise inference dynamics | arXiv: 2507.06722
on the importance of gaussianizing representations | arXiv: 2505.00685
on the power of context-enhanced learning in llms | arXiv: 2503.01821
on the robustness of reward models for language model alignment | arXiv: 2505.07271
on the role of label noise in the feature learning process | arXiv: 2505.18909
on the vulnerability of applying retrieval-augmented generation within knowledge | arXiv: 2409.17275
on understanding attention-based in-context learning for categorical data | arXiv: 2405.17248
one image is worth a thousand words a usability preservable text-image collabora | arXiv: 2505.11131
one wave to explain them all a unifying perspective on feature attribution | arXiv: 2410.01482
online pre-training for offline-to-online reinforcement learning | arXiv: 2507.08387
online sparsification of bipartite-like clusters in graphs | arXiv: 2508.05437
ood-chameleon is algorithm selection for ood generalization learnable | arXiv: 2410.02735
open source planning control system with language agents for autonomous scientif | arXiv: 2507.07257
open your eyes vision enhances message passing neural networks in link predictio | arXiv: 2505.08266
open-det an efficient learning framework for open-ended detection | arXiv: 2505.20639
optimal and practical batched linear bandit algorithm | arXiv: 2507.08438
optimal auction design in the joint advertising | arXiv: 2507.07418
optimal sensor scheduling and selection for continuous-discrete kalman filtering | arXiv: 2507.11240
optimization over sparse support-preserving sets two-step projection with global | arXiv: 2506.08558
optimizing language models for inference time objectives using reinforcement lea | arXiv: 2503.19595
or-bench an over-refusal benchmark for large language models | arXiv: 2405.20947
origin identification for text-guided image-to-image diffusion models | arXiv: 2501.02376
orthorank token selection via sink token orthogonality for efficient llm inferen | arXiv: 2507.03865
out-of-distribution detection methods answer the wrong questions | arXiv: 2507.01831
outlier gradient analysis efficiently identifying detrimental training samples f | arXiv: 2405.03869
overcoming multi-step complexity in multimodal theory-of-mind reasoning a scalab | arXiv: 2506.01301
pac learning with improvements | arXiv: 2503.03184
pak-ucb contextual bandit an online learning approach to prompt-aware selection | arXiv: 2410.13287
parallelcomp parallel long-context compressor for length extrapolation | arXiv: 2502.14317
parameter-efficient fine-tuning of state space models | arXiv: 2410.09016
parity requires unified input dependence and negative eigenvalues in ssms | arXiv: 2508.07395
parm multi-objective test-time alignment via preference-aware autoregressive rew | arXiv: 2505.06274
parrot multilingual visual instruction tuning | arXiv: 2406.02539
pde-transformer efficient and versatile transformers for physics simulations | arXiv: 2505.24717
pencil long thoughts with short memory | arXiv: 2503.14337
peptune de novo generation of therapeutic peptides with multi-objective-guided d | arXiv: 2412.17780
performance plateaus in inference-time scaling for text-to-image diffusion witho | arXiv: 2506.12633
permutation equivariant neural networks for symmetric tensors | arXiv: 2503.11276
persistent topological features in large language models | arXiv: 2410.11042
pessimism principle can be effective towards a framework for zero-shot transfer | arXiv: 2505.18447
phantomwiki on-demand datasets for reasoning and retrieval evaluation | arXiv: 2502.20377
physicsnerf physics-guided 3d reconstruction from sparse views | arXiv: 2505.23481
pigdreamer privileged information guided world models for safe partially observa | arXiv: 2508.02159
piloting structure-based drug design via modality-specific optimal schedule | arXiv: 2505.07286
poisonbench assessing large language model vulnerability to data poisoning | arXiv: 2410.08811
polyconf unlocking polymer conformation generation through hierarchical generati | arXiv: 2504.08859
popri private federated learning using preference-optimized synthetic data | arXiv: 2504.16438
poqd performance-oriented query decomposer for multi-vector retrieval | arXiv: 2505.19189
position ai evaluation should learn from how we test humans | arXiv: 2306.10512
position all current generative fidelity and diversity metrics are flawed | arXiv: 2505.22450
position causal machine learning requires rigorous synthetic experiments for bro | arXiv: 2508.08883
position dont use the clt in llm evals with fewer than a few hundred datapoints | arXiv: 2503.01747
position lifetime tuning is incompatible with continual reinforcement learning | arXiv: 2404.02113
position solve layerwise linear models first to understand neural dynamical phen | arXiv: 2502.21009
position the future of bayesian prediction is prior-fitted | arXiv: 2505.23947
position theory of mind benchmarks are broken for large language models | arXiv: 2412.19726
position uncertainty quantification needs reassessment for large-language model | arXiv: 2505.22655
position we need an algorithmic understanding of generative ai | arXiv: 2507.07544
positional attention expressivity and learnability of algorithmic computation | arXiv: 2410.01686
positional encoding meets persistent homology on graphs | arXiv: 2506.05814
ppo-mi efficient black-box model inversion via proximal policy optimization | arXiv: 2502.14370
practical principles for ai cost and compute accounting | arXiv: 2502.15873
prediction via shapley value regression | arXiv: 2505.04775
prediction-powered adaptive shrinkage estimation | arXiv: 2502.14166
predictive data selection the data that predicts is the data that teaches | arXiv: 2503.00808
preference adaptive and sequential text-to-image generation | arXiv: 2412.10419
preference optimization for combinatorial optimization problems | arXiv: 2505.08735
principal-agent bandit games with self-interested and exploratory learning agent | arXiv: 2412.16318
principled algorithms for optimizing generalized metrics in binary classificatio | arXiv: 2512.23133
privacy amplification through synthetic data insights from linear regression | arXiv: 2506.05101
privacy-shielded image compression defending against exploitation from vision-la | arXiv: 2506.15201
