Research map / Vision-language and multimodal / Robot action and navigation
TechCoach: Towards Technical-Point-Aware Descriptive Action Coaching
Yuan-Ming Li, An-Lan Wang, Ling-An Zeng, Kun-Yu Lin, Yu-Ming Tang, Weishi Zheng
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 TechPoint-level alignment loss L_Align (ablation) | TechCoach · EE4D-DescCoach | Spearman rho: 70.2 -> 72.5 | Improved |
| add TechPoint-level alignment loss L_Align (ablation) | TechCoach · EE4D-DescCoach | CIDEr: 14.3 -> 14.6 | Improved |
| combine L_MSE and L_MTM without L_Align (ablation) | TechCoach · EE4D-DescCoach | Spearman rho / BERTScore: rho 70.8 -> 70.2, BERT 66.8 -> 65.7 | Hurt |
| add TechPoint inputs and designs except L_Align (to SwinBERT) (ablation) | SwinBERT · EE4D-DescCoach | GPT-M score: +3.04 | Improved |
| add TechPoint inputs and designs except L_Align (to SwinBERT) (ablation) | SwinBERT · EE4D-DescCoach | GPT-Q score: +1.45 | Improved |
| add L_Align on top of other designs (ablation) | TechCoach · EE4D-DescCoach | GPT-Q score: +3.88 | Improved |
| use TechPoint-level alignment instead of binary classification (TP-CLS) (ablation) | TechCoach · EE4D-DescCoach | CIDEr / BERTScore: CIDEr 13.7 -> 14.6, BERT 64.7 -> 66.9 | Improved |
| use TP-CLS binary classification instead of no alignment (ablation) | TechCoach · EE4D-DescCoach | Spearman rho: 70.2 -> 72.3 | Improved |
| scale InternVL2 from 8B to 76B | InternVL2 · EE4D-DescCoach | GPT-Q score: 29.81 -> 32.60 | Improved |
| scale InternVL2 from 8B to 76B | InternVL2 · EE4D-DescCoach | BLEU: 17.88 -> 10.81 | Hurt |
Closest papers by meaning
- Watch and Learn: Learning to Use Computers from Online Videos (CVPR 2026 · similarity 0.83)
- ExAct: A Video-Language Benchmark for Expert Action Analysis (NeurIPS 2025 · similarity 0.82)
- ExpertAF: Expert Actionable Feedback from Video (CVPR 2025 · similarity 0.82)
- FineParser: A Fine-grained Spatio-temporal Action Parser for Human-centric Action Quality Assessment (CVPR 2024 · similarity 0.82)
- Procedure Knowledge Decoupled Distillation Strategy for Procedure Planning in Instructional Videos (AAAI 2025 · similarity 0.82)
- CigTime: Corrective Instruction Generation Through Inverse Motion Editing (NeurIPS 2024 · similarity 0.82)
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.
