English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Safe reinforcement learning has many variants and it is still an openresearch problem. Here, we focus on how to use action guidance by means of anon-expert demonstrator to avoid catastrophic events in a domain with sparse,delayed, and deceptive rewards: the recently-proposed multi-agent benchmark ofPommerman. This domain is very challenging for reinforcement learning (RL) ---past work has shown that model-free RL algorithms fail to achieve significantlearning. In this paper, we shed light into the reasons behind this failure byexemplifying and analyzing the high rate of catastrophic events (i.e.,suicides) that happen under random exploration in this domain. While model-freerandom exploration is typically futile, we propose a new framework where even anon-expert simulated demonstrator, e.g., planning algorithms such as MonteCarlo tree search with small number of rollouts, can be integrated toasynchronous distributed deep reinforcement learning methods. Compared tovanilla deep RL algorithms, our proposed methods both learn faster and convergeto better policies on a two-player mini version of the Pommerman game.

Safer Deep RL with Shallow MCTS: A Case Study in Pommerman