Submodular maximization under a Gaussian model selects small benchmark subsets that outperform random selection for imputing leaderboard scores, with mutual information better than entropy at small sizes.
Since the greedy algorithmgrows S one element at a time, we can maintain a Cholesky factorization ofΣ SS via rank-one updates at costO(k 2n)overall
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Submodular Benchmark Selection
Submodular maximization under a Gaussian model selects small benchmark subsets that outperform random selection for imputing leaderboard scores, with mutual information better than entropy at small sizes.