AITopics | layer tanh network

Collaborating Authors

layer tanh network

Information about AI from the News, Publications, and Conferences

Automatic Classification – Tagging and Summarization – Customizable Filtering and Analysis

If you are looking for an answer to the question What is Artificial Intelligence? and you only have a minute, then here's the definition the Association for the Advancement of Artificial Intelligence offers on its home page: "the scientific understanding of the mechanisms underlying thought and intelligent behavior and their embodiment in machines."

However, if you are fortunate enough to have more than a minute, then please get ready to embark upon an exciting journey exploring AI (but beware, it could last a lifetime) …

Reviews: Direct Feedback Alignment Provides Learning in Deep Neural Networks

Neural Information Processing SystemsJan-20-2025, 20:41:35 GMT

This paper demonstrates the surprising result that error information need not flow backward through several layers to enable good learning in a deep network. The scheme proposed in the paper is definitely biologically plausible (which is not to say that the brain necessarily uses this scheme). The paper could be strengthened through a more complete analysis of the learned filters in lower layers. Do they still visually resemble the filters produced through backpropagation? In particular, it seems important to understand why DFA attains the same training error as BP, but worse test error (particularly on CIFAR).

deep neural network, direct feedback alignment provide learning, layer tanh network, (4 more...)

Neural Information Processing Systems

Technology: Information Technology > Artificial Intelligence > Machine Learning > Neural Networks > Deep Learning (0.40)

Add feedback