TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Learning to Communicate with Deep Multi-Agent Reinforcement Learning

73 pointsby seycombiabout 8 years ago

2 comments

taliesinbabout 8 years ago
I think it&#x27;s really cool that these kinds of agent models are catching on in the RL world, with this, and OpenAI&#x27;s recent paper. And with <a href="https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1703.03864" rel="nofollow">https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1703.03864</a> makes me wonder how close we are to seeing deep learning resurrecting good ol&#x27; A-life. If I may divert for a moment to wistfully recall an old project...<p>When I was a teen I read Steven Levy&#x27;s book &quot;Artifical Life&quot;. It captured my imagination in a way that the biomorph program in a Blind Watchmaker also had. Ever since then I&#x27;ve played with toy alife models. Some of the attraction is just aesthetic - it&#x27;s fun to set these things up and design the rules, kinda like a higher-order god-game. Another is the enticing fantasy that something like this will one day lead to AGI. A third is the occasionally fulfilled desire to see the fabled emergent behavior of evolutionary computations, where the agents learn to do something you never expected, or exploit some funky loophole to their advantage.<p>When I was an undergrad I created an alife system as an honors project. It was an ambitious project, with sexually or asexually reproducing agents controlled by recurrent neural nets. The agents had perception and the ability to move, reproduce, attack, co-operate, or interact with objects. It had an entity framework to make it possible to design things like programmable Skinner boxes, traps, feeders, it had feeding schedules, it had live visualization and a slick Qt interface to edit the entities inline or take control of specific agents and visualize their brains.<p>I was so proud of it.<p>My assigned supervisor, who was an austere Georgian category theorist, had zero interest. He compared it unfavorably to a virus. Anyway, it was so dispiriting an experience that I kind of swore off Alife after that. But the fun I had made me realize I didn&#x27;t want to be a mathematician, so I suppose it played a useful role in my life after all.
评论 #13896703 未加载
评论 #13898154 未加载
pebblexeabout 8 years ago
Reminds me of this: <a href="https:&#x2F;&#x2F;fatiherikli.github.io&#x2F;language-evolution-simulation&#x2F;" rel="nofollow">https:&#x2F;&#x2F;fatiherikli.github.io&#x2F;language-evolution-simulation&#x2F;</a>
评论 #13898178 未加载