Boyang Li
Associate Professor
Studies computational narrative intelligence, multimodal learning and machine learning.
Boyang Li's lab at NTU works on multimodal & grounded language, vision-language & multimodal and data management & big data. 53 papers since 2023, 18 at top venues, mostly CVPR and NeurIPS. 8 PhD students. 32 alumni, most often to Meta and Abridge AI.
At a glance
- papers, 2023–2026
- 14/ year
- at T1 venues
- 34%
- PhD students
- 8
- MS students
- 0
- alumni listed
- 32
Papers by yearfrom the lab homepage and DBLP, 2023–2026
- 13 papers2026 · T1 2 · collecting
- 17 papers2025 · T1 7
- 11 papers2024 · T1 5
- 12 papers2023 · T1 4
T1 venuesOtherVenues: CVPR, NeurIPS, ACL, ACM MM, NAACL
Research focuspaper titles that use each topic's words
- Multimodal & grounded language17 papers
- Vision-language & multimodal17 papers
- Data management & big data7 papers
- narrative0 papers
- multimodal learning0 papers
Paper list
From the lab homepage and DBLP.
202613 papers
- —Beyond Semantic Similarity: Appraisal-Guided Chain-of-Thought Reasoning and Retrieval for Multimodal Emotional Support Conversations
- —Co2In: a Bi-level Memory Incremental Learning Framework with Knowledge Encoding, Consolidation, and Integration
- Pattern RecognitionGELD: A unified neural model for efficiently solving traveling salesman problems across different scales (opens in a new tab)
- ICML WorkshopInjecting Knowledge from Social Science Journals to Improve Indonesian Cultural Understanding by LLMs
- ECCV (T1)InstrAct: Towards Action-Centric Understanding in Instructional Videos (opens in a new tab)
- WACVLearning to Animate Images from A Few Videos to Portray Delicate Human Actions (opens in a new tab)
- —Learning to Refuse: Refusal-Aware Reinforcement Fine-Tuning for Hard-Irrelevant Queries in Video Temporal Grounding (opens in a new tab)
- —On the Difficulty of Learning a Meta-network for Training Data Selection (opens in a new tab)
- —Paint Outside the Box: Synthesizing and Selecting Training Data for Visual Grounding
- —Revisiting the Role of Pretrained Weights in Model Merging: On Near-Optimality within the Core Subspace (opens in a new tab)
- ICML WorkshopSEA-MU: Cultural Meme Understanding Benchmark for Southeast Asia
- —The 32nd SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) (AI for Sciences Track)
202517 papers
- —A Visual Annotation-Free Method That Rivals Fully Supervised Methods for Grounded Multimodal Named Entity Recognition (opens in a new tab)
- —Black Swan: Abductive and Defeasible Video Reasoning in Unexpected Events (opens in a new tab)
- CVPR (T1)Black Swan: Abductive and Defeasible Video Reasoning in Unpredictable Events (opens in a new tab)
- ICML (T1)CAT Merging: A Training-Free Approach for Resolving Conflicts in Model Merging (opens in a new tab)
- —CharMoral: A Character Morality Dataset for Morally Dynamic Character Analysis in Long-Form Narratives
- IUIConversational Explanations: Discussing Explainable AI with Non-AI Experts (opens in a new tab)
- IEEE Trans. Multim.Copycat vs. Original: Multi-modal Pretraining and Variable Importance in Box-office Prediction (opens in a new tab)
- CVPR (T1)Enhancing Vision-Language Compositional Understanding with Multimodal Synthetic Data (opens in a new tab)
- Neural NetworksImproving generalization of neural vehicle routing problem solvers through the lens of model architecture (opens in a new tab)
- WACVLocal Masked Reconstruction for Efficient Self-Supervised Learning on High-resolution Images (opens in a new tab)
- ACL (T1)SPHERE: Unveiling Spatial Blind Spots in Vision-Language Models Through Hierarchical Evaluation (opens in a new tab)
