Eric Blais

TL;DR: This work proposes sampling-based incremental visualization algorithms that reveal the "salient" features of the visualization quickly---with a 46× speedup relative to baselines---while minimizing error, thus enabling rapid and error-free decision making.

TL;DR: For example, the authors showed that testing unions of d intervals can be done with O(1) label requests in our setting, whereas it is known that requiring Omega(d) labeled examples for learning (and Omega(sqrt{d}) for passive testing [KR00] where the algorithm must pay for every example drawn from D).

TL;DR: In this paper, it was shown that there is an exponential gap between the query complexity of adaptive and non-adaptive algorithms for testing regular linear threshold functions (LTFs) for monotonicity.

TL;DR: The main result is a non-adaptive algorithm for testing k-juntas with $\tilde{O}(k^{3/2})/\epsilon$ queries, which disproves the conjecture that $O(k^2/2) + O(k/k\big) = 1$ and shows that the query complexity of non- Adaptive algorithms for testing juntas has a lower bound of $\min \big(\tilde{\Omega}

TL;DR: It is proved that the sample complexity is essentially determined by a fundamental operator in the theory of interpolation of Banach spaces, known as Peetre's K-functional, which stems from an unexpected connection to functional analysis and refined concentration of measure inequalities, which arise naturally in the reduction.