ReTVL leverages sparse retry event supervision to capture local mistake-recovery dynamics, outperforming progress-based models in fine-grained value estimation.

Xinyao Qin, Junjie Lu, Kaixin Wang et al.

Developed sample complexity bounds for unbalanced entropic OT, demonstrating geometric stability and high-probability guarantees.

Francisco Andrade, Gabriel Peyré, Clarice Poon

MemClaw addresses four failure modes in multi-agent LLM systems through governed shared memory, ensuring efficient knowledge management.

Yanki Margalit, Nurit Cohen-Inger, Erni Avram et al.

NatureBench evaluates AI coding agents on 90 scientific tasks; strongest model surpasses SOTA in only 17.8%.

Yuru Wang, Lejun Cheng, Yuxin Zuo et al.

PETRA leverages energy-domain classification and synthetic supervision to improve petroleum engineering retrieval, boosting first-stage nDCG from 0.703 to 0.763.

Kirill Dubovikov, Omar El Mansouri, Hachem Madmoun et al.

TuringViT introduces Turing Linear Attention, VISTA-Curation, and dynamic-resolution pretraining, achieving high performance with only 10% of the data and improved efficiency.

Qiman Wu, Hanlin Chen, Lyujie Chen et al.

FlowR2A learns reward-conditioned action distributions via flow matching, unifying dense reward supervision with proposal generation, achieving state-of-the-art in multimodal driving planning.

Xirui Li, Zhe Liu, Xiaoqing Ye et al.

Proposes optimization-guided diffusion model to enhance physical feasibility in robot control, improving grasp task success by 20%.

Sabrina Bodmer, René Zurbrügg, Tifanny Portela et al.

Introduced TR-CIE sampler to improve sampling quality in discrete flow matching under limited function evaluations.

Feiyang Fu, Hehe Fan

Introduces blockwise policy-drift gating, improving OPD pass rates in math reasoning tasks.

Liwen Zheng, Haiyun Jiang

ChartWalker constructs hierarchical knowledge graphs and uses structure-aware sampling to improve cross-chart multi-hop reasoning, boosting multimodal QA.

Ning Tang, Chenghan Xie, Hanyang Yuan et al.

RIFT-Bench provides a dynamic red-teaming method for unified evaluation of diverse agentic AI systems.

Yarin Yerushalmi Levi, Roy Betser, Amit Giloni et al.

Proposes a semantic browsing framework for controllable diversity in image generation, leveraging rich textual representations to enable structured, interpretable exploration.

Sara Dorfman, Maya Vishnevsky, Omer Dahary et al.

Proposes Success Visitation Matching (SVM) rewards, transforming sparse rewards into dense signals, doubling RL finetuning speed in robotic tasks.

Raymond Tsao, Andrew Wagenmaker, Sergey Levine

See2Act employs diffusion models for active perception, coupling viewpoint inference with action prediction, achieving 95% success in occluded manipulation tasks.

Kuancheng Wang, Vaibhav Saxena, Shuo Cheng et al.

Proposes an iterative self-filtering method using CLIP to improve vision-language data quality without pretraining, boosting downstream performance.

Andrei Liviu Nicolicioiu, Sarvjeet Singh Ghotra, Morgane M. Moss et al.

LangMAP extends UnigramLM with language-specific MAP estimation, enabling multi-language tokenization from a shared vocabulary, improving morphological boundary alignment.

Clara Meister, Suchir Salhan, Andrzej Szablewski et al.

HoloAgent-0 integrates 3D spatial memory with multi-layered skill scheduling, enabling long-horizon autonomous navigation and manipulation.

Xiaolin Zhou, Liu Liu, Tingyang Xiao et al.

Structure-aware variance reduction for unbiased randomized Hamiltonian simulation reduces sampling costs by up to 96%.

Joshua W. Dai, Fredrik Hasselgren, Chusei Kiumi

VeriEvol employs iterative prompt evolution and offline answer verification to scale visual math reasoning data from 10K to 250K, boosting accuracy by 19.31%.

Haoling Li, Kai Zheng, Jie Wu et al.