ProPose PrOntoQA, a logic-based synthetic dataset, to systematically analyze LLM reasoning, revealing strengths in single-step deduction but weaknesses in proof planning.

Abulhair Saparov, He He

Proves Lipschitz continuity of the Schrödinger map in entropic OT with C^{k+1} cost, covering multi-marginal cases.

Guillaume Carlier, Lénaïc Chizat, Maxime Laborde

DecAF jointly decodes logical forms and direct answers using text retrieval, achieving SOTA on WebQSP, FreebaseQA, GrailQA with 79-82% Hits@1.

Donghan Yu, Sheng Zhang, Patrick Ng et al.

This study compares FNO, ResNet, and U-Net architectures for multi-scale PDE modeling, demonstrating a single surrogate's strong generalization across parameters.

Jayesh K. Gupta, Johannes Brandstetter

Proposes InterFlow, a stochastic interpolant-based continuous-time normalizing flow, optimizing path length and surpassing traditional methods in high-resolution image generation.

Michael S. Albergo, Eric Vanden-Eijnden

ASPiRe accelerates reinforcement learning using adaptive skill priors, significantly improving learning efficiency.

Mengda Xu, Manuela Veloso, Shuran Song

Proposes 'Context Distillation' to internalize reasoning, instructions, and examples, boosting model performance by 9-30% on key benchmarks.

Charlie Snell, Dan Klein, Ruiqi Zhong

LiMBeR method linearly maps image features to text prompts, enhancing visual question answering performance.

Jack Merullo, Louis Castricato, Carsten Eickhoff et al.

Proposes a causal-model-based framework for quantifying harm, integrating decision-theoretic probability weighting, applicable at individual and societal levels.

Sander Beckers, Hana Chockler, Joseph Y. Halpern

Spotlight model uses vision-language integration for mobile UI understanding, surpassing existing methods.

Gang Li, Yang Li

This paper proves that two-layer neural networks trained with SGD converge their first-layer weights to the k-dimensional principal subspace spanned by the target model's index vectors, enabling low-dimensional feature learning.

Alireza Mousavi-Hosseini, Sejun Park, Manuela Girotti et al.

DINOSAUR leverages self-supervised feature reconstruction with Slot Attention, outperforming existing models, scalable to COCO and PASCAL VOC datasets.

Maximilian Seitzer, Max Horn, Andrii Zadaianchuk et al.

Spikformer combines Spiking Neural Network with Transformer, achieving 74.81% accuracy on ImageNet with 4 time steps.

Zhaokun Zhou, Yuesheng Zhu, Chao He et al.

Make-A-Video generates videos from text using text-image data and unsupervised video learning.

Uriel Singer, Adam Polyak, Thomas Hayes et al.

DiGress employs discrete diffusion for graph generation, achieving up to 99% validity on molecular datasets and scaling to 1.3 million molecules.

Clement Vignac, Igor Krawczuk, Antoine Siraudin et al.

Optimized global average treatment effect (GATE) estimation using IPW and OLS methods, achieving 22% higher consumption effect estimates.

Stefan Faridani, Paul Niehaus

Introduced STCAT, a one-stage framework achieving significant performance gains on VidSTG and HC-STVG datasets.

Yang Jin, Yongzhi Li, Zehuan Yuan et al.

Proposed a data-adaptive kernel Bayesian method within computational graph framework for one-shot learning of SDEs, improving likelihood by 15% on benchmark models.

Matthieu Darcy, Boumediene Hamzi, Giulia Livieri et al.

Proposes induction heads as the primary mechanism for in-context learning in large transformers, supported by six lines of evidence including training phase shifts and causal ablations.

Catherine Olsson, Nelson Elhage, Neel Nanda et al.

Introduced a Hybrid Physics-Informed Neural Network model for predicting char formation and burning degree in polymer pyrolysis and ablation.

Aref Ghaderi, Ramin Akbari, Yang Chen et al.