- ICMESynopses of Movie Narratives: a Video-Language Dataset for Story Understanding (opens in a new tab)
202411 papers
- EMNLP (T1)Concept-skill Transferability-based Data Selection for Large Vision-Language Models (opens in a new tab)
- AAAI (T1)Distilling Autoregressive Models to Obtain High-Performance Non-Autoregressive Solvers for Vehicle Routing Problems with Faster Inference Speed (opens in a new tab)
- —Diversify, Rationalize, and Combine: Ensembling Multiple QA Strategies for Zero-shot Knowledge-based VQA (opens in a new tab)
- CVPR (T1)Emergent Open-Vocabulary Semantic Segmentation from Off-the-shelf Vision-Language Models (opens in a new tab)
- NAACL (T1)Event Causality Is Key to Computational Story Understanding (opens in a new tab)
- —Generating Synthetic Datasets for Few-shot Prompt Tuning
- Int. J. Hum. Comput. Interact.May I Ask a Follow-up Question? Understanding the Benefits of Conversations in Neural Network Explainability (opens in a new tab)
- Findings of EMNLPMultilingual Synopses of Movie Narratives: A Dataset for Vision-Language Story Understanding (opens in a new tab)
- IEEE Trans. Neural Networks Learn. Syst.Reinforcement Learning-Based Nonautoregressive Solver for Traveling Salesman Problems (opens in a new tab)
- —Training Data Recipe to Accelerate A* Search with Large Language Models (opens in a new tab)
- NAACL (T1)What Are We Measuring When We Evaluate Large Vision-Language Models? An Analysis of Latent Factors and Biases (opens in a new tab)
202312 papers
- ACM Comput. Surv.A Survey of Computer Vision Technologies In Urban and Controlled-environment Agriculture (opens in a new tab)
- CVPR (T1)From Images to Textual Prompts: Zero-shot Visual Question Answering with Frozen Large Language Models (opens in a new tab)
- —From Images to Textual Prompts: Zero-shot VQA with Frozen Large Language Models (opens in a new tab)
- Pattern Recognit. Lett.Improving Tail-Class Representation with Centroid Contrastive Learning (opens in a new tab)
- NeurIPS (T1)InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning (opens in a new tab)
- ACL (T1)Is GPT-3 a Good Data Annotator? (opens in a new tab)
- WSDM WorkshopMinimalist and High-performance Conversational Recommendation with Uncertainty Estimation for User Preference
- ICCV (T1)Mitigating and Evaluating Static Bias of Action Representations in the Background and the Foreground (opens in a new tab)
- —Movie Box Office Prediction With Self-Supervised and Visually Grounded Pretraining
- —The Annual Conference of the Association for Computational Linguistics (ACL)
- ACM MMTraining Multimedia Event Extraction With Generated Images and Captions (opens in a new tab)
- LGM3A@MMUnlocking Multimedia Capabilities of Gigantic Pretrained Language Models (opens in a new tab)
Where alumni went32 listed, 45% to industry
Industry 9Academia 11
Most often to
- Meta
- Abridge AI
- Amazon
Also to
- Beijing Jiaotong University
- Carnegie Mellon University
- ETH Zurich
Apply to NTUFall 2027
Not announced yet
Computing and Data Science PhD
- Round
- Main
- Opens
- —
- GRE
- not required
- English
- TOEFL/IELTS
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- USGeorgia Tech
Justin Romberg
- Learning theory & optimization
- GBOxford
Ian Horrocks
- Reasoning & neuro-symbolic AI
- USMichigan
Wei Hu
- Learning theory & optimization
- GBCambridge
Andreas Vlachos
- Multilingual, translation & linguistics
- USNYU
Rachel Greenstadt
- Trustworthy & safe ML
- CNSJTU
Yibo Yang
- Trustworthy & safe ML
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage and DBLP
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.