“I think it's really cool that these kinds of agent models are catching on in the RL world, with this, and OpenAI's recent paper. And with https://arxiv.org/abs/1703.03864 makes me wonder how close we are to seeing deep learning resurrecting good ol' A-life. If I may divert for a moment to wistfully recall an old project... When I was a teen I read Steven…”
Evolution Strategies as a Scalable Alternative to Reinforcement Learning
We explore the use of Evolution Strategies (ES), a class of black box optimization algorithms, as an alternative to popular MDP-based RL techniques such as Q-learning and Policy Gradients. Experiments on MuJoCo and Atari show that ES is a viable solution strategy that scales extremely well with the number of CPUs available: By using a novel communication strategy based on common random numbers, our ES implementation only needs to communicate scalars, making it possible to scale to over a thousand parallel workers. This allows us to solve 3D humanoid walking in 10 minutes and obtain competitive results on most Atari games after one hour of training. In addition, we highlight several advantages of ES as a black box optimization technique: it is invariant to action frequency and delayed rewards, tolerant of extremely long horizons, and does not need temporal discounting or value function approximation.
候选层:至少 2 个不同的人、在 2 个不同月份提到过,还没到主书架的 3 人 3 个月。没有做立场标注,所以只写“被提及”。
HN 书评
推荐在前,同类按时间 · 每篇最多保留 20 条,统计用全部提及“> we were able to solve one of the hardest MuJoCo tasks (a 3D humanoid) using 1,440 CPUs across 80 machines in only 10 minutes. As a comparison, in a typical setting 32 A3C workers on one machine would solve this task in about 10 hours. So using 80 times more machines makes you 60 times faster (assuming those are the same machines) "while performing bett…”
“The key problem with the TPU architecture for training is that its on-die buffer space is fairly limited and its memory bandwidth is currently nowhere near that of a GPU. This matters if you've got a lot of intermediate buffers that need to be alive at the same time for backpropagation. Also, there's the less fundamental matter of convenience: while there…”
- Evolution Strategies as a Scalable Alternative to Reinforcement Learning ↗ntoshev · 2017-03-14 · 3 分