Bilinear Compressive Security with Finitely Many Samples
Researchers show finite-sample key recovery against Bilinear Compressive Security depends on measurement-matrix sparsity.
A cs.CR paper analyzes Bilinear Compressive Security when an attacker sees only finitely many ciphertexts rather than the full distribution. The authors present a covariance-based key-recovery attack that estimates linear measurements from repeated encryptions of the same plaintext and then recovers the secret measurement matrix, with finite-sample guarantees and sample-complexity bounds. Experiments find roughly linear measurement growth for dense Gaussian matrices, unlike the information-theoretic worst case, while sparse-row matrices follow the predicted quadratic growth.
- BCS pairs compressed sensing with a fresh random convolution filter.
- Earlier proofs assumed an adversary sees the full ciphertext distribution.
- A covariance attack recovers measurements from repeated encryptions of one plaintext.
- Dense Gaussian matrices show roughly linear sample growth with dimension.
- Sparse-row matrices show quadratic growth consistent with theory.
From the original article
1 min read at arxiv.orgBilinear Compressive Security (BCS) is a lightweight encryption scheme that combines compressed sensing with a fresh random convolution filter to improve resistance against key-recovery attacks under key reuse.
By Axel Flinth · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.