How Much Scale Do You Need to Scale Both Data and the Model Size to Solve Continuous Learning?

There are many scenarios for continuous learning. You can have them as this classical setting or more challenging task agnostic, like data changing and the system is not told when they changed. And it's quite an open question right now. To what extent, say, pre-training a very large-scale system can solve continual learning? In some most trivial, asymptotic sense, of course, if you scale the size of your model to infinity.

Transcript

Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!

Get the app