Kai-Wei Chang
Professor
Develops natural language processing and machine learning methods, including vision-language models, robustness and fairness of language models.
Kai-Wei Chang's lab at UCLA works on large language models, interpretability, evaluation & safety of LMs and vision-language & multimodal. 152 papers since 2023, 104 at top venues, mostly ACL and EMNLP. 46 alumni, most often to Meta and Amazon. Applications close .
At a glance
- papers, 2023–2026
- 40/ year
- at T1 venues
- 68%
- PhD students
- 0
- MS students
- 0
- alumni listed
- 46
Papers by yearfrom the lab homepage, 2023–2026
- 29 papers2026 · T1 18 · collecting
- 44 papers2025 · T1 29
- 43 papers2024 · T1 30
- 36 papers2023 · T1 27
T1 venuesOtherVenues: ACL, EMNLP, NeurIPS, ICLR, Findings of ACL
Research focuspaper titles that use each topic's words
- Large language models77 papers
- Vision-language & multimodal54 papers
- Interpretability, evaluation & safety of LMs13 papers
- vision-language11 papers
- vision-language models6 papers
Paper list
From the lab homepage.
202629 papers
- COLMAgent Q-Mix: Selecting the Right Action for LLM Multi-Agent Systems through Reinforcement Learning
- ACL (T1)ARES: Adaptive Red-Teaming and End-to-End Repair of Policy-Reward System
- Findings of ACLAutoSUIT Bench - Automated Security UnIt Test Benchmark for LLM Coding
- Findings of ACLBeyond Facts: Benchmarking Distributional Reading Comprehension in Large Language Models
- EACLBLUR: A Bi-Level Optimization Approach for LLM Unlearning
- Findings of ACLBRIEF-Pro: Universal Context Compression with Short-to-Long Synthesis for Fast and Accurate Multi-Hop Reasoning
- ICLR (T1)ContextNav: Towards Agentic Multimodal In-Context Learning
- NeurIPS (T1)Does Inference-Time Reasoning Really Improve Video Understanding?
- ACL (T1)Dynamic Generation of Multi LLM Agents Communication Topologies with Graph Diffusion Models
- ACL (T1)From Narrow Unlearning to Emergent Misalignment in LLMs
- COLMFronTalk: Benchmarking Front-End Development as Conversational Code Generation with Multi-Modal Feedback
- ACL (T1)Gold-Medal-Level Olympiad Geometry Solving with Efficient Heuristic Auxiliary Constructions
202544 papers
- NeurIPS (T1)3DLLM-Mem: Long-Term Spatial-Temporal Memory for Embodied 3D Large Language Model
- COLMA Meta-Evaluation of Measuring LLM Misgendering
- NAACL-FindingBRIEF: Bridging Retrieval and Inference for Multi-hop Reasoning via Compression
- ACL-FindingComparing Bad Apples to Good Oranges: Aligning Large Language Models via Joint Preference Optimization
- ICML (T1)Contradiction Retrieval via Contrastive Learning with Sparsity
- ICML (T1)Contrastive Visual Data Augmentation
- ICLR (T1)Controllable Generation via Locally Constrained Resampling
- COLMCustomize Multi-modal RAI Guardrails with Precedent-based predictions
- EMNLP (T1)DiCoRe: Enhancing Zero-shot Event Detection via Divergent-Convergent LLM Reasoning
- Findings of ACLDRS: Deep Question Reformulation With Structured Output
- NeurIPS (T1)Embodied Web Agents: Bridging Physical-Digital Realms for Integrated Agent Intelligence
- CVPR (T1)Exploring Visual Vulnerabilities via Multi-Loss Adversarial Search for Jailbreaking Vision-Language Models
202443 papers
- ACL (T1)Agent Lumos: Unified and Modular Training for Open-Source Language Agents
- American Journal of NeuroradiologyAI-Assisted Summarization of Radiologic Reports: Evaluating GPT3davinci, BARTcnn, LongT5booksum, LEDbooksum, LEDlegal, and LEDclinical
- Findings of ACLAre LLMs Capable of Data-based Statistical and Causal Reasoning? Benchmarking Advanced Quantitative Reasoning with Data
- Findings of NAACLAre you talking to ['xem'] or ['x', 'em']? On Tokenization and Addressing Misgendering in LLMs with Pronoun Tokenization Parity
- EMNLP-FindingAttribute Controlled Fine-tuning for Large Language Models: A Case Study on Detoxification
- LRECCan small language models help large language models reason better?: LM-guided chain-of-thought
- NAACL (T1)CASA: Causality-driven Argument Sufficiency Assessment
- ICML (T1)Characterizing Truthfulness in Large Language Model Generations with Local Intrinsic Dimension
- ICLR (T1)CoBIT: A Contrastive Bi-directional Image-Text Generation Model
- NAACL (T1)Contextual Label Projection for Cross-Lingual Structured Prediction
- ICML (T1)ConTextual: Evaluating Context-Sensitive Text-Rich Visual Reasoning in Large Multimodal Models
- EMNLP (T1)Control Large Language Models via Divide and Conquer
202336 papers
- EMNLP (T1)"What's 'up' with vision-language models? Investigating their struggle to understand spatial relations."
- NeurIPS (T1)A Pseudo-Semantic Loss for Deep Generative Models with Logical Constraints
- ACL (T1)A Survey of Deep Learning for Mathematical Reasoning
- EMNLP (T1)Active Instruction Tuning: Improving Cross-Task Generalization by Training on Prompt Sensitive Tasks
- EMNLP-FindingAre Personalized Stochastic Parrots More Dangerous? Evaluating Persona Biases in Dialogue Systems
- ACL-FindingAVATAR: A Parallel Corpus for Java-Python Program Translation
- NeurIPS (T1)AVIS: Autonomous Visual Information Seeking with Large Language Models
- NeurIPS (T1)Chameleon: Plug-and-Play Compositional Reasoning with Large Language Models
- ICCV (T1)CleanCLIP: Mitigating Data Poisoning Attacks in Multimodal Contrastive Learning
- Findings of EMNLPDataset Bias Mitigation in Multiple-Choice Visual Question Answering and Beyond
- NeurIPS (T1)DesCo: Learning Object Recognition with Rich Language Descriptions
- ICLR (T1)Dynamic Prompt Learning via Policy Gradient for Semi-structured Mathematical Reasoning
Where alumni went46 listed, 41% to industry
Industry 9Academia 13
Most often to
- Meta
- Amazon
- Anthropic
Also to
- Stanford University
- USC
- Caltech
Apply to UCLAFall 2027
Computer Science PhD
- Round
- Single round (Dec 15)
- Opens
- —
- GRE
- optional
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- USUC Berkeley
Dan Klein
- Multilingual, translation & linguistics
- USMIT
Paul Liang
21 T1 - USUSC
Jieyu Zhao
- Interpretability, evaluation & safety of LMs
- USUCLA
Nanyun Peng
- Large language models
- CNTsinghua
Ning Ding
- Large language models
- GBImperial
Edoardo Maria Ponti
- Large language models
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
UCLAcloses