Using large language models as user agents to evaluate task-oriented dialogue systems, enhancing diversity and task completion rates.

Taaha Kazi, Ruiliang Lyu, Sizhe Zhou et al.

EgoVid-5M provides 5M curated clips, while EgoDreamer jointly controls egocentric generation with text and kinematics.

Xiaofeng Wang, Kang Zhao, Feng Liu et al.

GaussianAnything employs point cloud structured latent space with diffusion models for multi-modal 3D generation, enabling high-quality, editable outputs.

Yushi Lan, Shangchen Zhou, Zhaoyang Lyu et al.

Spider 2.0 framework evaluates language models on enterprise text-to-SQL workflows, solving only 21.3% of tasks.

Fangyu Lei, Jixuan Chen, Yuxiao Ye et al.

Contextualized evaluations synthesize context to assess language model responses to underspecified queries, significantly altering evaluation conclusions.

Chaitanya Malaviya, Joseph Chee Chang, Dan Roth et al.

Introduces continual memorization framework; REMIX data mixing significantly reduces language model forgetting.

Howard Chen, Jiayi Geng, Adithya Bhaskar et al.

RFG replaces backpropagation with randomized weight perturbations, reaching 0.0347 Poisson error and an estimated 66% lower training cost.

Ruyin Wan, Qian Zhang, George Em Karniadakis

DNCIT combines learned image embeddings with nonparametric CITs and validates MRI–behavior null findings in UK Biobank data.

Marco Simnacher, Xiangnan Xu, Hani Park et al.

Proposes Image2Text2Image framework using diffusion models for label-free image caption evaluation, achieving high correlation with human judgment.

Jia-Hong Huang, Hongyi Zhu, Yixian Shen et al.

Multi-modal self-supervised learning enhances CVD prediction, improving balanced accuracy by 7.6% on limited data using ECG, CMR, and clinical info.

Francesco Girlanda, Olga Demler, Bjoern Menze et al.

Introduces Run-Length Tokenization (RLT), accelerating video Transformer training by 30% with only a 0.1% accuracy drop.

Rohan Choudhury, Guanglei Zhu, Sihan Liu et al.

This study evaluates 17 leading LLMs on their ability to follow multiple information threads through near-million-token contexts, revealing performance degradation but strong thread safety.

Jonathan Roberts, Kai Han, Samuel Albanie

SG-I2V enables zero-shot trajectory control in image-to-video generation using pre-trained diffusion models, achieving high-quality, controllable videos without fine-tuning.

Koichi Namekata, Sherwin Bahmani, Ziyi Wu et al.

DINO-WM leverages pre-trained visual patch features for zero-shot planning, outperforming state-of-the-art models with 56% LPIPS improvement and 45% success rate increase.

Gaoyue Zhou, Hengkai Pan, Yann LeCun et al.

FrontierMath benchmark tests advanced mathematical reasoning; AI solves less than 2%, exposing a large gap with human experts.

Elliot Glazer, Ege Erdil, Tamay Besiroglu et al.

Proposes SCaLR, a self-calibrated listwise reranking framework using explicit relevance scores, improving efficiency and global comparability in large candidate sets.

Ruiyang Ren, Yuhao Wang, Kun Zhou et al.

SimpleQA benchmark evaluates language models on short factual questions, challenging and easy to grade.

Jason Wei, Nguyen Karina, Hyung Won Chung et al.

Long-context RAG enhances LLM performance, but only a few models maintain accuracy above 64k tokens.

Quinn Leng, Jacob Portes, Sam Havens et al.

Robots using large language models can enhance healthcare efficiency, addressing aging and workforce shortages.

Souren Pashangpour, Goldie Nejat

This report analyzes rapid AI capability growth via scaling and interpretability limits, emphasizing resource constraints and safety challenges.

Yoshua Bengio, Sören Mindermann, Daniel Privitera et al.