Monoculture Robust Learning
We introduce a notion of monoculture robustness in Probably Approximately Correct (PAC) learning, as a way to mitigate the problem of algorithmic monoculture (Kleinberg and Raghavan, 2021): the tendency of models sharing the same components (e.g., training data) to fail on the same inputs. Specifically, we ask whether a learner with access to a single dataset can output $k$ individually accurate hypotheses whose joint failure probability is comparable to training on $k$ independent datasets. We say that such a learning procedure achieves monoculture robustness, and we quantify this by bounding the conjunctive error that all $k$ hypotheses misclassify points in an individual sense. We design methods for achieving this type of guarantee by drawing a surprising connection between replicability (Impagliazzo et al., 2022) and monoculture robustness. Specifically, we show that if the base learner is $ρ$-replicable, then running it independently $k$ times on the same sample yields individual monoculture error that shrinks exponentially with $k$, controlled by the single-run error of the base learner at the point and $ρ$. We complement these positive results with a lower bound: for individual monoculture robustness, we prove a worst-case lower bound showing that, for learners, driving the individual conjunctive error down to $γ$ can require a sample-size multiplier of $t = Ω(\log(1/γ))$.
Authors
- Jon Michael Kleinberg (ORCID: https://orcid.org/0000-0002-1929-2512)
- Omer Reingold (ORCID: https://orcid.org/0000-0003-4997-1716)
- Lee Cohen
Publication Details
- Journal
- arXiv (Cornell University)
- Published
- 2026-11-18
- DOI
- https://doi.org/10.48550/arxiv.2611.24881
- Primary Topic
- Machine Learning and Algorithms
- Type
- preprint