private model personalization revisited | arXiv: 2506.19220
probabilistic interactive 3d segmentation with hierarchical neural processes | arXiv: 2505.01726
probably approximately global robustness certification | arXiv: 2511.06495
product of experts with llms boosting performance on arc is a matter of perspect | arXiv: 2505.07859
progressive tempering sampler with diffusion | arXiv: 2506.05231
promoting ensemble diversity with interactive bayesian distributional robustness | arXiv: 2506.07247
proofcompass enhancing specialized provers with llm guidance | arXiv: 2507.14335
protein structure tokenization benchmarking and new recipe | arXiv: 2503.00089
protriever end-to-end differentiable protein homology search for fitness predict | arXiv: 2506.08954
provable benefit of random permutations over uniform sampling in stochastic coor | arXiv: 2505.23152
provable in-context vector arithmetic via retrieving task concepts | arXiv: 2508.09820
provable maximum entropy manifold exploration via diffusion models | arXiv: 2506.15385
provably cost-sensitive adversarial defense via randomized smoothing | arXiv: 2310.08732
provably efficient algorithm for best scoring rule identification in online prin | arXiv: 2505.17379
provably improving generalization of few-shot models with synthetic data | arXiv: 2505.24190
proxy-fda proxy-based feature distribution alignment for fine-tuning vision foun | arXiv: 2505.24088
putnam-axiom a functional and static benchmark for measuring higher level mathem | arXiv: 2508.08292
q-resafe assessing safety risks and quantization-aware safety patching for quant | arXiv: 2506.20251
qmamba on first exploration of vision mamba for image quality assessment | arXiv: 2406.09546
quadratic upper bound for boosting robustness | arXiv: 2601.13645
quantum algorithms for finite-horizon markov decision processes | arXiv: 2508.05712
quantum optimization via gradient-based hamiltonian descent | arXiv: 2505.14670
quest enhancing estimates of quantile-based distributional measures using model | arXiv: 2507.05220
qure query-relevant retrieval through hard negative sampling in composed image r | arXiv: 2507.12416
r3dm enabling role discovery and diversity through dynamics models in multi-agen | arXiv: 2505.24265
radio rate-distortion optimization for large language model compression | arXiv: 2505.03031
raising the bar investigating the values of large language models via generative | arXiv: 2406.14230
random feature representation boosting | arXiv: 2501.18283
random initialization of gated sparse adapters | arXiv: 2511.01794
random registers for cross-domain few-shot learning | arXiv: 2506.02843
randomized dimensionality reduction for euclidean maximization and diversity mea | arXiv: 2506.00165
ranked entropy minimization for continual test-time adaptation | arXiv: 2505.16441
ranked from within ranking large multimodal models without labels | arXiv: 2412.06461
rapid long-context inference with retrieval-augmented speculative decoding | arXiv: 2502.20330
raptor scalable train-free embeddings for 3d medical volumes leveraging pretrain | arXiv: 2507.08254
rate causal explainability of reward models with imperfect counterfactuals | arXiv: 2410.11348
re-imagine symbolic benchmark synthesis for reasoning evaluation | arXiv: 2506.15455
re-ranking reasoning context with tree search makes large vision-language models | arXiv: 2506.07785
reactivation empirical ntk dynamics under task shifts | arXiv: 2507.16039
reasoning limitations of multimodal large language models a case study of bongar | arXiv: 2411.01173
reasoning through execution unifying process and outcome rewards for code genera | arXiv: 2412.15118
recommendations and reporting checklist for rigorous transparent human baselines | arXiv: 2506.13776
recommendations with sparse comparison data provably fast convergence for noncon | arXiv: 2502.20033
refersplat referring segmentation in 3d gaussian splatting | arXiv: 2508.08252
reframe layer caching for accelerated inference in real-time rendering | arXiv: 2506.13814
regress dont guess -- a regression-like loss on number tokens for language model | arXiv: 2411.02083
regression for the mean auto-evaluation and inference with few labels through po | arXiv: 2411.12665
reimagining parameter space exploration with diffusion models | arXiv: 2506.17807
rejecting hallucinated state targets during planning | arXiv: 2410.07096
relative error fair clustering in the weak-strong oracle model | arXiv: 2506.12287
reliable algorithm selection for machine learning-guided design | arXiv: 2503.20767
representation shattering in transformers a synthetic study with knowledge editi | arXiv: 2410.17194
representative language generation | arXiv: 2505.21819
resampling augmentation for time series contrastive learning application to remo | arXiv: 2506.18587
researchtown simulator of human research community | arXiv: 2412.17767
residual matrix transformers scaling the size of the residual stream | arXiv: 2506.22696
restoregrad signal restoration using conditional denoising diffusion models with | arXiv: 2502.13574
rethink the role of deep learning towards large-scale quantum systems | arXiv: 2505.13852
rethinking addressing in language models via contexualized equivariant positiona | arXiv: 2501.00712
rethinking aleatoric and epistemic uncertainty | arXiv: 2412.20892
rethinking explainable machine learning as applied statistics | arXiv: 2402.02870
rethinking external slow-thinking from snowball errors to probability of correct | arXiv: 2501.15602
rethinking the bias of foundation model under long-tailed distribution | arXiv: 2501.15955
rethinking the stability-plasticity trade-off in continual learning from an arch | arXiv: 2506.03951
retraining with predicted hard labels provably increases model accuracy | arXiv: 2406.11206
retraining-free merging of sparse moe via hierarchical clustering | arXiv: 2410.08589
revealing weaknesses in text watermarking through self-information rewrite attac | arXiv: 2505.05190
review remask refine r3 process-guided block diffusion for text generation | arXiv: 2507.08018
revise learning to refine at test-time via intrinsic self-verification | arXiv: 2502.14565
revisiting continuity of image tokens for cross-domain few-shot learning | arXiv: 2506.03110
