REVIEW 5 cited by
Accounting for Variance in Machine Learning Benchmarks
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Accounting for Variance in Machine Learning Benchmarks
read the original abstract
Strong empirical evidence that one machine-learning algorithm A outperforms another one B ideally calls for multiple trials optimizing the learning pipeline over sources of variation such as data sampling, data augmentation, parameter initialization, and hyperparameters choices. This is prohibitively expensive, and corners are cut to reach conclusions. We model the whole benchmarking process, revealing that variance due to data sampling, parameter initialization and hyperparameter choice impact markedly the results. We analyze the predominant comparison methods used today in the light of this variance. We show a counter-intuitive result that adding more sources of variation to an imperfect estimator approaches better the ideal estimator at a 51 times reduction in compute cost. Building on these results, we study the error rate of detecting improvements, on five different deep-learning tasks/architectures. This study leads us to propose recommendations for performance comparisons.
Forward citations
Cited by 5 Pith papers
-
A Pre-Registered Causal Partition of Self-Consistency Elicitation and Reward Design in RLVR
Derives an exact telescoping decomposition of the naive RLVR reward-design estimator into null, elicitation, and reward-design terms on a tabular-GRPO simulator, measures the components across prior strengths, and val...
-
Grokking Is Conditional and Fragile: A Fully-Tractable, Multi-Seed Study at 12K Parameters
In a fully tractable 12K Llama-style model, grokking is a conditional fragile phase transition gated by coverage (tracking modulus more than structure), weight decay, and floating-point reduction order, so evidence mu...
-
From Forecasting Leaderboards to Deployment Decisions: A Fail-Closed Certification Protocol
Presents a fail-closed certification protocol for determining when forecasting leaderboard winners are deployment-actionable, using a traffic dataset to show friction-induced reversals and an audit to prevent overclaiming.
-
SafetyRepro: Configuration-Conditional Rank Instability on Alignment Benchmarks
Configuration choices alone flip pairwise safety verdicts on every tested alignment benchmark, isolated via a finite-envelope proposition linking disagreement rate to strict ordering reversal.
-
QuickScope: Certifying Hard Questions in Dynamic LLM Benchmarks
QuickScope uses modified COUP Bayesian optimization to find truly difficult questions in dynamic LLM benchmarks more sample-efficiently than baselines while cutting false positives.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.