Papers by Hengyuan Zhang

13 papers
Fine-grained Contrastive Learning for Definition Generation (2022.aacl-main)

Copied to clipboard

Challenge: Recent pre-trained transformer-based definition generation models lack effective representation learning to contain full semantic components of the given word, leading to under-specific definitions.
Approach: They propose a novel contrastive learning method that encourages the model to capture more detailed semantic representations from the definition sequence encoding.
Outcome: The proposed method could generate more specific definitions compared with state-of-the-art models.
ShifCon: Enhancing Non-Dominant Language Capabilities with a Shift-based Multilingual Contrastive Framework (2025.acl-long)

Copied to clipboard

Challenge: Experiments show that ShifCon significantly enhances the performance of non-dominant languages due to the imbalance in training data across languages.
Approach: They propose a Shift-based multilingual Contrastive framework that aligns the internal forward process of other languages toward that of the dominant one.
Outcome: The proposed framework significantly improves performance of non-dominant languages, particularly for low-resource ones.
Training LLMs for Divide-and-Conquer Reasoning Elevates Test-Time Scalability (2026.acl-long)

Copied to clipboard

Challenge: Large language models have demonstrated strong reasoning capabilities through step-by-step chain-of-thought (CoT) reasoning, but their strictly sequential nature constrains test-time scalability.
Approach: They propose an end-to-end reinforcement learning framework to enhance LLMs' DAC-style reasoning capacity by decomposing a problem into subproblems and solving them sequentially.
Outcome: The proposed model surpasses CoT by 8.6% and 6.3% on competition-level benchmarks and is available at the [github.com/MasterVito/DAC-RL].
Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Models (2026.findings-acl)

Copied to clipboard

Challenge: Existing literature on mechanistic interpretation (MI) treats it as an observational science, leaving practical applications underexplored.
Approach: They propose a survey structured around the pipeline to identify and improve MI models.
Outcome: The proposed framework enables tangible improvements in Alignment, Capability, and Efficiency.
Multi-level Contrastive Learning for Script-based Character Understanding (2023.emnlp-main)

Copied to clipboard

Challenge: Scripts are written text for plays, movies, or broadcasts.
Approach: They propose a multi-level contrastive learning framework to capture characters’ global information in a fine-grained manner.
Outcome: The proposed framework improves on three character understanding sub-tasks by a considerable margin.
Quantifying and Improving the Robustness of Retrieval-Augmented Language Models Against Spurious Features in Grounding Data (2026.acl-long)

Copied to clipboard

Challenge: Existing studies on robustness to explicit noise (e.g., document semantics) but overlook implicit noise (spurious features).
Approach: They propose a framework to quantify the robustness of RAGs against spurious features by integrating a data synthesis pipeline and a taxonomy.
Outcome: The proposed framework quantifies the robustness of RALMs against spurious features.
TreeReview: A Dynamic Tree of Questions Framework for Deep and Efficient LLM-based Scientific Peer Review (2025.emnlp-main)

Copied to clipboard

Challenge: Large Language Models (LLMs) have shown significant potential in assisting peer review, but current methods struggle to generate thorough and insightful reviews while maintaining efficiency.
Approach: They propose a framework that models paper review as a hierarchical and bidirectional question-answering process.
Outcome: The proposed framework outperforms baselines on full review generation and actionable feedback comments generation tasks while reducing LLM token usage by up to 80% compared to computationally intensive approaches.
BPO: Towards Balanced Preference Optimization between Knowledge Breadth and Depth in Alignment (2025.naacl-long)

Copied to clipboard

Challenge: Reinforcement Learning with Human Feedback (RLHF) is the key to the success of large language models (LLMs) in recent years.
Approach: They propose a method to balance the number of prompts and responses to improve knowledge breadth and knowledge depth by introducing gradient-based clustering to estimate the knowledge informativeness and usefulness of each augmented sample.
Outcome: The proposed method outperforms baseline methods while maintaining training efficiency.
Balancing Speciality and Versatility: a Coarse to Fine Framework for Supervised Fine-tuning Large Language Model (2024.findings-acl)

Copied to clipboard

Challenge: Aligned Large Language Models exhibit remarkable versatility, capable of handling diverse real-world tasks.
Approach: They propose a coarse to fine framework to fine-tune aligned Large Language Models to achieve a balance between speciality and versatility.
Outcome: The proposed framework outperforms baseline methods across diverse tasks and model scales.
Multitasking Framework for Unsupervised Simple Definition Generation (2022.acl-long)

Copied to clipboard

Challenge: Existing definition generation tasks require a dictionary with complex definitions and a corpus containing arbitrary simple texts to generate them.
Approach: They propose a multitasking framework SimpDefiner that only requires a standard dictionary with complex definitions and a corpus containing arbitrary simple texts.
Outcome: The proposed framework outperforms the baseline model by a 1.77 SARI score on the English dataset, and raises the proportion of the low level (HSK level 1-3) words in Chinese definitions by 3.87%.
Chain-of-Reasoning: Towards Unified Mathematical Reasoning in Large Language Models via a Multi-Paradigm Perspective (2025.acl-long)

Copied to clipboard

Challenge: Existing work shows that LLMs rely on single-paradigm reasoning that limits their effectiveness across diverse tasks.
Approach: They propose a new framework that integrates multiple reasoning paradigms to enable synergistic collaboration.
Outcome: The proposed model outperforms current SOTA models in theorem proving tasks and the MATH benchmark in arithmetic tasks.
Find Your Optimal Teacher: Personalized Data Synthesis via Router-Guided Multi-Teacher Distillation (2026.acl-long)

Copied to clipboard

Challenge: Existing studies show that stronger models are not always optimal teachers, suggesting a mismatch between the teacher’s output and the student’s learning ability.
Approach: They propose a method that routes each prompt to its optimal teacher via a query-level router that jointly considers the student models’ learnability and teacher models’ response quality.
Outcome: The proposed method outperforms baselines on six benchmarks including instruct tuning and math reasoning settings.
GuiLoMo: Allocating Experts and Ranks for LoRA-MoE via Bilevel Optimization with GuidedSelection Vectors (2025.findings-emnlp)

Copied to clipboard

Challenge: Low-Rank Adaptation (LoRA) methods are efficient for a large language model with reduced computational costs.
Approach: They propose a layer-wise expert numbers and ranks allocation strategy with GuidedSelection Vectors.
Outcome: The proposed method achieves superior or comparable performance to all baselines on three backbone models.

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