(IEEE TKDE2025) Learnable Game-theoretic Policy Optimization for Data-centric Self-explanation Rationalization Paper Code

Yunxiao Zhao, Zhiqiang Wang, Xingtong Yu, Xiaoli Li, Jiye Liang, Ru Li.

• In this paper, we systematically revisit cooperative rationalization from a novel game-theoretic perspective and identify the fundamental cause of rationale collapse. We then propose a novel approach, Game-theoretic Policy Optimization oriented RATionalization (PoRAT), which progressively introduces policy interventions to address the game equilibrium in the cooperative game process, thereby guiding the model toward a more optimal solution state.

(ACL2024 Main) AGR: Reinforced Causal Agent Guided Self-explaining Rationalization Paper Code

Yunxiao Zhao, Zhiqiang Wang, Xiaoli Li, Jiye Liang, Ru Li.

• We propose a novel approach AGR (Agent-Guided Rationalization), guiding the next action of the model based on its current training state. We introduce causal intervention calculus to quantify the causal effects inherent during rationale training, and utilize the reinforcement learning process to refine their learning bias.

(DASFAA2025 Main) Explaining Black-box Language Models with Knowledge Probing Systems: A Post-hoc Explanation Perspective Paper Code Tool

Yunxiao Zhao, Hao Xu, Zhiqiang Wang, Xiaoli Li, Jiye Liang, Ru Li.

• Pre-trained Language Models (PLMs) are trained on large amounts of unlabeled data, and they exhibit remarkable reasoning skills. However, the trustworthiness challenges have become increasingly evident. To alleviate this problem, we propose a novel knowledge-guided probing approach called KnowProb in a post-hoc explanation way, which aims to probe whether black-box PLMs understand implicit knowledge beyond the given text, rather than focusing only on the surface-level content of the text.

(EMNLP2026 Main) CuRLRank: Curriculum-Guided Reinforcement Learning for Reasoning-Intensive Reranking Paper Code

Xin Kang, Yunxiao Zhao*, Jiayang Zhang, Juncai Li, Zhichao Yan, XiaoQi Han, Ru Li, Delvin Ce Zhang. (* = corresponding author)

• We propose CuRLRank, which consists of (1) CuRL, an adaptive curriculum reinforcement learning approach that enables LLMs to develop complex reasoning capabilities across varying levels of difficulty, and (2) IFRS, an iterative filtering ranking strategy to achieve efficient and robust ranking.

(EMNLP2026 Main) Responsibility-Aware Scheduling for Selective Rationalization Paper Code

Dengchao Zhang, Yunxiao Zhao*, Zhichao Yan, Yue Fan, Shaoru Guo, Ru Li*. (* = corresponding author)

• We propose ReSR (Responsibility-Aware Scheduling for Rationalization), a training-control framework that keeps the standard select-then-predict path while adaptively scheduling generator and predictor updates. ReSR uses online, probe, and reference diagnostics with rationale-structure signals to attribute failures to generator-side rationale quality or predictor lag; a virtual-runtime scheduler then converts this attribution into generator/predictor debts and allocates the next optimization slice.

(EMNLP2026 Findings) MulEG: Multi-Level Explanation-Guided Feature Attribute Graph for Zero-Shot Tabular Classification Paper Code

Jiayang Zhang, Yunxiao Zhao*, Xin Kang, Zhichao Yan, Boxiang Ma, Qinghua Chai, Xingtong Yu, Ru Li*. (* = corresponding author)

• We propose a novel method (MulEG), which harnesses multi-level textual explanations with a feature attribute graph to model both tabular topology and cell-element relations.

(COLING2025 Main) LOG: A Local-to-Global Optimization Approach for Retrieval-based Explainable Multi-Hop Question Answering Paper Code

Hao Xu, Yunxiao Zhao*, Jiayang Zhang, Zhiqiang Wang, Ru Li. (* = corresponding author)

• Explainable multi-hop question answering aims to utilize multi-source intensive documents retrieved to derive the answer. However, it is challenging to model the importance of knowledge retrieved. In this paper, we propose LOG, a novel optimized retrieval method to discover more beneficial knowledge from a local-to-global perspective, facilitating multi-hop reasoning, notably for long-chain reasoning.

(JMIR2024) A Comprehensive Overview of CFN from a Commonsense Perspective Paper Code

Ru Li, Yunxiao Zhao, Zhiqiang Wang, Xuefeng Su, Shaoru Guo, Yong Guan, Xiaoqi Han, Hongyan Zhao.

• We conduct a comprehensive overview of Chinese FrameNet from a commonsense perspective, covering topics such as scenario commonsense representation, Chinese FrameNet resources, and its applications. We also summarize recent breakthroughs and identify future research directions.

(ACL2026 Main) Leibniz: Theory-of-Mind Driven Neuro-Symbolic Logical Reasoning via Multi-Agent Collaboration Paper Code

Yue Fan, Hu zhang, Yunxiao Zhao, Guangjun Zhang, Hao ZHAN, Ru Li, Hongye Tan, Wang Yuanlong

• We propose Leibniz, a theory-of-mind driven neuro-symbolic reasoning framework. We construct a bidirectional reasoning model based on multi-agent collaboration, which characterizes the reasoning process from two complementary perspectives.

(ACL2026 Main) Suggest-Verify-Revise: A Three-Stage Document-Level Event Causality Identification with Narrative Consistency Paper Code

Ya Su, Hu Zhang, Dan Qiao, Yujie Wang, Yunxiao Zhao, Yue Fan, Shike Li, Ru Li, Hongye Tan

• Document-level Event Causality Identification aims to identify causal relations among multiple events within unstructured text. To model the overall narrative backbone in the propagation of causal dependencies and the role differentiation of events within multi-cause/multi-effect structures, we propose a suggest-verify-revise approach for document-level event causality identification with narrative consistency.

(ACL2021 Findings) GCRC: A New Challenging MRC Dataset from Gaokao Chinese for Explainable Evaluation Paper Code Leaderboard

Hongye Tan, Xiaoyue Wang, Yu Ji, Ru Li, Xiaoli Li, Zhiwei Hu, Yunxiao Zhao, Xiaoqi Han.

• In this paper, we propose GCRC, a new explainable evaluation dataset with challenging and high-quality multi-choice questions, collected from Gaokao Chinese (Chinese subject from the National College Entrance Examination of China).