AOEPT: Breaking the Implicit Modality-Reduction Bottleneck in Modality Missing Prompt Tuning

Jian Lang, Rongpei Hong, Ting Zhong, Fan Zhou^# (^# corresponding author)

International Conference on Machine Learning (ICML) 2026

Modality-missing prompt tuning for Multimodal Transformers can unintentionally restrict reasoning to observed-modality subspaces. AOEPT introduces modal-contextualized prompts that recover missing-modality information sources with minimal overhead...

[Code] [Paper]

LEAF: Towards Lightweight Explainable Hateful Video Detection via Self-Grounding CoT Guided Stage-Wise Distillation

Jian Lang, Rongpei Hong, Meihui Zhong, Kaiju Li, Ting Zhong, Qiang Gao, Fan Zhou^# (^# corresponding author)

Findings of the Association for Computational Linguistics (ACL Finding) 2026

Hateful video detection remains hard to trust because existing systems are often opaque, while LMM explanations are costly and biased toward benign predictions. LEAF distills self-grounding CoT explanations from LMMs into lightweight SMMs for accurate and interpretable HVD...

TAMEing Long Contexts in Personalization: Towards Training-Free and State-Aware MLLM Personalized Assistant

Rongpei Hong, Jian Lang, Ting Zhong, Fan Zhou

ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) 2026

Multimodal Large Language Model (MLLM) Personalization is a critical research problem that facilitates personalized dialogues with MLLMs targeting specific entities (known as personalized concepts). However, existing methods and benchmarks focus on ...

[Code] [Paper]

Borrowing Eyes for the Blind Spot: Overcoming Data Scarcity in Malicious Video Detection via Cross-Domain Retrieval Augmentation

Rongpei Hong*, Jian Lang*, Ting Zhong, Fan Zhou (* equal contribution)

International Conference on Computer Vision (ICCV) 2025

The rapid proliferation of online video-sharing platforms has accelerated the spread of malicious videos, creating an urgent need for robust detection methods. However, the performance and generalizability of existing detection approaches are severely limited ...

[Code] [Paper]

REDEEMing Modality Information Loss: Retrieval-Guided Conditional Generation for Severely Modality Missing Learning

Jian Lang, Rongpei Hong, Zhangtao Cheng, Ting Zhong, Yong Wang, Fan Zhou

ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) 2025

Traditional multimodal learning approaches often assume that all modalities are available during both the training and inference phases. However, this assumption is often impractical in real-world ...

[Code] [Paper]

Following Clues, Approaching the Truth: Explainable Micro-Video Rumor Detection via Chain-of-Thought Reasoning

Rongpei Hong, Jian Lang, Jin Xu, Zhangtao Cheng, Ting Zhong, Fan Zhou

The ACM Web Conference (WWW) 2025

The rapid spread of rumor content on online micro-video platforms poses significant threats to public health and safety. However, existing Micro-Video Rumor Detection (MVRD) methods are generally black-box, which lacks transparency and makes it difficult to understand the reasoning behind classification decisions...

[Code] [Paper]