Forgetting learning is an incremental learning rule in associative memories. With it, the recent learning items can be encoded and the old learning items will be forgotten. In this paper, the storage behavior of bidirectional associative memory (BAM) under the forgetting learning is first analyzed. That is, "Can the most recent k learning item be stored as a fixed point?". We then discuss the way to choose the forgetting constant in the forgetting learning such that BAM can correctly store the most recent learning items as many as possible. The magnitude of the weights under the forgetting learning is also discussed. lastly, we investigate the error correction capability of BAM under forgetting learning. Simulations are provided to verify the theoretical analysis.
展开▼
