Papers by Chenliang Li

35 papers
Incentivizing In-depth Reasoning over Long Contexts with Process Advantage Shaping (2026.findings-acl)

Copied to clipboard

Challenge: Reinforcement Learning with Verifiable Rewards (RLVR) has proven effective in enhancing LLMs’ short-context reasoning but falters in long-contemporal scenarios requiring precise grounding and multi-hop reasoning.
Approach: They propose a framework that constructs high-difficulty, multi-hop long-context QA pairs with inherent reasoning chains to overcome this bottleneck.
Outcome: The proposed framework outperforms RLVR baselines and matches frontier LLMs while using far fewer parameters.
CharacterCraft: Bridging the Literature-Reality Dialogue Gap for Practical Role-Playing Agents (2025.findings-emnlp)

Copied to clipboard

Challenge: Existing dialogue datasets have a bias between query distributions and real-world user language usage.
Approach: They propose a framework for Chinese role-playing and a robust evaluation method . they propose specialized Chinese dialogue extraction model and specialized memory retrieval module .
Outcome: The proposed framework extracts character dialogue from novels and ensures high data quality.
SocialBench: Sociality Evaluation of Role-Playing Conversational Agents (2024.findings-acl)

Copied to clipboard

Challenge: Existing studies on role-playing agents have focused on enhancing their conversational capability, role-specific knowledge and style, but there has been a gap in assessing their social intelligence.
Approach: They propose a benchmark to evaluate the sociality of role-playing agents using LLMs.
Outcome: The proposed benchmark is constructed from various sources and covers a wide range of 500 characters and over 6,000 question prompts and 30,800 multi-turn role-playing utterances.
Writing-RL: Advancing Long-form Writing via Adaptive Curriculum Reinforcement Learning (2026.acl-long)

Copied to clipboard

Challenge: Recent advances in Large Language Models (LLMs) have enabled strong performance in long-form writing, but current training paradigms remain limited.
Approach: They propose an Adaptive Curriculum Reinforcement Learning framework to advance long-form writing capabilities beyond SFT.
Outcome: Experiments on 7B-scale writer models show that Writing-RL improves long-form writing performance over strong SFT baselines.
A Deep Relevance Model for Zero-Shot Document Filtering (P18-1)

Copied to clipboard

Challenge: Existing methods for document classification do not consider document filtering . existing methods do not include document filter.
Approach: They propose a deep relevance model for zero-shot document filtering called DAZER . they use word embeddings to extract the relevance signals from word embeds .
Outcome: The proposed model outperforms existing models on two document collections . it estimates the relevance between a document and a category by using seed words .
S2SPMN: A Simple and Effective Framework for Response Generation with Relevant Information (D18-1)

Copied to clipboard

Challenge: Existing work on how to generate relevant and informative responses is focusing on how dialogue systems generate information from large dialogue corpus.
Approach: They propose to use dialogue corpus to generate relevant responses by using prototypes to extract semantic information from PMN.
Outcome: The proposed model outperforms classical and strong baseline models in generating relevant and informative responses.
Guiding Generation for Abstractive Text Summarization Based on Key Information Guide Network (N18-2)

Copied to clipboard

Challenge: Abstractive text summarization models are hard to be controlled in the process of generation, which leads to a lack of key information.
Approach: They propose a guiding generation model that combines extractive and abstractive methods to generate text summarization.
Outcome: The proposed model improves on the CNN/Daily Mail dataset.
PALM: Pre-training an Autoencoding&Autoregressive Language Model for Context-conditioned Generation (2020.emnlp-main)

Copied to clipboard

Challenge: Existing techniques for natural language understanding and generation use autoencoding and/or autoregressive objectives to train models.
Approach: They propose a self-supervised pre-training scheme that pre-trains an autoencoding and autoregressive language model on a large unlabeled corpus for generating new text conditioned on context.
Outcome: The proposed scheme achieves state-of-the-art results on a variety of language generation benchmarks covering generative question answering, abstractive summarization and conversational response generation.
Incorporating External Knowledge into Machine Reading for Generative Question Answering (D19-1)

Copied to clipboard

Challenge: Existing knowledge-aware QA models do not have commonsense and background knowledge to answer nontrivial questions.
Approach: They propose a new neural model which exploits external knowledge to generate answers in natural language for a given question with context.
Outcome: The proposed model improves answer quality over existing models without knowledge and knowledge-aware models, a study shows . state officials in Hawaii confirmed that president Barack Obama was born in the U.S.
E2E-VLP: End-to-End Vision-Language Pre-training Enhanced by Visual Learning (2021.acl-long)

