arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2602.18690q-bio.NCcs.CVcs.LG
收藏

神经场作为世界模型

Neural Fields as World Models

Luddy School of Informatics, Computing, and Engineering, Indiana University, Bloomington(信息学、计算与工程学院,印第安纳大学,布卢明顿);Cognitive Science Program, Indiana University, Bloomington(认知科学项目,印第安纳大学,布卢明顿)

Joshua Nunley

AI总结:

提出同构世界模型,利用运动门控神经场在空间图中进行物理预测,实现离线任务学习和身体相关表征。

摘要:

人类可以在离线状态下预演可能的未来,例如在心理练习和可能的梦境中,这表明世界模型可能支持远离环境的学习。标准的机器学习世界模型将视觉输入压缩为潜在向量,丢弃了感觉皮层的空间结构特征。我们提出了同构世界模型:一种保持感觉拓扑结构的架构,使得物理预测成为几何传播而非抽象状态转换。我们通过运动门控神经场实现这一想法,其中活动通过局部侧向连接演化,运动命令乘性地调制特定通道。在三个实验中,相同的架构学习了无“瞬移”的弹道预测,通过将任务误差通过冻结的学习世界模型传播,改进了离线接球策略,并在没有身体标签的情况下发展出身体选择性的运动通道。这些结果提供了初步证据,表明物理预测、离线任务学习和身体相关表征共享一个共同的计算基础:空间地图内的动作条件预测。

英文摘要:

Humans rehearse possible futures offline, as in mental practice and perhaps dreaming, suggesting that world models may support task learning away from the environment. Standard machine learning world models compress visual input into latent vectors, discarding the spatial structure that characterizes sensory cortex. We propose isomorphic world models: architectures that preserve sensory topology, so physics prediction becomes geometric propagation rather than abstract state transition. We implement this idea with motor-gated neural fields, where activity evolves through local lateral connectivity and motor commands multiplicatively modulate specific channels. Across three experiments, the same architecture learns ballistic prediction without ``teleporting,'' improves a catching policy offline by propagating task error through a frozen learned world model, and develops body-selective motor channels without body labels. These results provide preliminary evidence that physical prediction, offline task learning, and body-linked representation share a common computational substrate: action-conditional prediction within a spatial map.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