研究方向：自然语言处理、多模态分析、认知计算

具体包括：情感分析、内容生成、大模型轻量化等

1. 具备推免资格，计算机相关专业，具有良好的语文、数学、英语、编程能力；

2. 发多少A不重要，希望学生有自驱力、有责任感、有同理心；

3. 期待未来彼此信任、并肩作战、共同成长；

081203-计算机应用技术
083900-网络空间安全

自然语言处理，多模态分析
信息内容安全

   

2022-10~现在, 中国科学院信息工程研究所, 研究员
2016-10~2022-10,中国科学院信息工程研究所, 副研究员
2014-07~2016-10,中国科学院信息工程研究所, 助理研究员

2024-12-28-今,中文信息学会, 计算语言学专委会委员
2024-11-30-今,中文信息学会, 大模型与生成专委会委员
2024-10-18-今,中文信息学会, 信息检索专委会委员
2024-07-01-今,中文信息学会, 情感计算专委会委员
2024-05-17-今,中国指挥与控制学会, 大模型与决策智能专委会委员
2023-07-01-今,中文信息学会, 青年工作委员会委员
2022-06-30-2027-07-30,中国指挥与控制学会, 信息融合专委会委员

信息检索导论
文献阅读
深度学习与自然语言处理

   

（1） 细粒度比拟句情绪分析, 一等奖, 其他, 2025
（2） 结合用户画像信息中文情绪分类评测, 一等奖, 其他, 2024
（3） 中国科学院优秀导师, 院级, 2024
（4） 科学技术进步奖, 一等奖, 部委级, 2022
（5） 全国第二届军事智能-机器阅读挑战赛, 一等奖, 国家级, 2019
（6） F5000优秀论文, 部委级, 2019
（7） 全国第一届军事智能-机器阅读挑战赛, 一等奖, 国家级, 2018

