Research map / Video / Video-language understanding
Adaptive Hierarchical Graph Reasoning With Semantic Coherence for Video-and-Language Inference
Juncheng Li, Siliang Tang, Linchao Zhu, Haochen Shi, Xuanwen Huang, Fei Wu, Yi Yang, Yueting Zhuang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add hierarchical graph network (AHGN) over MTS (ablation) | MTS · VIOLIN test | accuracy: 67.60 -> 69.76 | Improved |
| add cross-modal OT coherence loss (ablation) | AHGN · VIOLIN test | accuracy: 69.76 -> 70.19 | Improved |
| add cross-level mutual information loss (ablation) | AHGN · VIOLIN test | accuracy: 69.76 -> 70.47 | Improved |
| add both coherence losses (ablation) | AHGN · VIOLIN test | accuracy: 69.76 -> 71.38 | Improved |
| use adaptive instead of fixed graph structure (ablation) | Backbone · VIOLIN test | accuracy: 64.65 -> 65.32 | Improved |
| use adaptive instead of fixed graph structure (ablation) | AHGN · VIOLIN test | accuracy: 68.91 -> 69.76 | Improved |
| add gated inter-modal message passing (GER) (ablation) | Backbone · VIOLIN test | accuracy: 65.32 -> 67.07 | Improved |
| add gated intra-modal message passing (GRA) (ablation) | Backbone · VIOLIN test | accuracy: 65.32 -> 66.74 | Improved |
| add temporal-level reasoning (ablation) | Backbone · VIOLIN test | accuracy: 65.32 -> 66.93 | Improved |
| add temporal-level reasoning to GER + GRA (ablation) | Backbone + GER + GRA · VIOLIN test | accuracy: 68.03 -> 69.76 | Improved |
| replace GloVe with BERT text features (ablation) | MTS · VIOLIN test | accuracy: 60.33 -> 67.60 | Improved |
| replace ResNet image features with detection features (ablation) | MTS · VIOLIN test | accuracy: 67.60 -> 67.84 | No clear effect |
Closest papers by meaning
- Visual Graph Arena: Evaluating Visual Conceptualization of Vision and Multimodal Large Language Models (ICML 2025 · similarity 0.87)
- VideoEspresso: A Large-Scale Chain-of-Thought Dataset for Fine-Grained Video Reasoning via Core Frame Selection (CVPR 2025 · similarity 0.87)
- OpenCoF: Learning to Reason Through Video Generation (NeurIPS 2026 · similarity 0.87)
- Scene Graph Thinking: Reinforcing Structured Visual Reasoning for Multimodal Large Language Models (ICML 2026 · similarity 0.87)
- Vgent: Graph-based Retrieval-Reasoning-Augmented Generation For Long Video Understanding (NeurIPS 2025 · similarity 0.87)
- Look, Remember and Reason: Grounded Reasoning in Videos with Language Models (ICLR 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
