TalkRL: The Reinforcement Learning Podcast cover image

Max Schwarzer

TalkRL: The Reinforcement Learning Podcast

CHAPTER

The Evolution of DQN Networks

The original DQN network was pretty simple. CNN and fully connected. And then what BBF has, and pretty much every paper after that, used almost exactly the same network for Atari 100K. The overall capacity you get is dramatically higher. So we're not using anything like a transformer. It's pretty much just, like I said, it's just a ResNet.

00:00
Transcript
Play full episode

Remember Everything You Learn from Podcasts

Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
App store bannerPlay store banner