revisiting diffusion models from generative pre-training to one-step generation | arXiv: 2506.09376
revisiting instance-optimal cluster recovery in the labeled stochastic block mod | arXiv: 2306.12968
revisiting the predictability of performative social events | arXiv: 2503.11713
revisiting unbiased implicit variational inference | arXiv: 2506.03839
revolve optimizing ai systems by tracking response evolution in textual optimiza | arXiv: 2412.03092
reward-augmented data enhances direct preference alignment of llms | arXiv: 2410.08067
reward-free world models for online imitation learning | arXiv: 2410.14081
riflex a free lunch for length extrapolation in video diffusion transformers | arXiv: 2502.15894
right now wrong then non-stationary direct preference optimization under prefere | arXiv: 2407.18676
risk and cross validation in ridge regression with correlated samples | arXiv: 2408.04607
rlthf targeted human feedback for llm alignment | arXiv: 2502.13417
robot-gated interactive imitation learning with adaptive intervention mechanism | arXiv: 2506.09176
robust learning of diverse code edits | arXiv: 2503.03656
robust multi-bit text watermark with llm-based paraphrasers | arXiv: 2412.03123
robust multimodal large language models against modality conflict | arXiv: 2507.07151
robust noise attenuation via adaptive pooling of transformer outputs | arXiv: 2506.09215
robust offline reinforcement learning with linearly structured f-divergence regu | arXiv: 2411.18612
rocketkv accelerating long-context llm inference via two-stage kv cache compress | arXiv: 2502.14051
roll the dice look before you leap going beyond the creative limits of next-toke | arXiv: 2504.15266
rollingq reviving the cooperation dynamics in multimodal transformer | arXiv: 2506.11465
rulebreakers challenging llms at the crossroads between formal logic and human-l | arXiv: 2410.16502
runtime analysis of evolutionary nas for multiclass classification | arXiv: 2506.06019
sada stability-guided adaptive diffusion acceleration | arXiv: 2507.17135
saebench a comprehensive benchmark for sparse autoencoders in language model int | arXiv: 2503.09532
safe delta consistently preserving safety when fine-tuning llms on diverse datas | arXiv: 2505.12038
safe finding sparse and flat minima to improve pruning | arXiv: 2506.06866
safemap robust hd map construction from incomplete observations | arXiv: 2507.00861
safer a calibrated risk-aware multimodal recommendation model for dynamic treatm | arXiv: 2506.06649
safety alignment can be not superficial with explicit safety signals | arXiv: 2505.17072
safety certificate against latent variables with partially unidentifiable dynami | arXiv: 2506.17927
safetyanalyst interpretable transparent and steerable safety moderation for ai b | arXiv: 2410.16665
sample complexity of distributionally robust off-dynamics reinforcement learning | arXiv: 2511.05396
sample efficient demonstration selection for in-context learning | arXiv: 2506.08607
sampling from binary quadratic distributions via stochastic localization | arXiv: 2505.19438
sassha sharpness-aware adaptive second-order optimization with stable hessian ap | arXiv: 2502.18153
scalable equilibrium sampling with sequential boltzmann generators | arXiv: 2502.18462
scalable generation of spatial transcriptomics from histology images via whole-s | arXiv: 2506.05361
scalable non-equivariant 3d molecule generation via rotational alignment | arXiv: 2506.10186
scaling inference-efficient language models | arXiv: 2501.18107
scaling large motion models with million-level human motions | arXiv: 2410.03311
scaling value iteration networks to 5000 layers for extreme long-term planning | arXiv: 2406.08404
scaling video-language models to 10k frames via hierarchical differential distil | arXiv: 2504.02438
score matching with missing data | arXiv: 2506.00557
scssl-bench benchmarking self-supervised learning for single-cell data | arXiv: 2506.10031
sdp-crown efficient bound propagation for neural network verification with tight | arXiv: 2506.06665
se3-equivariant diffusion policy in spherical fourier space | arXiv: 2507.01723
secemb sparsity-aware secure federated learning of on-device recommender system | arXiv: 2505.12453
self-consistency preference optimization | arXiv: 2411.04109
self-disentanglement and re-composition for cross-domain few-shot segmentation | arXiv: 2506.02677
self-organizing visual prototypes for non-parametric representation learning | arXiv: 2505.21533
semantic shift estimation via dual-projection and classifier reconstruction for | arXiv: 2503.05423
sensei semantic exploration guided by foundation models to learn versatile world | arXiv: 2503.01584
separating knowledge and perception with procedural data | arXiv: 2508.11697
sepllm accelerate large language models by compressing one segment into one sepa | arXiv: 2412.12094
set valued predictions for robust domain generalization | arXiv: 2507.03146
sgd jittering a training strategy for robust and accurate model-based architectu | arXiv: 2410.14667
shielded diffusion generating novel and diverse images using sparse repellency | arXiv: 2410.06025
simple and critical iterative denoising a recasting of discrete diffusion in gra | arXiv: 2503.21592
simplemix frustratingly simple mixing of off- and on-policy data in language mod | arXiv: 2505.02363
sk-vqa synthetic knowledge generation at scale for training context-augmented mu | arXiv: 2406.19593
sketch to adapt fine-tunable sketches for efficient llm adaptation | arXiv: 2410.06364
sketch-plan-generalize learning and planning with neuro-symbolic programmatic re | arXiv: 2404.07774
sliding puzzles gym a scalable benchmark for state representation in visual rein | arXiv: 2410.14038
slim one-shot quantization and sparsity with low-rank approximation for llm weig | arXiv: 2410.09615
slimllm accurate structured pruning for large language models | arXiv: 2505.22689
smoothed preference optimization via renoise inversion for aligning diffusion mo | arXiv: 2506.02698
soft reasoning navigating solution spaces in large language models through contr | arXiv: 2505.24688
softmax is not enough for sharp size generalisation | arXiv: 2410.01104
solving probabilistic verification problems of neural networks using branch and | arXiv: 2405.17556
solving zero-sum convex markov games | arXiv: 2506.16120
