Papers by Kang Zhu

30 papers
Aligning VLM Assistants with Personalized Situated Cognition (2025.acl-long)

Copied to clipboard

Challenge: Existing studies on vision-language models aligned with general human objectives have not been successful because people with diversified backgrounds have different cognition even in the same situation.
Approach: They propose to characterize individuals based on the sociological concept of Role-Set and then evaluate their actions to see whether personalized alignment is achieved.
Outcome: The proposed framework constructs a cognition-aware and action-based reward model for personalized alignment.
Look Light, Think Heavy: What Multimodal Chain-of-Thought Reasoning Can and Cannot Do (2026.acl-long)

Copied to clipboard

Challenge: Existing open-source models often yield only marginal overall improvements, possibly due to an overemphasis on mathematical reasoning at the expense of broader capabilities.
Approach: They evaluate 12 multimodal tasks using 14 non-reasoning models and 8 reasoning models.
Outcome: The proposed method is effective in multimodal reasoning tasks, the authors show . they show that it lacks the ability to maintain deep visual introspection throughout the reasoning process.
ZipVoice-Dialog: Non-Autoregressive Spoken Dialogue Generation with Flow Matching (2026.findings-acl)

Copied to clipboard

Challenge: Existing autoregressive models for dialogue generation suffer from high latency and stability issues.
Approach: They propose a non-autoregressive (NAR) zero-shot spoken dialogue generation model based on flow-matching.
Outcome: The proposed model outperforms existing models in speech generation due to poor speech intelligibility and turn-taking precision.
TexSmart: A System for Enhanced Natural Language Understanding (2021.acl-demo)

Copied to clipboard

Challenge: TexSmart supports fine-grained named entity recognition (NER) Large-scale fine-granular entity types are expected to provide richer semantic information for downstream NLP applications.
Approach: They introduce TexSmart, a text understanding system that supports fine-grained named entity recognition (NER) and enhanced semantic analysis functionalities.
Outcome: The proposed system supports fine-grained named entity recognition (NER) and enhanced semantic analysis functions.
UTBoost: Rigorous Evaluation of Coding Agents on SWE-Bench (2025.acl-long)

Copied to clipboard

Challenge: Large language models have enabled the development of coding agents for real-world code generation.
Approach: They propose a novel LLM-driven test case generator that analyzes codebases and dependencies to generate test cases for real-world Python projects.
Outcome: The proposed framework improves the performance of SWE-Bench by analyzing codebases and dependencies.
Chinese Lexical Substitution: Dataset and Method (2023.emnlp-main)

Copied to clipboard

Challenge: Existing benchmarks for lexical substitution (LS) are limited and limited in coverage . despite extensive research on Lexical Substitution in various languages, there is limited evidence for LS in Chinese.
Approach: They propose to use human and machine collaboration to construct a Chinese LS dataset . they combine four unsupervised LS methods to generate candidate substitutes .
Outcome: The proposed method outperforms existing benchmarks on the Chinese lexical substitution task.
INSPIRED: Toward Sociable Recommendation Dialog Systems (2020.emnlp-main)

Copied to clipboard

Challenge: Existing studies on recommendation dialog systems lack a study on communication strategies used by human speakers for making successful and persuasive recommendations.
Approach: They propose to annotate a dataset of human-human movie recommendation dialogs with sociable recommendation strategies.
Outcome: The proposed model outperforms the baseline model in automatic and human evaluation.
BigVideo: A Large-scale Video Subtitle Translation Dataset for Multimodal Machine Translation (2023.findings-acl)

Copied to clipboard

Challenge: Existing datasets focus on captions describing images or videos, which are not large and diverse enough.
Approach: They propose a large-scale video subtitle translation dataset to facilitate multi-modality machine translation.
Outcome: The proposed dataset is 10 times larger than the widely used *How2* and *VaTeX* datasets.
TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities (2023.acl-demo)

Copied to clipboard

Challenge: Several pre-training models of different modalities are showing a rising trend of homogeneity in their model structures.
Approach: They propose a toolkit that supports pre-training models of different modalities.
Outcome: The proposed toolkit can match the performance of the original implementations on text, vision, and audio benchmarks.
Self-Specialization: Uncovering Latent Expertise within Large Language Models (2024.findings-acl)

Copied to clipboard

Challenge: Recent studies have demonstrated the effectiveness of self-alignment in which a large language model is aligned to follow general instructions using instructional data generated from the model itself.
Approach: They propose to use human-written seeds to align large language models to follow general instructions to achieve cross-task generalization.
Outcome: The proposed model outperforms base models and models that are generally instruction-tuned or have been adapted to the target domain by a large margin.
I²B-LPO: Latent Policy Optimization via Iterative Information Bottleneck (2026.acl-long)

