MIND: Benchmarking Memory Consistency and Action Control in World Models
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866917268005322752 |
|---|---|
| author | Ye, Yixuan Lu, Xuanyu Jiang, Yuxin Gu, Yuchao Zhao, Rui Liang, Qiwei Pan, Jiachun Zhang, Fengda Wu, Weijia Wang, Alex Jinpeng |
| author_facet | Ye, Yixuan Lu, Xuanyu Jiang, Yuxin Gu, Yuchao Zhao, Rui Liang, Qiwei Pan, Jiachun Zhang, Fengda Wu, Weijia Wang, Alex Jinpeng |
| contents | World models aim to understand, remember, and predict dynamic visual environments, yet a unified benchmark for evaluating their fundamental abilities remains lacking. To address this gap, we introduce MIND, the first open-domain closed-loop revisited benchmark for evaluating Memory consIstency and action coNtrol in worlD models. MIND contains 250 high-quality videos at 1080p and 24 FPS, including 100 (first-person) + 100 (third-person) video clips under a shared action space and 25 + 25 clips across varied action spaces covering eight diverse scenes. We design an efficient evaluation framework to measure two core abilities: memory consistency and action control, capturing temporal stability and contextual coherence across viewpoints. Furthermore, we design various action spaces, including different character movement speeds and camera rotation angles, to evaluate the action generalization capability across different action spaces under shared scenes. To facilitate future performance benchmarking on MIND, we introduce MIND-World, a novel interactive Video-to-World baseline. Extensive experiments demonstrate the completeness of MIND and reveal key challenges in current world models, including the difficulty of maintaining long-term memory consistency and generalizing across action spaces. Code: https://github.com/CSU-JPG/MIND. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_08025 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | MIND: Benchmarking Memory Consistency and Action Control in World Models Ye, Yixuan Lu, Xuanyu Jiang, Yuxin Gu, Yuchao Zhao, Rui Liang, Qiwei Pan, Jiachun Zhang, Fengda Wu, Weijia Wang, Alex Jinpeng Computer Vision and Pattern Recognition Artificial Intelligence World models aim to understand, remember, and predict dynamic visual environments, yet a unified benchmark for evaluating their fundamental abilities remains lacking. To address this gap, we introduce MIND, the first open-domain closed-loop revisited benchmark for evaluating Memory consIstency and action coNtrol in worlD models. MIND contains 250 high-quality videos at 1080p and 24 FPS, including 100 (first-person) + 100 (third-person) video clips under a shared action space and 25 + 25 clips across varied action spaces covering eight diverse scenes. We design an efficient evaluation framework to measure two core abilities: memory consistency and action control, capturing temporal stability and contextual coherence across viewpoints. Furthermore, we design various action spaces, including different character movement speeds and camera rotation angles, to evaluate the action generalization capability across different action spaces under shared scenes. To facilitate future performance benchmarking on MIND, we introduce MIND-World, a novel interactive Video-to-World baseline. Extensive experiments demonstrate the completeness of MIND and reveal key challenges in current world models, including the difficulty of maintaining long-term memory consistency and generalizing across action spaces. Code: https://github.com/CSU-JPG/MIND. |
| title | MIND: Benchmarking Memory Consistency and Action Control in World Models |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2602.08025 |