Papers by Peng Fang

17 papers
Model Composition for Multimodal Large Language Models (2024.acl-long)

Copied to clipboard

Challenge: Existing methods for creating versatile MLLMs rely on joint training with paired instruction data, which is resource-intensive and challenging to extend to new modalities.
Approach: They propose a new paradigm for multimodal large language models by reusing modality encoders and merging LLM parameters.
Outcome: The proposed model retains the modal understanding capabilities of each original model.
LearnAlign: Data Selection for LLM Reinforcement Learning with Improved Gradient Alignment (2026.findings-acl)

Copied to clipboard

Challenge: Reinforcement learning with verifiable rewards (RLVR) is a key technique for enhancing LLMs’ reasoning abilities, yet its data inefficiency remains a major bottleneck.
Approach: They propose a gradient-alignment-based method which intelligently selects the learnable and representative training reasoning data for RLVR post-training.
Outcome: Experiments on five reasoning benchmarks show that the proposed method significantly reduces training data requirements while improving performance.
A Semi-supervised Scalable Unified Framework for E-commerce Query Classification (2025.acl-industry)

Copied to clipboard

Challenge: Existing query classification methods rely on posterior click behavior to construct training samples, resulting in insufficient prior information for modeling.
Approach: They propose a semi-supervised scaleable unified framework that integrates enhanced modules to unify query classification tasks.
Outcome: The proposed framework outperforms the state-of-the-art models in offline and online A/B experiments.
C-World: A Computer Use Agent Environment Creator (2026.acl-long)

Copied to clipboard

Challenge: C-World enables users to build agent environments on demand.
Approach: They propose a system that enables users to build agent environments on demand.
Outcome: The proposed system outperforms baselines on 119k samples and achieves Spearman = 0.883 ranking correlation with real execution.
If an LLM Were a Character, Would It Know Its Own Story? Evaluating Lifelong Learning in LLMs (2026.acl-long)

Copied to clipboard

Challenge: Existing benchmarks for large language models (LLMs) fail to capture these dynamics, focusing on static, open-ended evaluations.
Approach: They propose a benchmark to assess lifelong learning in large language models . they use two episodic datasets rich in narrative structure and character interactions .
Outcome: Experiments on LLMs show that non-parametric methods outperform parametric ones in managing stateful learning.
CoT-RAG: Integrating Chain of Thought and Retrieval-Augmented Generation to Enhance Reasoning in Large Language Models (2025.findings-emnlp)

Copied to clipboard

Challenge: Chain-of-thought reasoning has two key limitations: lack of reliability when solely relying on LLM-generated reasoning chains and interference from natural language reasoning steps with the models’ inference logic.
Approach: They propose a chain-of-thought reasoning framework with three key designs to address these issues.
Outcome: The proposed framework improves the performance of large language models on complex tasks by incorporating knowledge graphs and learnable knowledge case-aware RAG.
Unifying Language Agent Algorithms with Graph-based Orchestration Engine for Reproducible Agent Research (2025.acl-demo)

Copied to clipboard

Challenge: Language agents powered by large language models (LLMs) have demonstrated remarkable capabilities in understanding, reasoning, and executing complex tasks.
Approach: They propose a flexible framework that addresses engineering overhead and insufficient evaluation frameworks for fair comparison.
Outcome: The proposed framework simplifies language agent development and establishes a foundation for reproducible agent research.
CaKE: Circuit-aware Editing Enables Generalizable Knowledge Learners (2025.emnlp-main)

Copied to clipboard

Challenge: Existing knowledge editing methods fail to generalize updates to multi-hop reasoning tasks . Existing methods only edit single or a few model layers, inadequately integrate updated knowledge into reasoning pathways.
Approach: They propose a circuit-aware method that enhances the effective integration of updated knowledge in large language models by leveraging curated data samples guided by their analysis.
Outcome: The proposed method improves accuracy and accuracy of 20% on the MQuAKE dataset while requiring less memory.
Enhancing Input-Label Mapping in In-Context Learning with Contrastive Decoding (2025.acl-short)

Copied to clipboard

Challenge: Prior research has found that large language models overlook input-label mapping information in ICL, relying more on their pre-trained knowledge.
Approach: They propose a novel method that contrasts input-label mappings between positive and negative in-context examples to improve model performance.
Outcome: The proposed method improves performance on 7 natural language understanding tasks without additional training.
Position-Aware Depth Decay Decoding (D3): Boosting Large Language Model Inference Efficiency (2025.findings-acl)

Copied to clipboard

Challenge: Recent dynamic computation methods show that not all components are required for inference, enabling a training-free pipeline.
Approach: They propose a token-position aware layer skipping framework to save 1.5x times operations efficiently while maintaining performance.
Outcome: The proposed algorithm achieves 1.5x speedup on large language models with no retraining and with comparable performance on the GSM8K and BBH benchmarks.
HierDiffuse: Progressive Diffusion for Robust Interest Fusion in CTR Prediction (2025.emnlp-industry)

Copied to clipboard

Challenge: Existing approaches fuse long-term behavioral profiles and short-term interactions, suffering from representational misalignment and noise in transient signals.
Approach: They propose a framework that redefines interest fusion as a hierarchical denoising process through diffusion models.
Outcome: The proposed framework redefines interest fusion as a hierarchical denoising process through diffusion models.
KELE: A Multi-Agent Framework for Structured Socratic Teaching with Large Language Models (2025.findings-emnlp)

Copied to clipboard

Challenge: Socratic teaching places high demands on teachers’ expertise and real-time feedback capabilities, making it difficult to scale in large educational settings.
Approach: They propose a multi-agent framework for structured Socratic teaching with LLMs that integrates a structured SocRule and a consultant-teacher collaborative teaching mechanism.
Outcome: The proposed framework outperforms existing LLMs in natural language generation and dialogue comprehension in the classroom.
Tokenization Falling Short: On Subword Robustness in Large Language Models (2024.findings-emnlp)

Copied to clipboard

Challenge: Language models typically tokenize raw text into sequences of subword identifiers from a predefined vocabulary.
Approach: They propose to tokenize raw text into sequences of subword identifiers from a predefined vocabulary . they also investigate the challenges and their impact on large language models .
Outcome: The proposed model can mitigate tokenization issues, but still suffer from typos and other variations.
Rethinking Cross-Subject Data Splitting for Brain-to-Text Decoding (2025.emnlp-main)

Copied to clipboard

Challenge: Recent studies have successfully decoded natural language from non-invasive brain signals . current dataset splitting methods suffer from data leakage problem .
Approach: They propose a right cross-subject data splitting criterion without data leakage for decoding fMRI and EEG signal to text.
Outcome: The proposed method overfits and overestimates brain-to-text decoding models.
Vision-Language Introspection: Mitigating Overconfident Hallucinations in MLLMs via Interpretable Bi-Causal Steering (2026.acl-long)

Copied to clipboard

Challenge: Existing approaches to overcome object hallucination are limited . Existing mitigations include costly retraining and a training-free inference framework .
Approach: They propose a training-free inference framework that simulates a metacognitive self-correction process.
Outcome: The proposed framework reduces object hallucination rates by 12.67% on MMHal-Bench and improves accuracy by 5.8% on POPE.
CAFE: Retrieval Head-based Coarse-to-Fine Information Seeking to Enhance Multi-Document QA Capability (2025.emnlp-main)

Copied to clipboard

Challenge: Existing methods to extend context length of Large Language Models (LLMs) still struggle with retrieval and reasoning in long context inputs.
Approach: They propose a coarse-to-fine method to enhance multi-document question-answering capacities by removing background and distracting documents.
Outcome: Experiments show that CAFE outperforms baseline methods on multiple documents.
ConvLab-2: An Open-Source Toolkit for Building, Evaluating, and Diagnosing Dialogue Systems (2020.acl-demos)

Copied to clipboard

Challenge: ConvLab-2 inherits Convlab's framework but integrates more powerful dialogue models and supports more datasets.
Approach: They present ConvLab-2, an open-source toolkit that enables researchers to build task-oriented dialogue systems with state-of-the-art models and perform an end-to-end evaluation.
Outcome: The new tool inherits ConvLab's framework and extends it by integrating many recently proposed state-of-the-art dialogue 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