sorbet a neuromorphic hardware-compatible transformer-based spiking language mod | arXiv: 2409.15298
sortformer a novel approach for permutation-resolved speaker supervision in spee | arXiv: 2409.06656
sounding that object interactive object-aware image to audio generation | arXiv: 2506.04214
space your genomic profile predictor is a powerful dna foundation model | arXiv: 2506.01833
sparse causal discovery with generative intervention for unsupervised graph doma | arXiv: 2507.07621
sparse spectral training and inference on euclidean and hyperbolic neural networ | arXiv: 2405.15481
sparse training from random initialization aligning lottery ticket masks using w | arXiv: 2505.05143
sparse-pivot dynamic correlation clustering for node insertions | arXiv: 2507.01830
sparselora accelerating llm fine-tuning with contextual sparsity | arXiv: 2506.16500
sparsevlm visual token sparsification for efficient vision-language model infere | arXiv: 2410.04417
speculative decoding in decentralized llm inference turning communication latenc | arXiv: 2511.11733
sphinx structural prediction using hypergraph inference network | arXiv: 2410.03208
spikevideoformer an efficient spike-driven video transformer with hamming attent | arXiv: 2505.10352
star attention efficient llm inference over long sequences | arXiv: 2411.17116
star learning diverse robot skill abstractions through rotation-augmented vector | arXiv: 2506.03863
statistical and computational guarantees of kernel max-sliced wasserstein distan | arXiv: 2405.15441
stealing that free lunch exposing the limits of dyna-style reinforcement learnin | arXiv: 2412.14312
stealix model stealing via prompt evolution | arXiv: 2506.05867
steer llm latents for hallucination detection | arXiv: 2503.01917
steering protein language models | arXiv: 2509.07983
stochastic encodings for active feature acquisition | arXiv: 2508.01957
stofm a multi-scale foundation model for spatial transcriptomics | arXiv: 2507.11588
strategic fusion optimizes transformer compression | arXiv: 2501.03273
streamline without sacrifice -- squeeze out computation redundancy in lmm | arXiv: 2505.15816
subspace optimization for large language models with convergence guarantees | arXiv: 2410.11289
suica learning super-high dimensional sparse implicit neural representations for | arXiv: 2412.01124
suitability filter a statistical framework for classifier evaluation in real-wor | arXiv: 2505.22356
sum-of-parts self-attributing neural networks with end-to-end learning of featur | arXiv: 2310.16316
supercharging graph transformers with advective diffusion | arXiv: 2310.06417
supernova event dataset interpreting large language models personality through c | arXiv: 2506.12189
symmetry-aware gflownets | arXiv: 2506.02685
symmetry-robust 3d orientation estimation | arXiv: 2410.02101
syndacate a synthetic dataset for evaluating part-whole hierarchical inference | arXiv: 2506.17558
synonymous variational inference for perceptual image compression | arXiv: 2505.22438
synthesizing images on perceptual boundaries of anns for uncovering and manipula | arXiv: 2505.03641
synthetic face datasets generation via latent space exploration from brownian id | arXiv: 2405.00228
synthetic perception can generated images unlock latent visual prior for text-ce | arXiv: 2506.17623
system-aware unlearning algorithms use lesser forget faster | arXiv: 2506.06073
t1 advancing language model reasoning through reinforcement learning and inferen | arXiv: 2501.11651
tabflex scaling tabular learning to millions with linear attention | arXiv: 2506.05584
tackling view-dependent semantics in 3d language gaussian splatting | arXiv: 2505.24746
tamas benchmarking adversarial risks in multi-agent llm systems | arXiv: 2511.05269
taming diffusion for dataset distillation with high representativeness | arXiv: 2505.18399
taming knowledge conflicts in language models | arXiv: 2503.10996
taming rectified flow for inversion and editing | arXiv: 2411.04746
tango clustering with typicality-aware nonlocal mode-seeking and graph-cut optim | arXiv: 2408.10084
targeted unlearning with single layer unlearning gradient | arXiv: 2407.11867
task-agnostic pre-training and task-guided fine-tuning for versatile diffusion p | arXiv: 2409.19949
tcp-diffusion a multi-modal diffusion model for global tropical cyclone precipit | arXiv: 2410.13175
teaching llms to speak spectroscopy | arXiv: 2508.10075
teaching physical awareness to llms through sounds | arXiv: 2506.08524
temporal query network for efficient multivariate time series forecasting | arXiv: 2505.12917
test-time adaptation with binary feedback | arXiv: 2505.18514
test-time canonicalization by foundation models for robust perception | arXiv: 2507.10375
test-time training provably improves transformers as in-context learners | arXiv: 2503.11842
text-to-lora instant transformer adaption | arXiv: 2506.06105
tgdpo harnessing token-level reward guidance for enhancing direct preference opt | arXiv: 2506.14574
the best of both worlds bridging quality and diversity in data selection with bi | arXiv: 2410.12458
the brains bitter lesson scaling speech decoding with self-supervised learning | arXiv: 2406.04328
the butterfly effect neural network training trajectories are highly sensitive t | arXiv: 2506.13234
the canarys echo auditing privacy risks of llm-generated synthetic text | arXiv: 2502.14921
the challenge of teaching reasoning to llms without rl or distillation | arXiv: 2507.09850
the courage to stop overcoming sunk cost fallacy in deep reinforcement learning | arXiv: 2506.13672
the dark side of the forces assessing non-conservative force models for atomisti | arXiv: 2412.11569
the devil is in the details tackling unimodal spurious correlations for generali | arXiv: 2503.03122
the diffusion duality | arXiv: 2506.10892
the disparate benefits of deep ensembles | arXiv: 2410.13831
the double-ellipsoid geometry of clip | arXiv: 2411.14517
the four color theorem for cell instance segmentation | arXiv: 2506.09724
the impact of on-policy parallelized data collection on deep reinforcement learn | arXiv: 2506.03404
the lock-in hypothesis stagnation by algorithm | arXiv: 2506.06166
the panaceas for improving low-rank decomposition in communication-efficient fed | arXiv: 2505.23176
