一作NeurIPS 2026Poster
Self-Improving World Modelling with Latent Actions
Yifu QIU · Zheng Zhao · Waylon Li · Yftah Ziser · Anna Korhonen · Shay Cohen · Edoardo Ponti
强化学习与后训练OpenReview
首次获取该候选人的完整档案及 StartLux 评估共 20 积分；每位用户单独解锁，之后查看免费。主动刷新 20 积分。
实时查找电子邮箱，费用不超过20积分；未找到邮箱不扣费。
Yifu QIU · Zheng Zhao · Waylon Li · Yftah Ziser · Anna Korhonen · Shay Cohen · Edoardo Ponti
Yifu QIU · Yftah Ziser · Anna Korhonen · Shay B Cohen · Edoardo Ponti
Can unified vision–language models (VLMs) perform forward dynamics prediction (FDP), i.e., predicting the future state (in image form) given the previous observation and an action (in language form)? We find that VLMs struggle to generate physically plausible transitions between frames from instr...