Multimodal Neurons in Artificial Neural Networks

1 code implementation • Distill 2021 • Gabriel Goh, Nick Cammarata, Chelsea Voss, Shan Carter, Michael Petrov, Ludwig Schubert, Alec Radford, Chris Olah

It’s the fact that you plug visual information into the rich tapestry of memory that brings it to life."

Weight Banding

no code implementations • Distill 2021 • Michael Petrov, Chelsea Voss, Ludwig Schubert, Nick Cammarata, Gabriel Goh, Chris Olah

Open up any ImageNet conv net and look at the weights in the last layer.

Visualizing Weights

no code implementations • Distill 2021 • Chelsea Voss, Nick Cammarata, Gabriel Goh, Michael Petrov, Ludwig Schubert, Ben Egan, Swee Kiat Lim, Chris Olah

Trying to understand artificial neural networks also has a lot in common with neuroscience, which tries to understand biological neural networks.

High-Low Frequency Detectors

no code implementations • Distill 2021 • Ludwig Schubert, Chelsea Voss, Nick Cammarata, Gabriel Goh, Chris Olah

Yet, when systematically characterizing the early layers of InceptionV1, we found a full fifteen neurons of mixed3a that appear to detect a high frequency pattern on one side, and a low frequency pattern on the other.

Vocal Bursts Intensity Prediction

Curve Detectors

no code implementations • Distill 2020 • Nick Cammarata, Gabriel Goh, Shan Carter, Ludwig Schubert, Michael Petrov, Chris Olah

Every vision model we've explored in detail contains neurons which detect curves.

Thread: Circuits

no code implementations • Distill 2020 • Nick Cammarata, Shan Carter, Gabriel Goh, Chris Olah, Michael Petrov, Ludwig Schubert, Chelsea Voss, Ben Egan, Swee Kiat Lim

To facilitate exploration of this direction, Distill is inviting a “thread” of short articles on circuits, interspersed with critical commentary by experts in adjacent fields.