Papers by Deren Lei

6 papers
Beyond Reasoning Gains: Mitigating General-Capability Forgetting in Large Reasoning Models (2026.findings-acl)

Copied to clipboard

Challenge: Reinforcement learning with verifiable rewards (RLVR) has delivered impressive gains in mathematical and multimodal reasoning . however, the recipe introduces a significant risk of capability regression, where models forget foundational skills after prolonged training without employing regularization strategies.
Approach: They propose a replay strategy with dynamic objective reweighting for general knowledge preservation using short-horizon signals of convergence and instability.
Outcome: The proposed method preserves general capabilities and improves reasoning . it can be applied to existing RLVR pipelines without training additional models or tuning .
Deep Reinforcement Learning with Distributional Semantic Rewards for Abstractive Summarization (D19-1)

Copied to clipboard

Challenge: Abstractive summarization tasks are often based on deep reinforcement learning (RL) but the traditional reward system Rouge-L simply looks for exact n-grams matches between candidates and annotated references, which makes the generated sentences repetitive and incoherent.
Approach: They propose to use distributional semantics to measure matching degrees instead of Rouge-L to generate sentences with n-grams matches.
Outcome: The proposed reward has superiority over the existing reward, despite the incoherence of the generated sentences.
Learning Collaborative Agents with Rule Guidance for Knowledge Graph Reasoning (2020.emnlp-main)

Copied to clipboard

Challenge: Walk-based models have shown their advantages in knowledge graph reasoning but are limited by their representations and generalizability.
Approach: They propose a walk-based model that leverages high-quality rules generated by symbolic-based methods to provide reward supervision for walk- based agents.
Outcome: Experiments on benchmark datasets show that RuleGuider improves the performance of walk-based models without losing interpretability.
Imposing Label-Relational Inductive Bias for Extremely Fine-Grained Entity Typing (N19-1)

Copied to clipboard

Challenge: Existing entity typing systems exploit type hierarchy provided by KB schema to model label correlations.
Approach: They propose a graph layer that encodes global label co-occurrence statistics and word-level similarities.
Outcome: The proposed model achieves a 15.3% relative F1 improvement on a large dataset with over 10,000 free-form types.
Extract, Denoise and Enforce: Evaluating and Improving Concept Preservation for Text-to-Text Generation (2021.emnlp-main)

Copied to clipboard

Challenge: Existing models for text-to-text generation do not explicitly focus on important concepts in the input and output.
Approach: They propose a framework to automatically extract, denoise, and enforce important input concepts as lexical constraints.
Outcome: The proposed framework performs comparably or better than its unconstrained counterpart on automatic metrics and receives better ratings in the human evaluation.
FactCG: Enhancing Fact Checkers with Graph-Based Multi-Hop Data (2025.naacl-long)

Copied to clipboard

Challenge: Prior research on training grounded factuality classification models to detect hallucinations in large language models (LLMs) has relied on public natural language inference (NLI) data and synthetic data.
Approach: They propose a method that leverages multi-hop reasoning on context graphs extracted from documents to generate complex multi-level claims without relying on LLMs to decide data labels.
Outcome: The proposed model outperforms GPT-4-o on the LLM-Aggrefact benchmark with much smaller model size.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations