Chapter
Discovering Functions within Neural Networks
37:30 - 44:37 (07:06)
Christopher Oleg's work on interpretability of neural networks allows for the identification and extraction of specific functions, such as detecting a wheel or mask for a car, within a larger network.
Clips
Neural networks are capable of discovering complex functions within programs, and can create human-like behaviors with the help of stochastic gradient descent algorithm.
37:30 - 44:37 (07:06)
Neural Networks
Summary
Neural networks are capable of discovering complex functions within programs, and can create human-like behaviors with the help of stochastic gradient descent algorithm. Furthermore, researchers have been able to identify individual detectors within these networks such as a detector for a car wheel.