Research map / Vision-language and multimodal / Multimodal LLMs
EgoPlan-Bench: Benchmarking Multimodal Large Language Models for Human-Level Planning
Yuehua Chen, Yuying Ge, Yixiao Ge, Mingyu Ding, Bohao Li, Rui Wang, Ruifeng Xu, Ying Shan, Xihui Liu
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 LoRA tuning (ablation) | Enhanced Video-LLaMA · EgoPlan-Val (all) | accuracy (%): 50.55 -> 51.83 | Improved |
| add contrastive loss (ablation) | Enhanced Video-LLaMA · EgoPlan-Val (all) | accuracy (%): 48.94 -> 51.83 | Improved |
| add diverse instructions (ablation) | Enhanced Video-LLaMA · EgoPlan-Val (all) | accuracy (%): 50.97 -> 51.83 | Improved |
| add action recognition auxiliary data (ablation) | Enhanced Video-LLaMA · EgoPlan-Val (all) | accuracy (%): 50.13 -> 51.83 | Improved |
| add action recognition auxiliary data (ablation) | Enhanced Video-LLaMA · EgoPlan-Val (out-of-domain) | accuracy (%): 44.31 -> 44.42 | No clear effect |
| add action recognition auxiliary data (ablation) | Enhanced Video-LLaMA · EgoPlan-Val (in-domain) | accuracy (%): 52.34 -> 54.65 | Improved |
| add contrastive loss (ablation) | Enhanced Video-LLaMA · EgoPlan-Val (out-of-domain) | accuracy (%): 40.52 -> 44.42 | Improved |
| add LoRA tuning (ablation) | Enhanced Video-LLaMA · EgoPlan-Val (out-of-domain) | accuracy (%): 42.58 -> 44.42 | Improved |
| instruction-tune Video-LLaMA on EgoPlan-IT | Video-LLaMA · EgoPlan-Val (all) | accuracy (%): 28.58 -> 51.83 | Improved |
| instruction-tune Video-LLaMA on EgoPlan-IT | Video-LLaMA · EgoPlan-Val (out-of-domain) | accuracy (%): 30.44 -> 44.42 | Improved |
| higher goal-answer semantic similarity (easier questions) | top three MLLMs · EgoPlan-Val | accuracy (%) | Improved |
| longer task progress (more historical actions) | MLLMs · EgoPlan-Val | accuracy (%) | Hurt |
Closest papers by meaning
- On Path to Multimodal Generalist: General-Level and General-Bench (ICML 2025 · similarity 0.91)
- LLMs Can Plan Only If We Tell Them (ICLR 2025 · similarity 0.88)
- EmbodiedBench: Comprehensive Benchmarking Multi-modal Large Language Models for Vision-Driven Embodied Agents (ICML 2025 · similarity 0.88)
- MDK12-Bench: A Multi-Discipline Benchmark for Evaluating Reasoning in Multimodal Large Language Models (AAAI 2026 · similarity 0.88)
- EgoBench: An Interactive Egocentric Multimodal Benchmark for Tool-Using Agents (NeurIPS 2026 · similarity 0.88)
- HumanSense: From Multimodal Perception to Empathetic Context-Aware Responses Through Reasoning MLLMs (AAAI 2026 · similarity 0.88)
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.