the price of freedom exploring expressivity and runtime tradeoffs in equivariant | arXiv: 2506.13523
the right to ai | arXiv: 2501.17899
the ripple effect on unforeseen complications of backdoor attacks | arXiv: 2505.11586
the sample complexity of online strategic decision making with information asymm | arXiv: 2506.09940
the sharpness disparity principle in transformers for accelerating language mode | arXiv: 2502.19002
theoretical guarantees on the best-of-n alignment policy | arXiv: 2401.01879
theoretical limitations of ensembles in the age of overparameterization | arXiv: 2410.16201
theoretical performance guarantees for partial domain adaptation via partial opt | arXiv: 2506.02712
theoretically unmasking inference attacks against ldp-protected clients in feder | arXiv: 2506.17292
thickness-aware e3-equivariant 3d mesh neural networks | arXiv: 2505.21572
tilted sharpness-aware minimization | arXiv: 2410.22656
time-aware world model for adaptive prediction and control | arXiv: 2506.08441
timedart a diffusion autoregressive transformer for self-supervised time series | arXiv: 2410.05711
timepoint accelerated time series alignment via self-supervised keypoint and des | arXiv: 2505.23475
timepro efficient multivariate long-term time series forecasting with variable- | arXiv: 2505.20774
timing temporality-aware integrated gradients for time series explanation | arXiv: 2506.05035
tined gnns-to-mlps by teacher injection and dirichlet energy distillation | arXiv: 2412.11180
to each metric its decoding post-hoc optimal decision rules of probabilistic hie | arXiv: 2506.01552
to steer or not to steer mechanistic error reduction with abstention for languag | arXiv: 2510.13290
tokenized bandit for llm decoding and alignment | arXiv: 2506.07276
toma token merge with attention for diffusion models | arXiv: 2509.10918
toping topologically interpretable graph learning via persistent rationale filtr | arXiv: 2510.05102
toward data-centric directed graph learning an entropy-driven approach | arXiv: 2505.00983
toward robust hyper-detailed image captioning a multiagent approach and dual eva | arXiv: 2412.15484
toward safe and human-aligned game conversational recommendation via multi-agent | arXiv: 2504.20094
towards a mechanistic explanation of diffusion model generalization | arXiv: 2411.19339
towards an optimal control perspective of resnet training | arXiv: 2506.21453
towards attributions of input variables in a coalition | arXiv: 2309.13411
towards benchmarking foundation models for tabular data with text | arXiv: 2507.07829
towards efficient online tuning of vlm agents via counterfactual soft reinforcem | arXiv: 2505.03792
towards flexible perception with visual memory | arXiv: 2408.08172
towards graph foundation models learning generalities across graphs via task-tre | arXiv: 2412.16441
towards llm agents for earth observation | arXiv: 2504.12110
towards long-horizon interpretability efficient and faithful multi-token attribu | arXiv: 2602.01914
towards practical defect-focused automated code review | arXiv: 2505.17928
towards rationale-answer alignment of lvlms via self-rationale calibration | arXiv: 2509.13919
towards robust influence functions with flat validation minima | arXiv: 2505.19097
towards trustworthy federated learning with untrusted participants | arXiv: 2505.01874
towards universal offline black-box optimization via learning language model emb | arXiv: 2506.07109
training a generally curious agent | arXiv: 2502.17543
training dynamics of in-context learning in linear attention | arXiv: 2501.16265
training flexible models of genetic variant effects from functional annotations | arXiv: 2506.19598
training software engineering agents and verifiers with swe-gym | arXiv: 2412.21139
transformative or conservative conservation laws for resnets and transformers | arXiv: 2506.06194
transformer-based spatial-temporal counterfactual outcomes estimation | arXiv: 2506.21154
transpl vq-code transition matrices for pseudo-labeling of time series unsupervi | arXiv: 2505.09955
tree-sliced wasserstein distance a geometric perspective | arXiv: 2406.13725
tree-sliced wasserstein distance with nonlinear projection | arXiv: 2505.00968
treelora efficient continual learning via layer-wise loras guided by a hierarchi | arXiv: 2506.10355
truly self-improving agents require intrinsic metacognitive learning | arXiv: 2506.05109
tuco measuring the contribution of fine-tuning to individual responses of llms | arXiv: 2506.23423
ui-evol automatic knowledge evolving for computer use agents | arXiv: 2505.21964
ui-vision a desktop-centric gui benchmark for visual perception and interaction | arXiv: 2503.15661
unable to forget proactive interference reveals working memory limits in llms be | arXiv: 2506.08184
understanding and mitigating memorization in diffusion models for tabular data | arXiv: 2412.11044
understanding and mitigating memorization in generative models via sharpness of | arXiv: 2412.04140
understanding and mitigating miscalibration in prompt tuning for vision-language | arXiv: 2410.02681
understanding mode connectivity via parameter space symmetry | arXiv: 2505.23681
understanding model ensemble in transferable adversarial attack | arXiv: 2410.06851
understanding sharpness dynamics in nn training with a minimalist example the ef | arXiv: 2506.06940
understanding synthetic context extension via retrieval heads | arXiv: 2410.22316
understanding the emergence of multimodal representation alignment | arXiv: 2502.16282
understanding the limits of deep tabular methods with temporal shift | arXiv: 2502.20260
understanding the statistical accuracy-communication trade-off in personalized f | arXiv: 2410.08934
unhippo uncertainty-aware initialization for state space models | arXiv: 2506.05065
unifews you need fewer operations for efficient graph neural networks | arXiv: 2403.13268
unifying specialized visual encoders for video language models | arXiv: 2501.01426
unimomo unified generative modeling of 3d molecules for de novo binder design | arXiv: 2503.19300
unisim a unified simulator for time-coarsened dynamics of biomolecules | arXiv: 2506.03157
universal neural optimal transport | arXiv: 2212.00133
universal retrieval for multimodal trajectory modeling | arXiv: 2506.22056
unlocking post-hoc dataset inference with synthetic data | arXiv: 2506.15271
