MétaCan
Menu
Back to cohort
Record W3161297530 · doi:10.1109/jsait.2021.3079856

Asynchronous Delayed Optimization With Time-Varying Minibatches

2021· article· en· W3161297530 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

VenueIEEE Journal on Selected Areas in Information Theory · 2021
Typearticle
Languageen
FieldComputer Science
TopicStochastic Gradient Optimization Techniques
Canadian institutionsUniversity of Toronto
FundersNatural Sciences and Engineering Research Council of CanadaHuawei Technologies
KeywordsAsynchronous communicationRegretComputer scienceVariable (mathematics)Mathematical optimizationProcess (computing)Mathematics

Abstract

fetched live from OpenAlex

Large-scale learning and optimization problems are often solved in parallel. In a master-worker distributed setup, worker nodes are most often assigned fixed-sized minibatches of data points to process. However, workers may take different amounts of time to complete their per-batch calculations. To deal with such variability in processing times, an alternative approach has recently been proposed wherein each worker is assigned a fixed duration to complete the calculations associated with each batch. This fixed-time approach results in time-varying minibatch sizes and has been shown to outperform the fixed minibatch approach in synchronous optimization. In this paper we make a number of contributions in the analysis and experimental verification of such systems. First, we formally present a system model of an asynchronous optimization scheme with variable-sized minibatches and derive the expected minibatch size. Second, we show that for our fixed-time asynchronous approach, the expected gradient staleness does not depend on the number of workers contrary to existing schemes. Third, we prove that for convex smooth objective functions the asynchronous variable minibatch method achieves the optimal regret and optimality gap bounds. Finally, we run experiments comparing the performances of the asynchronous fixed-time and fixed-minibatch methods. We present results for CIFAR-10 and ImageNet.

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)0.001
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Simulation or modeling · Consensus signal: Simulation or modeling
GenreCandidate signal: Methods · Consensus signal: Methods
Teacher disagreement score0.354
Threshold uncertainty score0.707

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0010.000
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0000.000
Bibliometrics0.0010.002
Science and technology studies0.0000.000
Scholarly communication0.0000.003
Open science0.0000.000
Research integrity0.0000.000
Insufficient payload (model declined to judge)0.0000.000

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 head0.006
GPT teacher head0.211
Teacher spread0.205 · 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