ML benchmarks and their pitfalls
On marginal efficiency gain in paperclip manufacture
August 16, 2020 — October 15, 2024
Your baseline
Your baseline has got me feeling fine
It’s filling up my mind
with apologies to Puretone
Machine learning’s gamified/Goodharted version of the replication crisis is the paper treadmill wherein something counts as a “novel result” if it performs on some conventional benchmarks. But how often does that demonstrate real progress and how often is it overfitting to benchmarks?
Oleg Trott on How to sneak up competition leaderboards.
Jörn-Henrik Jacobsen, Robert Geirhos, Claudio Michaelis: Shortcuts: How Neural Networks Love to Cheat.
Sanjeev Arora, Yi Zhang, Rip van Winkle’s Razor, a Simple New Estimate for Adaptive Data Analysis has a minimum description length approach to model meta-overfitting which I will not summarize except to recommend it for being extremely psychedelic.
1 Explicit connection to Goodhart’s law
Filip Piekniewski on the tendency to select bad target losses for convenience. Measuring Goodhart’s Law at OpenAI.
2 Measuring speed
Lots of algorithms claim to go fast, but that is a complicated claim on modern hardware. Stabilizer attempts to randomise things to give a “fair” comparison.
3 Incoming
MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering | OpenAI
We introduce MLE-bench, a benchmark for measuring how well AI agents perform at machine learning engineering. To this end, we curate 75 ML engineering-related competitions from Kaggle, creating a diverse set of challenging tasks that test real-world ML engineering skills such as training models, preparing datasets, and running experiments. We establish human baselines for each competition using Kaggle’s publicly available leaderboards. We use open-source agent scaffolds to evaluate several frontier language models on our benchmark, finding that the best-performing setup—OpenAI’s o1-preview with AIDE scaffolding—achieves at least the level of a Kaggle bronze medal in 16.9% of competitions. In addition to our main results, we investigate various forms of resource-scaling for AI agents and the impact of contamination from pre-training. We open-source our benchmark code (opens in a new window) to facilitate future research in understanding the ML engineering capabilities of AI agents.