unlocking the capabilities of large vision-language models for generalizable and | arXiv: 2503.14853
unlocking the power of rehearsal in continual learning a theoretical perspective | arXiv: 2506.00205
unmore unsupervised multi-object segmentation via center-boundary reasoning | arXiv: 2506.01778
unsupervised learning for class distribution mismatch | arXiv: 2505.06948
update your transformer to the latest release re-basin of task vectors | arXiv: 2505.22697
using multiple input modalities can improve data-efficiency and ood generalizati | arXiv: 2507.13385
validating mechanistic interpretations an axiomatic approach | arXiv: 2407.13594
video prediction policy a generalist robot policy with predictive visual represe | arXiv: 2412.14803
vineppo refining credit assignment in rl training of llms | arXiv: 2410.01679
vision graph prompting via semantic low-rank decomposition | arXiv: 2505.04121
vision-language model selection and reuse for downstream adaptation | arXiv: 2501.18271
vision-language models create cross-modal task representations | arXiv: 2410.22330
visionts visual masked autoencoders are free-lunch zero-shot time series forecas | arXiv: 2408.17253
visual generation without guidance | arXiv: 2501.15420
visual language models as zero-shot deepfake detectors | arXiv: 2507.22469
vocabtrim vocabulary pruning for efficient speculative decoding in llms | arXiv: 2506.22694
vtgaussian-slam rgbd slam for large scale scenes with splatting view-tied 3d gau | arXiv: 2506.02741
vulnerability-aware alignment mitigating uneven forgetting in harmful fine-tunin | arXiv: 2506.03850
wasserstein policy optimization | arXiv: 2505.00663
watch out your album on the inadvertent privacy memorization in multi-modal larg | arXiv: 2503.01208
wave weighted autoregressive varying gate for time series forecasting | arXiv: 2410.03159
weak-to-strong jailbreaking on large language models | arXiv: 2401.17256
weisfeiler and leman go gambling why expressive lottery tickets win | arXiv: 2506.03919
wgformer an se3-transformer driven by wasserstein gradient flows for molecular g | arXiv: 2410.09795
what has a foundation model found using inductive bias to probe for world models | arXiv: 2507.06952
what limits virtual agent application omnibench a scalable multi-dimensional ben | arXiv: 2506.08933
what makes an ensemble un interpretable | arXiv: 2506.08216
when bad data leads to good models | arXiv: 2505.04741
when can in-context learning generalize out of task distribution | arXiv: 2506.05574
when data-free knowledge distillation meets non-transferable teacher escaping ou | arXiv: 2507.04119
when diffusion models memorize inductive biases in probability flow of minimum-n | arXiv: 2506.19031
when every millisecond counts real-time anomaly detection via the multimodal asy | arXiv: 2506.17457
when model knowledge meets diffusion model diffusion-assisted data-free image sy | arXiv: 2506.15381
when will it fail anomaly to prompt for forecasting future anomalies in time ser | arXiv: 2506.23596
whitened clip as a likelihood surrogate of images and captions | arXiv: 2505.06934
why is spatial reasoning hard for vlms an attention mechanism perspective on foc | arXiv: 2503.01773
widening the network mitigates the impact of data heterogeneity on fedavg | arXiv: 2508.12576
wikibigedit understanding the limits of lifelong knowledge editing in llms | arXiv: 2503.05683
wildchat-50m a deep dive into the role of synthetic data in post-training | arXiv: 2501.18511
wilting trees interpreting the distance between mpnn embeddings | arXiv: 2505.24642
winner-takes-all for multivariate probabilistic time series forecasting | arXiv: 2506.05515
x hacking the threat of misguided automl | arXiv: 2401.08513
x-transfer attacks towards super transferable adversarial attacks on clip | arXiv: 2505.05528
xchemagents agentic ai for explainable quantum chemistry | arXiv: 2505.20574
zero-shot adaptation of parameter-efficient fine-tuning in diffusion models | arXiv: 2506.04244
zero-shot generalization of vision-based rl without data augmentation | arXiv: 2410.07441
connecting thompson sampling and ucb towards more efficient best-fixed action | arXiv: 2505.02383
on differential privacy for adaptively solving search problems | arXiv: 2506.05503
retraining with predicted hard labels provably increases model accurac | arXiv: 2406.11206
theoretically unmasking inference attacks against ldp-protected client data in | arXiv: 2506.17292
adaptivestep automatically dividing reasoning step through model confidence | arXiv: 2502.13943
dynamic benchmarking of reasoning capabilities in code large language models und | arXiv: 2503.04149
efficoder enhancing code generation in large language models through efficiency- | arXiv: 2410.10209
epicoder encompassing diversity and complexity in code generation | arXiv: 2501.04694
function-to-style guidance of llms for code translation | arXiv: 2507.11083
mind the gap a practical attack on gguf quantization | arXiv: 2505.23786
reasoning through execution unifying process and outcome rewards for code genera | arXiv: 2412.15118
robust learning of diverse code edits | arXiv: 2503.03656
sparselora accelerating llm fine-tuning with contextual sparsity | arXiv: 2506.16500
towards practical defect-focused automated code review | arXiv: 2505.17928
training software engineering agents and verifiers with swe-gym | arXiv: 2412.21139
agent warpp workflow adherence via runtime parallel personalization | arXiv: 2507.19543
investigating non-transitivity in llm-as-a-judge | arXiv: 2502.14074
position uncertainty quantification needs reassessment for large-language model | arXiv: 2505.22655
action-minimization meets generative modeling efficient transition pat | arXiv: 2504.18506
impact iterative mask-based parallel decoding for text-to-audio generation | arXiv: 2506.00736
when model knowledge meets diffusion model diffusion-assisted data-free image synthesis | arXiv: 2506.15381
epsilon-vae denoising as visual decoding | arXiv: 2410.04081
dont lag rag training-free adversarial detection using rag | arXiv: 2504.04858
poqd performance-oriented query decomposer for multi-vector retrieval | arXiv: 2505.19189
rapid long-context inference with retrieval-augmented speculative decoding | arXiv: 2502.20330