1. VecInfer: Efficient LLM Inference with Low-Bit KV Cache via Outlier-Suppressed Vector Quantization. ACL, 2026 (CCF A)
2. LWGR: Lagrangian-Constrained Personalized World Knowledge for Generative Recommendation. SIGIR, 2026 (CCF A)
3. ExDR: Explanation-driven Dynamic Retrieval Enhancement for Multimodal Fake News Detection. SIGIR, 2026 (CCF A)
4. Dynamic Early Exit in Reasoning Models. ICLR, 2026 (CCF A)
5. Blink: Dynamic Visual Token Resolution for Enhanced Multimodal Understanding. CVPR, 2026 (CCF A)
6. Test-time Prompt Intervention. AAAI, 2026 (CCF A)
7. Breaking the Trade-Off Between Faithfulness and Expressiveness for Large Language Models. AAAI, 2026 (CCF A)
8. System 1&2 Synergy via Dynamic Model Interpolation. IJCAI, 2026
9. Online Self-Calibration Against Hallucination in Vision-Language Models. IJCAI, 2026
10. Causal Path Alignment: Anchoring the Optimization Trajectory for Controllable In-Parameter Knowledge Editing. IJCAI, 2026
11. Absorbing Gradient Conflicts: Modeling Semantic Variance via Kent Distributions for Cross-Modal Hashing. IJCAI, 2026
12. DualECot: Toward Explainable Emotion Causality Analysis in Conversations. IEEE Transactions on Audio, Speech and Language Processing, 2026 (CCF B)
13. NeedleInATable: Exploring Long-Context Capability of Large Language Models towards Long-Structured Tables. NeurIPS, 2025 (CCF A)
14. Adapt Once, Thrive with Updates: Transferable Parameter-Efficient Fine-Tuning on Evolving Base Models. ACL, 2025 (CCF A)
15. BeamLoRA: Beam-Constraint Low-Rank Adaptation. ACL, 2025 (CCF A)
16. DIVE into MoE: Diversity-Enhanced Reconstruction of Large Language Models from Dense into Mixture-of-Experts. ACL, 2025 (CCF A)
17. PIG: Privacy Jailbreak Attack on LLMs via Gradient-based Iterative In-Context Optimization. ACL, 2025 (CCF A)
18. TailorKV: A Hybrid Framework for Long-Context Inference via Tailored KV Cache Optimization. Findings of ACL, 2025
19. TableDreamer: Progressive and Weakness-guided Data Synthesis from Scratch for Table Instruction Tuning. Findings of ACL, 2025
20. Incorporating Communication Style and Interaction of Speakers for Sarcasm Explanation in Dialogue. SIGIR, 2025 (CCF A)
21. Two-stage Adversarial Training for Deep Hashing via Representation Distillation. SIGIR, 2025 (CCF A)
22. AutoPrompt: Automated Red-Teaming of Text-to-Image Models via LLM-Diven Adversarial Prompts. ICCV, 2025 (CCF A)
23. Mitigating the Evolving Semantic Entanglement in Continual Learning of Vision-Language Models. ACM MM, 2025 (CCF A)
24. Categorical Attention: Fine-grained Language-guided Noise Filtering Network for Occluded Person Re-Identification. IJCAI, 2025 (CCF A)
25. Bridging the Gap: Aligning Language Model Generation with Structured Information Extraction via Controllable State Transition. WWW, 2025 (CCF A)
26. Trust-GRS: A Trustworthy Training Framework for Graph Neural Network Based Recommender Systems against Shilling Attacks. AAAI, 2025 (CCF A)
27. Relation Also Knows: Rethinking the Recall and Editing of Factual Associations in Auto-Regressive Transformer Language Models. AAAI, 2025 (CCF A)
28. Multimodal Hypothetical Summary for Retrieval-based Multi-image Question Answering. AAAI, 2025 (CCF A)
29. Weights-Rotated Preference Optimization for Large Language Models. EMNLP, 2025
30. CBP-Tuning: Efficient Local Customization for Black-box Large Language Models. EMNLP, 2025
31. Sibyl: Empowering Empathetic Dialogue Generation in Large Language Models via Sensible and Visionary Commonsense Inference. COLING, 2025
32. Towards One-to-Many Visual Question Answering. findings of EMNLP, 2024
33. Towards Flexible Evaluation for Generative Visual Question Answering. ACM MM, 2024 （CCF A）
34. Disrupting Diffusion: Token-Level Attention Erasure Attack against Diffusion-based Customization. ACM MM, 2024（CCF A）
35. Prediction Exposes Your Face: Black-box Model Inversion via Prediction Alignment. ECCV, 2024
36. Multimodal Table Understanding. ACL, 2024 （CCF A）
37. Light-PEFT: Lightening Parameter-Efficient Fine-Tuning via Early Pruning. findings of ACL, 2024 
38. Pruning Large Language Models to Intra-module Low-rank Structure with Transitional Activations. findings of ACL, 2024
39. A Pedestrian is Worth One Prompt: Towards Language Guidance Person Re-Identification. CVPR, 2024 （CCF A）
40. Object Attribute Matters in Visual Question Answering. AAAI, 2024 （CCF A）
41. SD4Privacy: Exploiting Stable Diffusion for Protecting Facial Privacy. ICME, 2024
42. Exploiting Vision-Language Model for Visible-Infrared Person Re-identification via Textual Modality Alignment. ICME, 2024
43. Privacy-Preserving Replay and Adaptive Relation Distillation for Camera Incremental Person Re-Identification. ICME. ICME, 2024
44. Multi-level Adaptive Contrastive Learning for Knowledge Internalization in Dialogue Generation. EMNLP, 2023
45. Compressing and Debiasing Vision-Language Pre-Trained Models for Visual Question Answering. EMNLP, 2023
46. An Empirical Study of Instruction-tuning Large Language Models in Chinese.  findings of EMNLP, 2023
47. Chain-of-Thought Reasoning in Tabular Language Models. findings of EMNLP, 2023
48. IM-TQA: A Chinese Table Question Answering Dataset with Implicit and Multi-type Table Structures. ACL, 2023 （CCF A）
49. Combo of Thinking and Observing for Outside-Knowledge VQA. ACL, 2023 （CCF A）
50. A Gradient Control Method for Backdoor Attacks on Parameter-Efficient Tuning. ACL, 2023 （CCF A）
51. Question-Interlocutor Scope Realized Graph Modeling over Key Utterances for Dialogue Reading Comprehension. Findings of ACL, 2023