Copied to clipboard

Challenge: Existing vision-language pre-training methods use a two-step training procedure to learn visual features from image-text pairs.
Approach: They propose a vision-language pre-trained model for V+L understanding and generation using a unified Transformer framework.
Outcome: The proposed model can learn visual representation and semantic alignments between image and text on visual-text pairs and on visual processing tasks.
Small LLMs Are Weak Tool Learners: A Multi-LLM Agent (2024.emnlp-main)

Copied to clipboard

Challenge: Large Language Models (LLMs) have revolutionized natural language processing with impressive capabilities, but they lack domain specificity, real-time information and face challenges in solving specialized problems.
Approach: They propose a multi-LLM approach that decomposes the aforementioned capabilities into a planner, caller, and summarizer.
Outcome: The proposed model outperforms existing models by demonstrating its effectiveness and advantages in tool learning.
Weak-to-Strong Honesty Alignment via Learning-to-Rank Supervision (2025.findings-acl)

Copied to clipboard

Challenge: Existing approaches to enhance honesty with prompt engineering and fine-tuning are limited by annotated data.
Approach: They propose a framework that enhances honesty through weak-to-strong generalization by training weak LLMs under weak supervision to improve their honesty.
Outcome: The proposed framework improves honesty in large models even with limited label data.
Unifying Latent and Lexicon Representations for Effective Video-Text Retrieval (2024.lrec-main)

Copied to clipboard

Challenge: Existing methods for video-text retrieval capture fine-grained semantic concepts . however, they lack the ability to capture finer-grain concepts such as objects and actions.
Approach: They propose a dual-encoder architecture for fast video-text retrieval that learns lexicon representations to capture fine-grained semantics.
Outcome: The proposed framework outperforms existing methods with 4.8% and 8.2% improvement on MSR-VTT and DiDeMo respectively.
Learning to Translate by Translating: Stabilizing the Dual Loop via Semantic-Aware Self-Evolution (2026.findings-acl)

Copied to clipboard

Challenge: Large Language Models (LLMs) have been successful in machine translation, but lack of high-quality parallel corpora and cost constrain scalability.
Approach: They propose an LLM-driven dual-learning framework that enables autonomous translation . they employ a robust semantic-aware reward function that balances adequacy with reconstruction fidelity .
Outcome: The proposed model outperforms larger models on benchmarks and achieves parity with state-of-the-art supervised baselines on mainstream benchmarks.
UReader: Universal OCR-free Visually-situated Language Understanding with Multimodal Large Language Model (2023.findings-emnlp)

Copied to clipboard

Challenge: Existing studies for visually-situated language understanding have shown shallow zero-shot visual text recognition ability when fed a low-resolution image with salient text information.
Approach: They propose a model for universal OCR-free visually-situated language understanding based on the Multimodal Large Language Model (MLLM) their model is jointly finetuned on a wide range of visually situated language understanding tasks via a unified instruction format.
Outcome: The proposed model achieves state-of-the-art ocr-free performance in 8 out of 10 visually-situated language understanding tasks across 5 domains: documents, tables, charts, natural images, and webpage screenshots.
Semantics-enhanced Cross-modal Masked Image Modeling for Vision-Language Pre-training (2024.lrec-main)

Copied to clipboard

Challenge: Existing methods for vision-language pre-training lack high-level semantics and text is not sufficiently involved in masked modeling.
Approach: They propose a semantics-enhanced cross-modal MIM framework for vision-language representation learning that harvests high-level semantics from global image features via self-supervised agreement learning and transfers them to local patch encodings by sharing the encode space.
Outcome: The proposed model achieves state-of-the-art or competitive performance on multiple vision-language tasks.
AIGuard: A Benchmark and Lightweight Detection for E-commerce AIGC Risks (2025.findings-acl)

Copied to clipboard

Challenge: Existing detection methods lack real-world scenarios and corresponding risk datasets . current MLLMs lack knowledge and have limited capability to detect the risk of AIGC content.
Approach: They propose a benchmark for AIGC risk detection in real-world e-commerce . it includes 253,420 image-text pairs across four critical categories .
Outcome: The proposed method achieves 9.68% higher recall than leading multimodal models while using only 25% of training resources.
Addressing Semantic Drift in Generative Question Answering with Auxiliary Extraction (2021.acl-short)

