We present a general framework for designing efficient algorithms for unsupervised learning problems, such as mixtures of Gaussians and subspace clustering. Our framework is based on a meta algorithm that learns arithmetic circuits in the presence of noise, using lower bounds. This builds upon the recent work of Garg, Kayal ... more >>>

Randomness extractors provide a generic way of converting sources of randomness that are
merely unpredictable into almost uniformly random bits. While in general, deterministic randomness
extraction is impossible, it is possible if the source has some structural constraints.
While much of the literature on deterministic extraction has focused on sources ... more >>>

We prove that for every $\alpha \in [1,1.5]$,
$$
\text{BPSPACE}[S]\subseteq \text{TISP}[2^{S^{\alpha}},S^{3-\alpha}]
$$
where $\text{BPSPACE}[S]$ corresponds to randomized space $O(S)$ computation, and $\text{TISP}[T,S]$ to time $poly(T)$, space $O(S)$ computation. Our result smoothly interpolates between the results of (Nisan STOC 1992) and (Saks and Zhou FOCS 1995), which prove $\text{BPSPACE}[S]$ is contained ... more >>>