Copied to clipboard

Challenge: Existing methods for large language model reasoning suffer from exploration collapse due to the semantic homogeneity of random rollouts.
Approach: They propose to use latent policy optimization via iterative information bottleneck to optimize reasoning trajectories by diversifying reasoning .
Outcome: Empirical results show that the proposed method achieves state-of-the-art performance with margins of up to 5.3% in accuracy and 7.4% in diversity metrics.
FoE: Forest of Errors Makes the First Solution the Best in Large Reasoning Models (2026.acl-long)

Copied to clipboard

Challenge: Recent Large Reasoning Models (LRMs) have demonstrated remarkable success in complex reasoning tasks.
Approach: They propose a self-guided efficient reasoning framework that reduces FoE by pruning subs.
Outcome: The proposed model outperforms eight competitive baselines while reducing token consumption by 37.7% 70.4%.
Guiding Medical Vision-Language Models with Diverse Visual Prompts: Framework Design and Comprehensive Exploration of Prompt Variations (2025.naacl-long)

Copied to clipboard

Challenge: Current vision-language models lack the ability to focus on specific areas designated by humans . a new framework that integrates medical entity extraction, visual prompt generation, and dataset adaptation is proposed to improve visual prompt-guided fine-tuning.
Approach: They propose to use visual prompts to guide and enhance formation of region-specific attention.
Outcome: The proposed framework outperforms state-of-the-art large vision-language models on medical datasets.
Leveraging Explicit Lexico-logical Alignments in Text-to-SQL Parsing (2022.acl-short)

Copied to clipboard

Challenge: Text-to-SQL parsing aims to parse natural language questions into SQL queries . current attention-based approaches can only model alignments at the token level .
Approach: They propose a method to leverage explicit lexico-logical alignments by identifying possible phrase-level alignments and injecting them as additional contexts into the parsing procedure.
Outcome: The proposed approach improves performance by 3.4% on Squall.
ParaLS: Lexical Substitution via Pretrained Paraphraser (2023.acl-long)

Copied to clipboard

Challenge: Lexical substitution (LS) is an extremely powerful technology that can be used as a backbone of various NLP applications such as writing assistance.
Approach: They propose two simple decoding strategies that focus on the variations of the target word during decoding to generate substitutes from a paraphraser.
Outcome: The proposed methods outperform state-of-the-art LS methods based on pre-trained language models on three benchmarks.
HMoE: Heterogeneous Mixture of Experts for Language Modeling (2025.emnlp-main)

Copied to clipboard

Challenge: Mixture of Experts (MoE) models use homogeneous experts with diverse capacities, resulting in a lack of expert specialization and parameter utilization.
Approach: They propose a framework where experts differ in size and possess diverse capacities . they propose HMoE to encourage frequent activation of smaller experts .
Outcome: The proposed framework outperforms homogeneous homogenous MoE models on evaluation benchmarks and achieves lower loss rate with fewer activated parameters.
NeuReasoner: Towards Explainable, Controllable, and Unified Reasoning via Mixture-of-Neurons (2026.acl-long)

Copied to clipboard

Challenge: Existing Large Reasoning Models (LRMs) lack explainability and controllability . Existing models target isolated levels without unification, while relying on RL .
Approach: They propose an explainable, controllable, and unified reasoning framework driven by MoN.
Outcome: The proposed framework achieves performance gains of 27.0% while reducing token consumption by 19.6% 63.3%.
Towards Graph-hop Retrieval and Reasoning in Complex Question Answering over Textual Database (2024.lrec-main)

Copied to clipboard

Challenge: Existing benchmarks for textual question answering only focus on single-chain or single-hop retrieval . Existing approaches to answer complex questions have limitations .
Approach: They propose to conduct Graph-Hop, a novel multi-chains and multi-hops retrieval paradigm in complex question answering.
Outcome: The proposed model provides explicit and fine-grained evidence graphs for complex question to support comprehensive and detailed reasoning.
JurisBench: A Deep Benchmark for Assessing Large Language Models in Professional Legal Practice (2026.acl-long)

Copied to clipboard

Challenge: Existing legal benchmarks evaluate isolated tasks or exam-style questions, failing to capture the procedural interdependencies and adjudicative rigor inherent in professional practice.
Approach: They propose a vertical, depth-oriented, domain-specific benchmark to evaluate Large Language Models (LLMs) in Chinese civil litigation.
Outcome: The proposed benchmarks show that large language models exhibit an "illusion of competence" the results highlight a critical gap between fluent linguistic output and judicial reliability .
STINMatch: Semi-Supervised Semantic-Topological Iteration Network for Financial Risk Detection via News Label Diffusion (2023.emnlp-main)

Copied to clipboard

Challenge: Commercial news provides rich semantics and timely information for automated financial risk detection.
Approach: They propose a semi-supervised Semantic-Topological Iteration Network, STINMatch, along with a news-enterprise knowledge graph to endorse the risk detection enhancement.
Outcome: The proposed model outperforms existing models in terms of generalization and semantics and annotation.
FocalOrder: Focal Preference Optimization for Reading Order Detection (2026.acl-long)

