AutoMLGen integrates knowledge base and MCGS to optimize end-to-end ML pipelines, achieving 36.4% medal rate within 12 hours, outperforming baselines.

Shangheng Du, Xiangchao Yan, Dengyang Jiang et al.

VRPAgent uses LLMs to generate heuristic operators, optimized via genetic algorithms, outperforming handcrafted methods on multiple VRP variants.

André Hottung, Federico Berto, Chuanbo Hua et al.

This paper introduces AI-Driven Research for Systems (ADRS), using OpenEvolve to automate algorithm optimization, achieving up to 5.0x speedup and 50% cost reduction.

Audrey Cheng, Shu Liu, Melissa Pan et al.

Proposes DeSA, a two-stage training framework that decouples search and answer generation, significantly improving recall and accuracy metrics.

Yiding Wang, Zhepei Wei, Xinyu Zhu et al.

JEF-Hinter distills offline trajectories into compact, context-aware hints, boosting web agent performance by 15-20% with minimal inference overhead.

Hadi Nekoei, Aman Jaiswal, Patrice Bechard et al.

OptAgent combines multi-agent shopping simulation with genetic search, improving Etsy query fitness 21.98% over original queries and 3.36% over BoN.

Divij Handa, David Blincoe, Orson Adams et al.

R-AIRL uses adversarial IRL to learn reasoning rewards from expert demonstrations, improving training, inference reranking, and error localization.

Claudio Fanconi, Nicolás Astorga, Mihaela van der Schaar

This paper compares four reward models (dORM, dPRM, gORM, gPRM) across 14 multi-domain tasks, finding gORM most robust and effective.

Dong Bok Lee, Seanie Lee, Sangwoo Park et al.

AgentFlux uses decoupled fine-tuning with LoRA adapters, boosting local tool-calling accuracy by 46% on MCP-Bench.

Rohan Kadekodi, Zhan Jin, Keisuke Kamahori et al.

Proposes SafeMind with a three-layer safety architecture, boosting safety rate by over 30% in embodied LLM agents evaluated on 5558 multimodal samples.

Ruolin Chen, Yinqian Sun, Jihang Wang et al.

The study introduces a 'Perception to Cognition' framework to analyze bottlenecks in MLLMs for vision-language interaction.

Chenyue Zhou, Mingxuan Wang, Yanbiao Ma et al.

Cogito, ergo ludo (CEL) learns games through reasoning and planning, mastering diverse grid-world tasks.

Sai Wang, Yu Wu, Zhongwen Xu

Proposes EELMA, an information-theoretic method to estimate language model agents' empowerment, strongly correlating with task performance.

Jinyeop Song, Jeff Gore, Max Kleiman-Weiner

Proposed ASIMOV-2.0 benchmark integrates real injury reports and generative models to evaluate AI's physical risk perception and intervention.

Abhishek Jindal, Dmitry Kalashnikov, R. Alex Hofer et al.

Semantic F1 scores incorporate label semantic similarity, enabling fair evaluation of fuzzy multi-label predictions, outperforming traditional F1.

Georgios Chochlakis, Jackson Trager, Vedant Jhaveri et al.

Meta-Memory employs joint semantic-spatial reasoning with LLMs to significantly improve robot spatial question-answering, achieving 67.8% success on SpaceLocQA, outperforming SOTA.

Yufan Mao, Hanjing Ye, Wenlong Dong et al.

Proposes an on-device AI framework with adaptive memory and tool management, reducing context size by over 6x while maintaining performance.

Sanidhya Vijayvargiya, Rahul Lokesh

Proposed stereo-aware attention and dynamic fusion framework, boosting audio-visual navigation success rate by over 40%.

Jia Li, Yinfeng Yu, Liejun Wang et al.

End-to-end LLM-based solver (7B) reduces optimality gap to 1.03-8.20% via SFT and FOARL.

Xia Jiang, Yaoxin Wu, Minshuo Li et al.

Defines LLM agent generalization via hierarchical domain-task ontology; introduces evaluation metrics and multi-component optimization strategies.

Minxing Zhang, Yi Yang, Roy Xie et al.