 |
Zhiwei Jiang 蒋智威PhD, Assistant Professor（姑苏青年教授 & 特聘研究员）
State Key Laboratory for Novel Software Technology, School of Intelligence Software and Engineering, Nanjing University, China Email: jzw AT nju.edu.cn Office: Room 224, East of Nanyong Building, Suzhou Campus |
About Me
-
I am currently a tenure-track assistant professor in the School of Intelligence Software and Engineering, Nanjing University.
I received my Ph.D. degree from Nanjing University in 2018 under the supervision of Prof. Qing Gu, and received my B.Eng. degree from Central South University in 2011.
-
I am recruiting Ph.D., Master and Undergraduate students.
现招收博士、硕士以及南京大学本科生，对自然语言处理方向感兴趣的同学，欢迎随时邮件联系。
Research Interests
- Representation Learning and Understanding
- Text Representation from LLMs
- Latent Reasoning for LLMs/MLLMs
- Representation Probing and Analysis
- Multimodal Representation Alignment
- Controllable & Trustworthy Generation
- Inference-Time Steering and Test-Time Adaptation
- Hallucination Detection and Mitigation
- Controllable Image/Video Generation
- Concept Erasure and Model Watermarking
Selected Publications
-
ScaleErasure: Inference-Time Minimal Intervention for Precise Concept Erasure in Next-Scale Autoregressive Image Generation,International Conference on Machine Learning (, ) , 2026.
-
Multi-Label Test-Time Adaptation with Bayesian Conditional Priors,International Conference on Machine Learning (, ) , 2026.
-
Focusing Condition: Inference-Time Self-Contrastive Steering Elicits Better Conditional Text Embeddings in LLMs,Annual Meeting of the Association for Computational Linguistics (, ) , 2026.
-
AEA: Adaptive Expert Allocation Improves Sentence Embeddings from Mixture-of-Experts LLM,Annual Meeting of the Association for Computational Linguistics (, ) , 2026.
-
Rethinking BCE Loss for Multi-Label Image Recognition with Fine-Tuning,The IEEE/CVF Conference on Computer Vision and Pattern Recognition (, ) , 2026.
-
RegionMarker: A Region-Triggered Semantic Watermarking Framework for Embedding-as-a-Service Copyright Protection,AAAI Conference on Artificial Intelligence (, ) , 2026.
-
Steering When Necessary: Flexible Steering Large Language Models with Backtracking,Annual Conference on Neural Information Processing Systems (, ) , 2025.
-
Advanced Sign Language Video Generation with Compressed and Quantized Multi-Condition Tokenization,Annual Conference on Neural Information Processing Systems (, ) , 2025.
-
Token Prepending: A Training-Free Approach for Eliciting Better Sentence Embeddings from LLMs,Annual Meeting of the Association for Computational Linguistics (, ) , 2025.
-
Contrastive Prompting Enhances Sentence Embeddings in LLMs through Inference-Time Steering,Annual Meeting of the Association for Computational Linguistics (, ) , 2025.
-
Cross-Modal Prototype Augmentation and Dual-Grained Prompt Learning for Social Media Popularity Prediction,ACM International Conference on Multimedia (, ) , 2025.
-
Hierarchical Vision-Language Reasoning for Multimodal Multiple-Choice Question Answering,ACM International Conference on Multimedia (, ) , 2025.
-
Implicit Location-Caption Alignment via Complementary Masking for Weakly-Supervised Dense Video Captioning,AAAI Conference on Artificial Intelligence (, ) , 2025.
-
Ensembling Diffusion Models via Adaptive Feature Aggregation,International Conference on Learning Representations (, ) , 2025.
-
AP-Adapter: Improving Generalization of Automatic Prompts on Unseen Text-to-Image Diffusion Models,Annual Conference on Neural Information Processing Systems (, ) , 2024.
-
Short Video Ordering via Position Decoding and Successor Prediction,International Conference on Research and Development in Information Retrieval (, ) , 2024.
-
Unifying Token and Span Level Supervisions for Few-Shot Sequence Labeling,ACM Transactions on Information Systems (, ) , 42(1): 32:1-32:27, 2024.
-
Improving Domain Generalization for Prompt-Aware Essay Scoring via Disentangled Representation Learning,Annual Meeting of the Association for Computational Linguistics (, ) , 12456-12470, 2023.
-
Aggregating Multiple Heuristic Signals as Supervision for Unsupervised Automated Essay Scoring,Annual Meeting of the Association for Computational Linguistics (, ) , 13999-14013, 2023.
-
Unsupervised Readability Assessment via Learning from Weak Readability Signals,International Conference on Research and Development in Information Retrieval (, ) , 1324-1334, 2023.
-
Learning Robust Multi-Modal Representation for Multi-Label Emotion Recognition via Adversarial Masking and Perturbation,The Web Conference (, ) , 1510-1518, 2023.
-
Learning Event-Specific Localization Preferences for Audio-Visual Event Localization,ACM International Conference on Multimedia (, ) , 3446-3454, 2023.
-
Controlling Class Layout for Deep Ordinal Classification via Constrained Proxies Learning,AAAI Conference on Artificial Intelligence (, ) , 2483-2491, 2023.
-
A Consistent Dual-MRC Framework for Emotion-Cause Pair Extraction,ACM Transactions on Information Systems (, ) , 41(4): 105:1-105:27, 2023.