arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2007.02686cs.NEcs.LG
收藏

随机网络中通过Hebbian可塑性的元学习

Meta-Learning through Hebbian Plasticity in Random Networks

IT University of Copenhagen(哥本哈根信息技术大学)

Elias Najarro, Sebastian Risi

摘要:

终身学习和适应性是生物制剂的两个定义方面。现代强化学习(RL)方法在解决复杂任务方面取得了显着进展,但一旦训练结束,找到的解决方案通常是静态的,无法适应新的信息或扰动。虽然生物大脑如何从经验中有效地学习和适应仍然没有完全理解,但人们认为突触可塑性在这一过程中起着重要作用。受这种生物学机制的启发,我们提出了一种搜索方法,而不是直接优化神经网络的权重参数,只搜索突触特定的赫布学习规则,使网络在代理的生命周期内不断自组织其权重。我们展示了我们的方法在几个强化学习任务与不同的感觉方式和超过450 K可训练的可塑性参数。我们发现,从完全随机的权重开始,所发现的赫布规则使智能体能够在动态的2D像素环境中导航;同样,它们允许模拟的3D四足机器人学习如何行走,同时适应训练过程中没有看到的形态学损伤,并且在不到100个时间步的情况下没有任何明确的奖励或错误信号。代码可在https://github.com/enajx/HebbianMetaLearning上获得。

英文摘要:

Lifelong learning and adaptability are two defining aspects of biological agents. Modern reinforcement learning (RL) approaches have shown significant progress in solving complex tasks, however once training is concluded, the found solutions are typically static and incapable of adapting to new information or perturbations. While it is still not completely understood how biological brains learn and adapt so efficiently from experience, it is believed that synaptic plasticity plays a prominent role in this process. Inspired by this biological mechanism, we propose a search method that, instead of optimizing the weight parameters of neural networks directly, only searches for synapse-specific Hebbian learning rules that allow the network to continuously self-organize its weights during the lifetime of the agent. We demonstrate our approach on several reinforcement learning tasks with different sensory modalities and more than 450K trainable plasticity parameters. We find that starting from completely random weights, the discovered Hebbian rules enable an agent to navigate a dynamical 2D-pixel environment; likewise they allow a simulated 3D quadrupedal robot to learn how to walk while adapting to morphological damage not seen during training and in the absence of any explicit reward or error signal in less than 100 timesteps. Code is available at https://github.com/enajx/HebbianMetaLearning.

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。

↑