AI Paper Atlas

Research map / Bandits and online learning / Online convex optimization and games

Dynamic Local Regret for Non-convex Online Forecasting

Sergul Aydore, Tianhao Zhu, Dean P. Foster

NeurIPS 2019 · Bandits and online learning › Online convex optimization and games · approach: Parameter-free non-stationary online learning · 2 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 dynamic exponentially time-smoothed SGD instead of online SGDLSTM · GEFCom2014 electrical loads, test setQL_grand: 14,612 -> 14,585Improved
Use dynamic exponentially time-smoothed SGD instead of static time-smoothed SGDLSTM · GEFCom2014 electrical loads, test setQL_grand: 14,595 -> 14,585Improved
Use dynamic exponentially time-smoothed SGD rather than online SGDLSTM · GEFCom2014 electrical loads, varying learning ratesrobustness to learning-rate selectionImproved
Use dynamic exponentially time-smoothed SGD rather than static time-smoothed SGDLSTM · GEFCom2014 electrical loads, varying window sizes and learning ratesstabilityImproved
Use dynamic exponentially time-smoothed SGD rather than offline SGD retrained on all past dataLSTM · GEFCom2014 electrical loads, varying window sizescomputation timeImproved
Evaluate historical losses at their corresponding historical parameters rather than at the current parameterDynamic Local Regret · Toy example, T=3local regret ranking of oracle versus stale policy: Static Local Regret: $40/9$ vs. $4/9$; Dynamic Local Regret: 0 vs. $10/9$Improved
Increase the STS-SGD window sizeSTS-SGD · GEFCom2014 electrical loads, varying learning ratesstabilityHurt
Increase the DTS-SGD window sizeDTS-SGD · GEFCom2014 electrical loads, varying learning ratesstabilityNo clear effect

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