Copied to clipboard

Challenge: Recent work focuses on question answering based on machine reading comprehension . current approaches treat QA as extracting a consecutive piece of text to a given question.
Approach: They propose a generative QA model that incorporates an extractive mechanism into a model.
Outcome: The proposed model improves quality and semantic accuracy over baseline models.
ModelScope-Agent: Building Your Customizable Agent System with Open-source Large Language Models (2023.emnlp-demo)

Copied to clipboard

Challenge: Large language models (LLMs) have demonstrated remarkable capabilities to comprehend human intentions, engage in reasoning, and design planning-like behavior.
Approach: They propose a framework that equips large language models with tool-use capabilities . they propose LLaMA and Chat-GLM as controllers, and a model-based agent framework .
Outcome: The proposed framework equips open-source LLMs with tool-use capabilities . it provides a user-friendly system library with a customizable engine design .
META-LORA: Memory-Efficient Sample Reweighting for Fine-Tuning Large Language Models (2025.coling-main)

Copied to clipboard

Challenge: Supervised fine-tuning (SFT) is widely adopted for tailoring large language models (LLMs) to specific downstream tasks.
Approach: They propose a memory-efficient method for automatic sample reweighting that learns to re-weight fine-tuning samples by minimizing the loss on a small, high-quality validation set.
Outcome: Meta-LoRA learns to reweight fine-tuning samples by minimizing the loss on a small, high-quality validation set through an end-to-end bi-level optimization framework based on meta-learning.
Joint Document-Level Event Extraction via Token-Token Bidirectional Event Completed Graph (2023.acl-long)

Copied to clipboard

Challenge: a joint exaction method can be used to extract document-level event records . it avoids inefficiency and error propagation issues in traditional pipeline methods .
Approach: They propose a joint exaction method that can avoid inefficiency and error propagation issues . they propose eType-Role1-Roul2 as the edge type to reveal which tokens play argument roles .
Outcome: The proposed method can avoid inefficiency and error propagation issues in traditional pipeline methods.
Transforming Visual Scene Graphs to Image Captions (2023.acl-long)

Copied to clipboard

Challenge: Existing approaches to generate captions using image captioning are based on multi-head attention (MHA)
Approach: They propose to transform scene graphs into more descriptive captions by using multi-head attention to build a Graph Neural Network (GNN) . they construct a Mixture-of-Expert (MOE)-based decoder where each expert is built on MHA for discriminating the graph embeddings to generate different kinds of words.
Outcome: The proposed framework can generate captions from multiple visual features and objects . it is based on a mixture-of-expert (MOE)-based decoder based upon MHA .
MUSEG: Reinforcing Video Temporal Understanding via Timestamp-Aware Multi-Segment Grounding (2026.acl-long)

Copied to clipboard

Challenge: Existing methods for MLLMs struggle with fine-grained temporal reasoning . despite advances in video understanding, current methods struggle with time-sensitive tasks .
Approach: They propose a time-stamp-aware multi-segment grounding method that enhances temporal understanding by introducing timestamps.
Outcome: The proposed method outperforms existing methods on time-sensitive tasks and generalizes well across diverse temporal understanding scenarios.
MATINF: A Jointly Labeled Large-Scale Dataset for Classification, Question Answering and Summarization (2020.acl-main)

Copied to clipboard

Challenge: Existing datasets labeled for one task hinder multi-task learning . task-specific data make models learn task-related leakage features rather than meaningful knowledge that could generalize to other tasks.
Approach: They propose to jointly label large-scale NLP dataset MATINF . it contains 1.07 million question-answer pairs with human-labeled categories .
Outcome: The proposed dataset is applicable for classification, question answering, and summarization.
Adaptive Feature Discrimination and Denoising for Asymmetric Text Matching (2022.coling-1)

Copied to clipboard

Challenge: Existing models focus on asymmetric text matching but rarely perform feature denoising . existing models focus only on recognizing discriminative features and filtering out irrelevant features .
Approach: They propose a novel adaptive feature discrimination and denoising model for asymmetric text matching . it explicitly distinguishes discriminative features and filters out irrelevant features in context .
Outcome: The proposed model achieves significant performance gains over current state-of-the-art models on four real-world datasets.
StructuralLM: Structural Pre-training for Form Understanding (2021.acl-long)

Copied to clipboard

Challenge: Existing pre-trained language models focus on text-only representation, neglecting cell-level layout information.
Approach: They propose a pre-training approach to leverage cell and layout information from scanned documents.
Outcome: The proposed model achieves state-of-the-art in various downstream tasks . it uses 2Dposition embeddings to model word-level layout information .
mPLUG: Effective and Efficient Vision-Language Learning by Cross-modal Skip-connections (2022.emnlp-main)