Copied to clipboard

Challenge: Existing methods for document comprehension rely on uniform supervision, resulting in a performance degradation in the intermediate sections.
Approach: They propose a framework driven by Focal Preference Optimization to detect reading order in document layouts.
Outcome: The proposed framework outperforms competing baselines and surpasses large-scale general VLMs.
LLaSE-G1: Incentivizing Generalization Capability for LLaMA-based Speech Enhancement (2025.acl-long)

Copied to clipboard

Challenge: Recent advances in language models have demonstrated strong capabilities in semantic understanding and contextual modeling.
Approach: They propose a LLaMA-based language model that incentivizes generalization capabilities for speech enhancement.
Outcome: The proposed language model outperforms prior task-specific discriminative and generative models in acoustic enhancement tasks.
Teams of LLM Agents can Exploit Zero-Day Vulnerabilities (2026.eacl-long)

Copied to clipboard

Challenge: Existing frameworks for LLM agents fail to exploit real-world zero-day vulnerabilities . prior work has shown that simple agents can hack mock "capture-the-flag" websites .
Approach: They propose a system of agents with a planning agent that can launch subagents to exploit real-world vulnerabilities.
Outcome: The proposed framework improves over previous frameworks by up to 4.3 on 14 real-world vulnerabilities.
SciMMIR: Benchmarking Scientific Multi-modal Information Retrieval (2024.findings-acl)

Copied to clipboard

Challenge: Multi-modal information retrieval (MMIR) is a rapidly evolving field . current benchmarks for image-text pairings overlook the scientific domain .
Approach: They develop a scientific domain-specific MMIR benchmark to evaluate image-text pairings using open-access research paper corpora.
Outcome: The proposed benchmarks are based on 530K image-text pairs extracted from scientific documents with detailed captions.
EasyQuant: An Efficient Data-free Quantization Algorithm for LLMs (2023.emnlp-main)

Copied to clipboard

Challenge: Recent work has shown that large language models are superior to conventional methods in various tasks.
Approach: They propose a data-independent quantization algorithm that leaves outliers in the weight and quantization ranges . they find the algorithm runs over 10 times faster than the data-dependent methods .
Outcome: The proposed method runs over 10 times faster than the data-dependent methods.
MeasHalu: Mitigation of Scientific Measurement Hallucinations for Large Language Models with Enhanced Reasoning (2026.findings-acl)

Copied to clipboard

Challenge: Large Language Models (LLMs) exhibit severe hallucinations, which undermine reliability of automated scientific document understanding systems.
Approach: They propose a framework for mitigating scientific measurement hallucinations through enhanced reasoning and targeted optimization.
Outcome: The proposed framework significantly reduces hallucination rates and improves overall accuracy on the MeasEval benchmark.
From Past To Path: Masked History Learning for Next-Item Prediction in Generative Recommendation (2026.acl-long)

Copied to clipboard

Challenge: Generative recommendation models inherently bias towards local contexts, failing to capture deeper historical dependencies necessary for understanding complex user intents.
Approach: They propose a training framework that shifts the objective from simple next-step prediction to deep comprehension of history by entropy-guided masking policy and a curriculum learning scheduler to enhance the framework.
Outcome: The proposed framework outperforms state-of-the-art generative models on three public datasets and shows that it is more accurate than current models.
Threads of Subtlety: Detecting Machine-Generated Texts Through Discourse Motifs (2024.acl-long)

Copied to clipboard

Challenge: Empirical findings show that although both LLMs and humans generate distinct discourse patterns influenced by specific domains, human-written texts exhibit more structural variability, reflecting the nuanced nature of human writing in different domains.
Approach: They propose a method to leverage hierarchical parse trees and recursive hypergraphs to uncover distinctive discourse patterns in texts written by humans and LLMs.
Outcome: The proposed method combines hierarchical parse trees and recursive hypergraphs to uncover distinctive discourse patterns in texts produced by both LLMs and humans.
FCDS: Fusing Constituency and Dependency Syntax into Document-Level Relation Extraction (2024.lrec-main)

Copied to clipboard

Challenge: Document-level Relation Extraction (DocRE) aims to identify relation labels between entities within a document.
Approach: They propose to fuse constituency and dependency syntax into DocRE to exploit the rich syntax information in the document.
Outcome: The proposed method is able to identify relation labels between entities within a document and is scalable.
Large Language Models are Better Reasoners with Self-Verification (2023.findings-emnlp)

Copied to clipboard

Challenge: Existing methods to solve complex natural language processing tasks require multiple steps to verify the answers.
Approach: They propose to use chain of thought prompting to solve reasoning tasks with large language models.
Outcome: The proposed method can improve reasoning performance on arithmetic, commonsense, and logical reasoning datasets.

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