AI Paper Atlas

Research map / Optimization / Distributed and compressed

Distributed Multitask Reinforcement Learning with Quadratic Convergence

Rasul Tutunov, Dongho Kim, Haitham Bou Ammar

NeurIPS 2018 · Optimization › Distributed and compressed · approach: Local and asynchronous SGD · 7 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use distributed SDD-Newton instead of ADD, ADMM, distributed averaging, Network Newton and sub-gradientsDistributed multitask RL solver · SM, DM, CP, HC, HR systems, random graph of 20 nodes and 50 edgestime to convergence (10^-5 error threshold)Improved
use distributed SDD-Newton instead of other distributed methods on different graph topologiesDistributed multitask RL solver · HC with small, medium, large random graphs and bar-bell graphiterations to 10^-5 consensus errorImproved
tighten accuracy requirement, with SDD-Newton communication growing with graph condition numberDistributed SDD-Newton · HC systemlocal communication exchangeImproved
use SDD-Newton instead of PG-ELLA, GO-MTL and Dist-ADMMMultitask RL · SM, CP, HC, DMjump-start improvement %Improved
use SDD-Newton instead of streaming models such as PG-ELLAMultitask RL · SM, CP, HC, DMasymptotic performance %Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy