In sparse recovery with mixed-quality data, the information-theoretic price of quality is bounded by two in the agnostic decoder case but can grow arbitrarily large when informed, while LASSO recovery depends only on average noise level.
(40) 21 Published as a conference paper at ICLR 2026 Similarly to (40), we obtain the following expression forM −∆2 i (θ/σ2 2): M−∆2 i (θ/σ2
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Price of Quality: Sufficient Conditions for Sparse Recovery using Mixed-Quality Data
In sparse recovery with mixed-quality data, the information-theoretic price of quality is bounded by two in the agnostic decoder case but can grow arbitrarily large when informed, while LASSO recovery depends only on average noise level.