News for hackers. Stuff that matters.
Re:Visual Q-Learning Demo in JavaScript from Karpathy (2013) by eigenvalue on Saturday November 25, 2023 @02:26PM (#38415883)
It was. Around the time this came out, something like half of the new ML papers were about reinforcement learning. The problem is that it’s incredibly slow and inefficient compared to any learning where you have access to a gradient and can use that to choose more targeted weight updates.

But there are certain applications where it’s the only good way of doing it (for example in games, where you don’t have access to a gradient over the space of how good a certain move is given the current game state, and it’s relatively quick and and efficient to simulate the evolution of the game).

Re:Visual Q-Learning Demo in JavaScript from Karpathy (2013) by rnimmer on Saturday November 25, 2023 @03:52PM (#38416579)
neuroevolution strategies are another approach to games, for what it's worth (since you said 'only').
Re:Visual Q-Learning Demo in JavaScript from Karpathy (2013) by Jagerbizzle on Saturday November 25, 2023 @05:44PM (#38417446)
For those of us like me who are unfamiliar, can you recommend any useful reading on the topic?
Re:Visual Q-Learning Demo in JavaScript from Karpathy (2013) by rnimmer on Sunday November 26, 2023 @11:39AM (#38422904)
Yes, take a look at Ken Stanley's web presence:

NEAT: https://www.cs.ucf.edu/~kstanley/neat.html

HyperNEAT: http://eplex.cs.ucf.edu/hyperNEATpage/HyperNEAT.html

There are explanations, links to the research papers, and links to implementations.

Here is a working example that runs in your browser using Javascript: https://liquidcarrot.io/example.flappy-bird/