14 papers
Temporal Evidence Forests for Long-Horizon Multi-Party Conversational Memory
Junyou Guo, Chongyang Tao, Yan He +2
Long-term memory has been widely adopted in LLM-based agents to support persistent interaction and personalized assistance. Long-term multi-party interactions, however, accumulate…
Beyond Entropy: Self-Diagnostic Multi-Role Token Optimization for Video Reasoning
Yudong Han, Yong Wang, Zaiquan Yang +4
Reinforcement learning with verifiable rewards has substantially advanced multimodal reasoning, yet it remains fundamentally limited by ambiguous token-level credit assignment. Whi…
DeFA: Dependency-Guided Failure Attribution for LLM Agents
Bo Deng, Xinlei Zheng, Yi Wei +6
Errors in LLM agent executions and their visible consequences can be separated by many steps, making decisive-error localization a matter of understanding both step content and ste…
FinEvo-Bench: A Longitudinal Benchmark for Self-Evolving Agents in Professional Financial Workflows
Bo Deng, Kang Zhou, Lifan Guo +6
Agents used over time encounter recurring professional work: each case requires different evidence and judgment, while the underlying workflow can be reused. Benchmarks built from…
UniQL: Towards Dialect-Universal Benchmarking for Text-to-SQL
Jianling Gao, Chongyang Tao, Jiayuan Bai +7
Existing text-to-SQL benchmarks are largely centered on SQLite, making it difficult to evaluate whether models can generalize across heterogeneous SQL dialects. However, real-world…
Unsat Core Prediction through Polarity-Aware Representation Learning over Clause-Literal Hypergraphs
Zhenchao Sun, Shuai Ma, Ping Lu +1
Graph neural networks have been widely used in Boolean satisfiability (SAT) tasks to learn structural information from SAT formulas. The goal of these studies is to solve SAT insta…
BCTuner: LLM-Guided Monte Carlo Tree Search for Efficient Blockchain Knob Tuning
Yaoyi Deng, Chongyang Tao, Mingxuan Li +4
Knob tuning plays a critical role in improving the performance of permissioned blockchains. However, efficient tuning remains challenging due to the architectural complexity of blo…
DEvo: Dual Difficulty-Aware Self-Evolution for Data-Efficient Reinforcement Learning
Ru Zhang, Renda Li, Ziyu Ma +4
Reinforcement learning (RL) has demonstrated potential for enhancing reasoning in large language models (LLMs). However, effective RL training, which requires medium-difficulty tra…
EvolSQL: Structure-Aware Evolution for Scalable Text-to-SQL Data Synthesis
Xuanguang Pan, Chongyang Tao, Jiayuan Bai +5
Training effective Text-to-SQL models remains challenging due to the scarcity of high-quality, diverse, and structurally complex datasets. Existing methods either rely on limited h…
AIR: Post-training Data Selection for Reasoning via Attention Head Influence
Jinrui Liu, Jeff Wu, Xuanguang Pan +3
LLMs achieve remarkable multi-step reasoning capabilities, yet effectively transferring these skills via post-training distillation remains challenging. Existing data selection met…
SetAD: Semi-Supervised Anomaly Learning in Contextual Sets
Jianling Gao, Chongyang Tao, Xuelian Lin +2
Semi-supervised anomaly detection (AD) has shown great promise by effectively leveraging limited labeled data. However, existing methods are typically structured around scoring ind…
Dynamic Stratified Contrastive Learning with Upstream Augmentation for MILP Branching
Tongkai Lu, Shuai Ma, Chongyang Tao
Mixed Integer Linear Programming (MILP) is a fundamental class of NP-hard problems that has garnered significant attention from both academia and industry. The Branch-and-Bound (B\…
JudgeSQL: Reasoning over SQL Candidates with Weighted Consensus Tournament
Jiayuan Bai, Xuan-guang Pan, Chongyang Tao +1
Text-to-SQL is a pivotal task that bridges natural language understanding and structured data access, yet it remains fundamentally challenging due to semantic ambiguity and complex…
Empowering Targeted Neighborhood Search via Hyper Tour for Large-Scale TSP
Tongkai Lu, Shuai Ma, Chongyang Tao
Traveling Salesman Problem (TSP) is a classic NP-hard problem that has garnered significant attention from both academia and industry. While neural-based methods have shown promise…