MétaCan
Menu
Back to cohort

Dynamic Regret Bounds without Lipschitz Continuity: Online Convex Optimization with Multiple Mirror Descent Steps

2022· article· en· W4294691466 on OpenAlex

Why this work is in the frame

A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.

affAt least one author lists a Canadian institution in the pinned OpenAlex snapshot.
fundA Canadian funder is recorded on the work.

Bibliographic record

Venue2022 American Control Conference (ACC) · 2022
Typearticle
Languageen
FieldDecision Sciences
TopicAdvanced Bandit Algorithms Research
Canadian institutionsUniversity of Toronto
FundersNatural Sciences and Engineering Research Council of Canada
KeywordsRegretLipschitz continuityPath (computing)Mathematical optimizationConvex optimizationUpper and lower boundsGradient descentMathematicsPath lengthSequence (biology)Regular polygonComputer scienceArtificial intelligenceStatisticsMathematical analysis

Abstract

fetched live from OpenAlex

We study the dynamic regret in online convex optimization (OCO), where the cost functions are revealed sequentially over time. Prior studies on the dynamic regret of OCO algorithms often require the cost functions to be Lipschitz continuous. However, the costs functions that arise in many applications may not satisfy this condition. In this work, we analyze the performance of Online Multiple Mirror Descent (OMMD), which can handle non-Lipschitz cost functions. OMMD is based on mirror descent but uses multiple mirror descent steps per online round. We first derive two upper bounds on the dynamic regret based on the path length and squared path length, and we further derive a third upper bound based on the cumulative optimal cost, which can be much smaller than the path length or squared path length especially when the sequence of minimizers fluctuates over time. We show that the dynamic regret of OMMD scales linearly with the minimum among the path length, squared path length, and cumulative optimal cost.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.002
metaresearch head score (Gemma)0.003
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesMeta-epidemiology (narrow), Insufficient payload (model declined to judge)
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Simulation or modeling · Consensus signal: none
GenreCandidate signal: Empirical · Consensus signal: none
Teacher disagreement score0.745
Threshold uncertainty score1.000

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0020.003
Meta-epidemiology (narrow)0.0010.000
Meta-epidemiology (broad)0.0010.000
Bibliometrics0.0010.003
Science and technology studies0.0010.001
Scholarly communication0.0010.001
Open science0.0020.001
Research integrity0.0000.001
Insufficient payload (model declined to judge)0.0040.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.048
GPT teacher head0.365
Teacher spread0.317 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it