A poly(d,k)-time algorithm learns mixtures of k heavy-tailed spherical distributions via high-dimensional sparse Fourier transforms without needing mean separation.
Hence, we can indeed pick in line 8 a bππβππ that |ππβ©π΅π 2πβ²(bππ)|β₯3 5 π > 2 5 π , and by Lemma A.2, there is aπ β²β[π]thatβ₯bππβππβ²β₯2β€3πβ²β€π
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DS 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Learning Mixture Models via Efficient High-dimensional Sparse Fourier Transforms
A poly(d,k)-time algorithm learns mixtures of k heavy-tailed spherical distributions via high-dimensional sparse Fourier transforms without needing mean separation.