rethinking addressing in language models via contexualized equivariant positiona | arXiv: 2501.00712
understanding synthetic context extension via retrieval heads | arXiv: 2410.22316
a cross modal knowledge distillation data augmentation recipe for improving tran | arXiv: 2505.21317
a reasoning-based approach to cryptic crossword clue solving | arXiv: 2506.04824
ab initio nonparametric variable selection for scalable symbolic regression with | arXiv: 2410.13681
avoiding leakage poisoning concept interventions under distribution shifts | arXiv: 2504.17921
concept-based unsupervised domain adaptation | arXiv: 2505.05195
configurable preference tuning with rubric-guided synthetic data | arXiv: 2506.11702
conformal prediction as bayesian quadrature | arXiv: 2502.13228
do sparse autoencoders generalize a case study of answerability | arXiv: 2502.19964
evaluating neuron explanations a unified framework with sanity checks | arXiv: 2506.05774
evolving prompts in-context an open-ended self-replicating perspective | arXiv: 2506.17930
explaining fast and slow abstraction and refinement of provable explanations | arXiv: 2506.08505
foundation molecular grammar multi-modal foundation models induce interpretable | arXiv: 2505.22948
inference-time decomposition of activations itda a scalable approach to interpre | arXiv: 2505.17769
leveraging predictive equivalence in decision trees | arXiv: 2506.14143
mib a mechanistic interpretability benchmark | arXiv: 2504.13151
modeling user behavior from adaptive surveys with supplemental context | arXiv: 2507.20919
near optimal decision trees in a split second | arXiv: 2502.15988
on the effect of uncertainty on layer-wise inference dynamics | arXiv: 2507.06722
on the power of context-enhanced learning in llms | arXiv: 2503.01821
position we need an algorithmic understanding of generative ai | arXiv: 2507.07544
rethinking explainable machine learning as applied statistics | arXiv: 2402.02870
safetyanalyst interpretable transparent and steerable safety moderation for ai b | arXiv: 2410.16665
slim one-shot quantization and sparsity with low-rank approximation for llm weig | arXiv: 2410.09615
supernova event dataset interpreting large language models personality through c | arXiv: 2506.12189
to steer or not to steer mechanistic error reduction with abstention for languag | arXiv: 2510.13290
towards attributions of input variables in a coalition | arXiv: 2309.13411
towards flexible perception with visual memory | arXiv: 2408.08172
what makes an ensemble un interpretable | arXiv: 2506.08216
why is spatial reasoning hard for vlms an attention mechanism perspective on foc | arXiv: 2503.01773
representation shattering in transformers a synthetic study with knowledge editi | arXiv: 2410.17194
wikibigedit understanding the limits of lifelong knowledge editing in llms | arXiv: 2503.05683
theorem-of-thought a multi-agent framework for abductive deductive and inductive | arXiv: 2506.07106
are llm belief updates consistent with bayes theorem | arXiv: 2507.17951
communicating activations between language model agents | arXiv: 2501.14082
cooperation of experts fusing heterogeneous information with large margin | arXiv: 2505.20853
correlated errors in large language models | arXiv: 2506.07962
cross-regularization adaptive model complexity through validation gradients | arXiv: 2506.19755
dilqr differentiable iterative linear quadratic regulator via implicit different | arXiv: 2506.17473
disentangling and integrating relational and sensory information in transformer | arXiv: 2405.16727
enigma interactive tools substantially assist lm agents in finding security vuln | arXiv: 2409.16165
evaluating llms across multi-cognitive levels from medical knowledge mastery to | arXiv: 2506.08349
faster and stronger when ann-snn conversion meets parallel spiking calculation | arXiv: 2412.13610
fedtail federated long-tailed domain generalization with sharpness-guided gradie | arXiv: 2506.08518
feedforward few-shot species range estimation | arXiv: 2502.14977
fully heteroscedastic count regression with deep double poisson networks | arXiv: 2406.09262
function encoders a principled approach to transfer learning in hilbert spaces | arXiv: 2501.18373
g-sim generative simulations with large language models and gradient-free calibr | arXiv: 2506.09272
gradient aligned regression via pairwise losses | arXiv: 2402.06104
how much can we forget about data contamination | arXiv: 2410.03249
hyperband-based bayesian optimization for black-box prompt selection | arXiv: 2412.07820
improved and oracle-efficient online ell 1-multicalibration | arXiv: 2505.17365
improving generalization with flat hilbert bayesian inference | arXiv: 2410.04196
improving the effective receptive field of message-passing neural networks | arXiv: 2505.23185
latent imputation before prediction a new computational paradigm for de novo pep | arXiv: 2505.17524
learning distribution-wise control in representation space for language models | arXiv: 2506.06686
learning safe strategies for value maximizing buyers in uniform price auctions | arXiv: 2406.03674
leveraging online olympiad-level math problems for llms training and contaminati | arXiv: 2501.14275
llm-srbench a new benchmark for scientific equation discovery with large languag | arXiv: 2504.10415
meek models shall inherit the earth | arXiv: 2507.07931
on temperature scaling and conformal prediction of deep classifiers | arXiv: 2402.05806
phantomwiki on-demand datasets for reasoning and retrieval evaluation | arXiv: 2502.20377
position ai evaluation should learn from how we test humans | arXiv: 2306.10512
promoting ensemble diversity with interactive bayesian distributional robustness | arXiv: 2506.07247
provably cost-sensitive adversarial defense via randomized smoothing | arXiv: 2310.08732
random registers for cross-domain few-shot learning | arXiv: 2506.02843
researchtown simulator of human research community | arXiv: 2412.17767
runtime analysis of evolutionary nas for multiclass classification | arXiv: 2506.06019
set valued predictions for robust domain generalization | arXiv: 2507.03146
the best of both worlds bridging quality and diversity in data selection with bi | arXiv: 2410.12458
ui-evol automatic knowledge evolving for computer use agents | arXiv: 2505.21964
unlocking post-hoc dataset inference with synthetic data | arXiv: 2506.15271