Copied to clipboard

Challenge: Existing pre-trained vision-language models suffer from inefficiency and linguistic signal overwhelmed by long visual sequences in cross-modal alignment.
Approach: They propose a vision-language foundation model with cross-modal skip-connections that can be pre-trained end-to-end on large-scale image-text pairs with both discriminative and generative objectives.
Outcome: The proposed model achieves state-of-the-art results on a wide range of vision-language downstream tasks, including image captioning, image-text retrieval, visual grounding and visual question answering.
Mitigating Language Confusion through Inference-time Intervention (2025.coling-main)

Copied to clipboard

Challenge: Existing methods to address the problem of language confusion are incontext learning and supervised fine-tuning (SFT) however, they consume context window space and require extensive data collection.
Approach: They propose a language-sensitive intervention that detects and assesses language confusion without additional complex mechanisms.
Outcome: The proposed method detects language confusion and assesses content quality without additional complex mechanisms.
UnihanLM: Coarse-to-Fine Chinese-Japanese Language Model Pretraining with the Unihan Database (2020.aacl-main)

Copied to clipboard

Challenge: Chinese and Japanese share many characters with similar surface morphology.
Approach: They propose a Chinese-Japanese pretrained masked language model with a coarse-to-fine training approach to exploit the shared knowledge across the languages.
Outcome: The proposed model is effective on mono- and cross-lingual Chinese and Japanese tasks.
Analyzing and Modeling LLM Response Lengths with Extreme Value Theory: Anchoring Effects and Hybrid Distributions (2025.emnlp-main)

Copied to clipboard

Challenge: Existing approaches treat length as an incidental output property rather than a statistically regular phenomenon worthy of rigorous modeling.
Approach: They propose a statistical framework for modeling and controlling large language model response lengths using extreme value theory and cross-validation on Qwen and DeepSeek architectures.
Outcome: The proposed model improves tail fit and generalizability while maintaining generalizzability.
Pre-train and Plug-in: Flexible Conditional Text Generation with Variational Auto-Encoders (2020.acl-main)

Copied to clipboard

Challenge: Existing conditional generation models cannot handle emerging conditions due to their joint end-to-end learning fashion.
Approach: They propose a framework for conditional text generation that decouples the text generation module from the condition representation module to allow "one-to-many" conditional generation.
Outcome: The proposed framework decouples the text generation module from the condition representation module to allow “one-to-many” conditional generation.
TRIPS: Efficient Vision-and-Language Pre-training with Text-Relevant Image Patch Selection (2022.emnlp-main)

Copied to clipboard

Challenge: Existing vision-and-language pre-training models suffer from long visual sequences . experimental results show that TRIPS gains a speedup of 40% over previous similar VLP models .
Approach: They propose an efficient vision-and-language pre-training model with text-relevant image patch selection, TRIPS, which reduces the visual sequence progressively with a text-guided patch-selection layer in the visual backbone for efficient training and inference.
Outcome: The proposed model can speed up training and inference by 40% over previous models.
Efficient Sparse Attention needs Adaptive Token Release (2024.findings-acl)

Copied to clipboard

Challenge: Large Language Models (LLMs) have demonstrated remarkable capabilities across a wide array of text-centric tasks, however, their ‘large’ scale introduces significant computational and storage challenges, particularly in managing the key-value states of the transformer, which limits their wider applicability.
Approach: They propose to release resources from caches and rebuild key-value states by a lightweight controller module to approximate an ideal top-K sparse attention.
Outcome: The proposed method achieves a significant throughput improvement of 221.8% over full attention and a model with 7 billion tokens.
Token-level Preference Self-Alignment Optimization for Multi-style Outline Controllable Generation (2025.findings-acl)

Copied to clipboard

Challenge: Existing attempts to outline generation are limited by response pair requirements and substantial computation costs.
Approach: They propose a token-level preference self-alignment optimization for outline controllable generation that extends the Bradley-Terry model from pair-wise to list-wise comparison.
Outcome: The proposed method outperforms existing methods by 19.28% in performance while requiring only 56.25% training time.
Dependency Graph Enhanced Dual-transformer Structure for Aspect-based Sentiment Classification (2020.acl-main)

Copied to clipboard

Challenge: Aspect-based sentiment classification is a popular task aimed at identifying the corresponding emotion of a given aspect.
Approach: They propose a dependency graph enhanced dual-transformer network to support mutual reinforcement between the flat representation learning and graph-based representation learning.
Outcome: The proposed model outperforms state-of-the-art methods on five datasets with a large margin.

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