52. Maximum Entropy Loss, the Silver Bullet Targeting Backdoor Attacks in Pre-trained Language Models. Findings of ACL, 2023
53. Divide, Conquer, and Combine: Mixture of Semantic-Independent Experts for Zero-Shot Dialogue State Tracking. ACL, 2023 （CCF A）
54. DiffusEmp: A Diffusion Model-Based Framework with Multi-Grained Control for Empathetic Response Generation. ACL, 2023 （CCF A）
55. Retrieve-and-Sample: Document-level Event Argument Extraction via Hybrid Retrieval Augmentation. ACL, 2023 （CCF A）
56. Explainable Hyperbolic Temporal Point Process for User-Item Interaction Sequence Generation. TOIS, 2023 （CCF A）
57. Multi-Aspect Heterogeneous Graph Augmentation. WWW, 2023 （CCF A）
58. Towards Polymorphic Adversarial Examples Generation for Short Text. ICASSP, 2023
59. SERI: SkEtching-Reasoning-Integrating Progressive Workflow for Empathetic Response Generation. ICASSP, 2023
60. Explainable Hyperbolic Temporal Point Process for User-Item Interaction Sequence Generation. ACM Transactions on Information Systems, 2022 （CCF A）
61. A Win-win Deal: Towards Sparse and Robust Pre-trained Language Models. NeurIPS, 2022  （CCF A）
62. Detach and Attach: Stylized Image Captioning without Paired Stylized Dataset. ACM MM, 2022 （CCF A）
63. Neutral Utterances are Also Causes: Enhancing Conversational Causal Emotion Entailment with Social Commonsense Knowledge. IJCAI, 2022 （CCF A）
64. Neural Label Search for Zero-Shot Multi-Lingual Extractive Summarization. ACL, 2022 （CCF A）
65. COST-EFF: Collaborative Optimization of Spatial and Temporal Efficiency with Slenderized Multi-exit Language Models. EMNLP, 2022
66. Language Prior Is Not the Only Shortcut: A Benchmark for Shortcut Learning in VQA. findings of EMNLP, 2022
67. Empathetic Dialogue Generation via Sensitive Emotion Recognition and Sensible Knowledge Selection. findings of EMNLP, 2022
68. Towards Robust Visual Question Answering: Making the Most of Biased Samples via Contrastive Learning. findings of EMNLP, 2022
69. Target Really Matters: Target-aware Contrastive Learning and Consistency Regularization for Few-shot Stance Detection. COLING, 2022
70. TAKE: Topic-shift Aware Knowledge sElection for Dialogue Generation. COLING, 2022
71. Slot Dependency Modeling for Zero-Shot Cross-Domain Dialogue State Tracking. COLING, 2022
72. CLIO: Role-interactive Multi-event Head Attention Network for Document-level Event Extraction. COLING, 2022
73. Learning to Win Lottery Tickets in BERT Transfer via Task-agnostic Mask Training. NAACL, 2022
74. Cross-Target Stance Detection via Refined Meta-Learning. ICASSP, 2022
75. Connecting Targets via Latent Topics and Contrastive Learning: A Unified Framework for Robust Zero-shot and Few-shot Stance Detection. ICASSP, 2022
76. Addressing Extraction and Generation Separately:  Key phrase Prediction with Pre-trained Language Models. IEEE Transactions on Audio, Speech and Language Processing, 2021
77. Sarcasm Detection with Commonsense Knowledge. IEEE Transactions on Audio, Speech and Language Processing, 2021
78. Past, Present, and Future: Conversational Emotion Recognition through Structural Modeling of Psychological Knowledge. Findings of EMNLP, 2021
79. Check It Again: Progressive Visual Question Answering via Visual Entailment. ACL, 2021 （CCF A）
80. Marginal Utility Diminishes: Exploring the Minimum Knowledge for BERT Knowledge Distillation. ACL, 2021 （CCF A）
81. Enhanching Zero-shot and Few-shot Stance Detection with Commonsense Knowledge Graph. Findings of  ACL, 2021 （CCF A）
82. ROSITA: Refined BERT Compression with Integrated Technique. AAAI, 2021 （CCF A）
83. Learning Class-Transductive Intent Representations for Zero-shot IntentDetection. IJCAI, 2021 （CCF A）
84. MACROBERT: Maximizing Certified Region of BERT to Adversarial Word Substitutions. DASFAA, 2021
85. Modeling Intra and Inter-modality Incongruity for Multi-Modal Sarcasm Detection. Findings of EMNLP, 2020
86. Joint Character-Level Word Embedding and Adversarial Stability Training to Defend Adversarial Text. AAAI, 2020 （CCF A）
87. Modeling the Incongruity between Sentence Snippets for Sarcasm Detection. ECAI, 2020
88. Generating Paraphrase with Topic as Prior Knowledge. CIKM, 2019
89. Ranking and Sampling in Open-Domain Question Answering. EMNLP, 2019
90. Learning Sentiment-Specific Word Embedding via Global Sentiment Representation. AAAI, 2018 （CCF A）
91. Deep Uniqueness-aware Hashing for Fine-grained Multi-label Image Retrieval. ICASSP, 2018
92. Deep Supervised Hashing for Multi-Label and Large-Scale Image Retrieval. ICMR, 2017
93. An Unsupervised Cross-Lingual Topic Model Framework for Sentiment Classification, IEEE/ACM Transactions on Audio, Speech and Language Processing, 2016
94. A Word Vector and Matrix Factorization Based Method for Opinion Lexicon Extraction. WWW, 2015
95. A Cross-Lingual Joint  Aspect/Sentiment Model for Sentiment Analysis. CIKM, 2014
96. Towards jointly extracting aspects and aspect-specific sentiment knowledge. CIKM, 2012
97. Find me opinion sources in blogosphere: a unified framework for opinionated blog feed retrieval. WSDM, 2012
98. Effective and efficient?bilingual sentiment lexicon extraction using collocation alignment. CIKM, 2012
99. Language-independent sentiment classification using three common words. CIKM, 2011
100. 文本后门攻击与防御综述. 计算机研究与发展，2023 （CCF A）
101. 智能回复系统研究综述. 信息安全学报，2020
102. 文本情绪分析综述. 计算机研究与发展，2018 （CCF A）
103. 一种面向大规模社会信息网络的多层社区发现算法. 计算机学报，2016 （CCF A）

（1） 情感分析进阶, Advanced Sentiment Analysis, 机械工业出版社, 2023-03, 第 1 作者
（2） 文本情感分析, Text Sentiment Analysis, 清华大学出版社, 2019-12, 第 1 作者

（ 1 ） 观点大模型, 负责人, 国家任务, 2024-03--2024-12
（ 2 ） 行业大模型, 负责人, 国家任务, 2024-01--2025-12
（ 3 ） 多模态情绪分析, 负责人, 国家任务, 2023-11--2025-11
（ 4 ） 大模型轻量化, 负责人, 国家任务, 2023-06--2025-06
（ 5 ） 大模型推理技术, 负责人, 境内委托项目, 2023-05--2024-05
（ 6 ） 大模型行业助理, 负责人, 国家任务, 2023-01--2024-12
（ 7 ） 情绪感知与知识增强的文本生成技术研究, 负责人, 国家任务, 2020-01--2023-12
（ 8 ） 识别与研判技术（二期）, 负责人, 国家任务, 2019-10--2020-09
（ 9 ） 网络信息平台（二期）, 参与, 国家任务, 2019-09--2021-09
（ 10 ） 态势感知技术研究, 负责人, 国家任务, 2019-02--2022-01
（ 11 ） 识别与消解技术（一期）, 负责人, 国家任务, 2018-10--2019-09
（ 12 ） 网络信息平台（一期）, 参与, 国家任务, 2016-05--2017-11
（ 13 ） 短文本情感分析关键技术研究, 负责人, 国家任务, 2016-01--2018-12

许同学 2017级硕士生 腾讯微信

耿同学 2017级硕士生 国家公务员

吴同学 2017级硕士生 百度搜索

刘同学 2017级直博生 阿里蚂蚁金服

佀同学 2019级硕转博 华为天才少年 

李同学 2019级硕转博 腾讯微信模式识别 

刘同学 2019级硕士生 北京大学 

李同学 2019级硕转博 香港中文大学 

谭同学 2020级硕士生 米哈游 

温同学 2020级硕士生 工商银行

梁同学 2020级硕士生 移动研究院 

王同学 2021级硕士生 华为鸿蒙

王同学 2022级硕士生 腾讯混元大模型

陈同学 2022级硕士生 字节跳动

申同学 2022级硕士生 小米mimo

孙同学 2023级硕士生 北京市公务员

郑同学 2021级博士生 蚂蚁百灵大模型