expert evaluation of llm world models a high-t c superconductivity case study | arXiv: 2511.03782
la rosa enhancing llm efficiency via layerwise rotated sparse activation | arXiv: 2507.01299
regress dont guess -- a regression-like loss on number tokens for language model | arXiv: 2411.02083
a square peg in a square hole meta-expert for long-tailed semi-supervised learni | arXiv: 2505.16341
algebra unveils deep learning -- an invitation to neuroalgebraic geometry | arXiv: 2501.18915
bayesian neural scaling law extrapolation with prior-data fitted networks | arXiv: 2505.23032
benign overfitting in token selection of attention mechanism | arXiv: 2409.17625
chameleon a flexible data-mixing framework for language model pretraining and fi | arXiv: 2505.24844
counting in small transformers the delicate interplay between attention and feed | arXiv: 2407.11542
density ratio estimation-based bayesian optimization with semi-supervised learni | arXiv: 2305.15612
dipllm fine-tuning llm for strategic decision-making in diplomacy | arXiv: 2506.09655
does data scaling lead to visual compositional generalization | arXiv: 2507.07102
evaluating morphological alignment of tokenizers in 70 languages | arXiv: 2507.06378
how to synthesize text data without model collapse | arXiv: 2412.14689
in-context adaptation to concept drift for learned database operations | arXiv: 2505.04404
inductive gradient adjustment for spectral bias in implicit neural representatio | arXiv: 2410.13271
language model developers should report train-test overlap | arXiv: 2410.08385
language models over canonical byte-pair encodings | arXiv: 2506.07956
large language models are demonstration pre-selectors for themselves | arXiv: 2506.06033
llm data selection and utilization via dynamic bi-level optimization | arXiv: 2507.16178
metadata conditioning accelerates language model pre-training | arXiv: 2501.01956
on the clean generalization and robust overfitting in adversarial training from | arXiv: 2306.01271
on the role of label noise in the feature learning process | arXiv: 2505.18909
position the future of bayesian prediction is prior-fitted | arXiv: 2505.23947
revisiting continuity of image tokens for cross-domain few-shot learning | arXiv: 2506.03110
the dark side of the forces assessing non-conservative force models for atomisti | arXiv: 2412.11569
the double-ellipsoid geometry of clip | arXiv: 2411.14517
the sharpness disparity principle in transformers for accelerating language mode | arXiv: 2502.19002
tokenized bandit for llm decoding and alignment | arXiv: 2506.07276
towards robust influence functions with flat validation minima | arXiv: 2505.19097
when can in-context learning generalize out of task distribution | arXiv: 2506.05574
whitened clip as a likelihood surrogate of images and captions | arXiv: 2505.06934
one missing piece for open-source reasoning models a dataset to mitigate cold-st | arXiv: 2506.02338
pcot persuasion-augmented chain of thought for detecting fake news and social me | arXiv: 2506.06842
pcot persuasion disinfo | arXiv: 2506.06842
putnam-axiom a functional and static benchmark for measuring higher level mathem | arXiv: 2508.08292
quire better cot | arXiv: 2405.18915
towards better chain-of-thought a reflection on effectiveness and faithfulness | arXiv: 2405.18915
became bayesian continual learning with adaptive model merging | arXiv: 2504.02666
cut out and replay a simple yet versatile strategy for multi-label online contin | arXiv: 2505.19680
improving continual learning performance and efficiency with auxiliary classifie | arXiv: 2403.07404
negmerge sign-consensual weight merging for machine unlearning | arXiv: 2410.05583
system-aware unlearning algorithms use lesser forget faster | arXiv: 2506.06073
unlocking the power of rehearsal in continual learning a theoretical perspective | arXiv: 2506.00205
training flexible models of genetic variant effects from functional annotations | arXiv: 2506.19598
fgfp a fractional gaussian filter and pruning for deep neural networks compression | arXiv: 2507.22527
m3-jepa multimodal alignment via multi-gate moe based on the joint-embedding pre | arXiv: 2409.05929
overcoming multi-step complexity in multimodal theory-of-mind reasoning | arXiv: 2506.01301
ui-evol automatic knowledge evolving for computer use agents | arXiv: 2505.21964
recommendations with sparse comparison data provably fast convergence for noncon | arXiv: 2502.20033
towards practical defect-focused automated code review | arXiv: 2505.17928
gradual transition from bellman optimality operator to bellman operator in | arXiv: 2506.05968
closed-form solutions a new perspective on solving differential equations | arXiv: 2405.14620
actionpiece contextually tokenizing action sequences generative recommendation | arXiv: 2502.13581
large language model llm-enabled in-context learning for wireless network optimi | arXiv: 2408.00214
defame dynamic evidence-based fact-checking with multimodal experts | arXiv: 2412.10510
dynamical phases of short-term memory mechanisms in rnns | arXiv: 2502.17433
is your llm-based multi-agent a reliable real-world planner exploring fraud dete | arXiv: 2505.16557
learning survival distributions with the asymmetric laplace distribution | arXiv: 2505.03712
or-bench an over-refusal benchmark for large language models | arXiv: 2405.20947
raising the bar investigating the values of large language models via generative | arXiv: 2406.14230
when bad data leads to good models | arXiv: 2505.04741
when will it fail anomaly to prompt for forecasting future anomalies in time seri | arXiv: 2506.23596
asymrnr video diffusion transformers acceleration with asymmetric reduction and | arXiv: 2412.11706
ca2-vdm efficient autoregressive video diffusion model with causal generation an | arXiv: 2411.16375
data-juicer sandbox a feedback-driven suite for multimodal data-model co-develop | arXiv: 2407.11784
diffusion adversarial post-training for one-step video generation | arXiv: 2501.08316
how far is video generation from world model a physical law perspective | arXiv: 2411.02385
mimicmotion high-quality human motion video generation with confidence-aware pos | arXiv: 2406.19680
riflex a free lunch for length extrapolation in video diffusion transformers | arXiv: 2502.15894