Curiosity-driven Exploration by Self-supervised Prediction · HackerPapers discussions

Curiosity-driven Exploration by Self-supervised Prediction — mentioned by andy12_

Tue, 10 Mar 2026 14:37:31 GMT

So, I have been thinking about this for a little while. Image a model f that takes a world x and makes a prediciton y. At a high-level, a traditional supervised model is trained like this f(x)=y' => loss(y',y) => how good was my prediction? Train f through backprop with that error. While a model trained with reinforcement learning is more similar to this. Where m(y) is the resulting world state of taking an action y the model predicted. f(x)=y' => m(y')=z => reward(z) => how good was the state I was in based on my actions? Train f with an algorithm like REINFORCE with the reward, as the world…

Paper and discussion history. A mention is not an endorsement.

Curiosity-driven Exploration by Self-supervised Prediction — mentioned by briga

Sun, 25 Nov 2018 20:25:42 GMT

Reinforcement learning has a pretty useful framework for thinking about this called the exploration-exploitation dilemma. In order for an agent to fully exploit it's environment it first has to explore that environment, and this is where curiosity becomes useful. In fact, last year there was a paper that tried to introduce curiosity to RL-based learning models [1] I suspect some sort of cross-pollination of these two areas of research would be useful. [1] https://www.google.com/url?q=https://arxiv.org/pdf/1705.0536...

Paper and discussion history. A mention is not an endorsement.