Coordination Among Neural Modules Through a Shared Global Workspace
Mila（米拉人工智能研究所）；University of Montreal（蒙特利尔大学）；Google Deepmind（谷歌DeepMind）；Max Planck Institute（马克斯·普朗克研究所）；Google Research（谷歌研究院）；Microsoft Research（微软研究院）；Indian Institute of Technology, Delhi（印度理工学院德里分校）
Anirudh Goyal, Aniket Didolkar, Alex Lamb, Kartikeya Badola, Nan Rosemary Ke, Nasim Rahaman, Jonathan Binas, Charles Blundell, Michael Mozer, Yoshua Bengio
Deep learning has seen a movement away from representing examples with a monolithic hidden state towards a richly structured state. For example, Transformers segment by position, and object-centric architectures decompose images into entities. In all these architectures, interactions between different elements are modeled via pairwise interactions: Transformers make use of self-attention to incorporate information from other positions; object-centric architectures make use of graph neural networks to model interactions among entities. However, pairwise interactions may not achieve global coordination or a coherent, integrated representation that can be used for downstream tasks. In cognitive science, a global workspace architecture has been proposed in which functionally specialized components share information through a common, bandwidth-limited communication channel. We explore the use of such a communication channel in the context of deep learning for modeling the structure of complex environments. The proposed method includes a shared workspace through which communication among different specialist modules takes place but due to limits on the communication bandwidth, specialist modules must compete for access. We show that capacity limitations have a rational basis in that (1) they encourage specialization and compositionality and (2) they facilitate the synchronization of otherwise independent specialists.
Semi-supervised Multimodal Representation Learning through a Global Workspace
CerCo, CNRS UMR 5549, Université de Toulouse and ANITI, Artificial and Natural Intelligence Toulouse Institute（CerCo、CNRS UMR 5549、图卢兹大学和ANITI人工智能与自然智能图卢兹研究所）
Efficient Coordination with the System-Level Shared State: An Embodied-AI Native Modular Framework
School of Science and Engineering, The Chinese University of Hong Kong, Shenzhen, 2001 Longxiang Boulevard, Shenzhen, China（香港中文大学（深圳）科学与工程学院）；School of Artificial Intelligence, The Chinese University of Hong Kong, Shenzhen, 2001 Longxiang Boulevard, Shenzhen, China（香港中文大学（深圳）人工智能学院）；Shenzhen Institute of Artificial Intelligence（深圳人工智能与机器人研究院）；International Digital Economy Academy, Shenzhen-Hong Kong Collaborative Innovation Center, Shenzhen, China（国际数字经济学院）；School of Data Science, The Chinese University of Hong Kong, Shenzhen, 2001 Longxiang Boulevard, Shenzhen, China（香港中文大学（深圳）数据科学学院）；The School of Computer Science, The University of Sydney, Sydney, Australia（悉尼大学计算机科学学院）
Multi-level Collaborative Distillation Meets Global Workspace Model: A Unified Framework for OCIL
School of Computer Science, Northwestern Polytechnical University（西北工业大学计算机学院）；School of Telecommunications Engineering, Xidian University（西安电子科技大学电信工程学院）
Celebrating Diversity in Shared Multi-Agent Reinforcement Learning
Tsinghua University（清华大学）
最近，深度多智能体强化学习（MARL）已经显示出解决复杂协作任务的前景。它的成功部分是因为代理之间的参数共享。然而，这种共享可能会导致代理人的行为相似，并限制他们的协调能力。在本文中，我们的目标是在共享多智能体强化学习的优化和表示中引入多样性。具体来说，我们提出了一个信息理论正则化，以最大限度地提高代理的身份和他们的轨迹之间的互信息，鼓励广泛的探索和多样化的个性化行为。在表示中，我们将代理特定的模块中的共享神经网络架构，这是正规化的L1范数，以促进学习共享的代理，同时保持必要的多样性。实证结果表明，我们的方法在Google Research Football和超级硬的星际争霸II微观管理任务上实现了最先进的性能。
