BHN: A Brain-like Heterogeneous Network
This work addresses representation learning in AI by mimicking brain-like unsupervised and cooperative processes, though it appears incremental as it builds on existing heterogeneous and self-supervised methods.
The paper tackles the challenge of learning distributed and global attention representations by proposing a brain-like heterogeneous network (BHN) that optimizes multiple objective functions in a minimax fashion, resulting in improved representations for image patches and video frames.
The human brain works in an unsupervised way, and more than one brain region is essential for lighting up intelligence. Inspired by this, we propose a brain-like heterogeneous network (BHN), which can cooperatively learn a lot of distributed representations and one global attention representation. By optimizing distributed, self-supervised, and gradient-isolated objective functions in a minimax fashion, our model improves its representations, which are generated from patches of pictures or frames of videos in experiments.