Search results
Quantitative Researcher, Citadel Securities. Other activities: AddisCoder JamCoders USVICoder workshop on chaining Simons program on probability, geometry, and computation in high dimensions. Notes on sketching and streaming algorithms from the TUM Summer School on Mathematical Methods for High-Dimensional Data Analysis. CS 270 (UC Berkeley).
UC Berkeley Berkeley, CA. Chair, Computer Science Division, Department of Electrical Engineering and Computer Sci-ences, July 2024–present. Professor, Department of Electrical Engineering and Computer Sciences, July 2019–present.
Proceedings of the 32nd Annual Conference on Advances in Neural Information Processing Systems (NeurIPS 2022), New Orleans, LA, November 28-December 9, 2022. Maryam Aliakbarpour, Andrew McGregor, Jelani Nelson, Erik Waingarten. Estimation of Entropy in Constant Space with Improved Sample Complexity. Proceedings of the 32nd Annual Conference on ...
Tuesday, Aug. 11 — JL optimality, point query via incoherence, sparse JL, FFT-based fast JL (bonus: not covered in lecture). Thursday, Aug. 13 — Gordon's theorem, numerical linear algebra, compressed sensing (RIP). BibTeX file. Professor, Department of EECS.
2016 TUM Summer School on Mathematical Methods for High-Dimensional Data Analysis, Notes on sketching and streaming. You may also find these lecture notes and these videos helpful (from lectures 1—3, 5, and 6). Monday, Jul. 18 — introduction, basic tail bounds, Morris' algorithm.
Xin Lyu 吕欣. I am a third-year Ph.D. student in the Theory Group at UC Berkeley, where I am fortunate to be advised by Avishay Tal and Jelani Nelson. My research interests lie in theoretical computer science in general.
Optimality of the Johnson-Lindenstrauss lemma Kasper Green Larsen Jelani Nelsony November 8, 2017 Abstract For any d;n 2 and 1=(minfn;dg)0:4999 < " < 1, we show the existence of a set of n vectors X ˆRd such that any embedding f : X !Rm satisfying 8x;y 2X; (1 ")kx yk2 2 kf(x) f(y)k 2
Jelani Nelson Huacheng Yuy October 5, 2020 Abstract Storing a counter incremented Ntimes would naively consume O(logN) bits of memory. In 1978 Morris described the very first streaming algorithm: the “Morris Counter” [Mor78]. His algorithm has been shown to use O(loglogN+log(1=")+log(1= ))bits of memory to provide a (1+")-approximation
Daniel M. Kane Jelani Nelsony Abstract We give two di erent and simple constructions for dimensionality reduction in ‘ 2 via linear mappings that are sparse: only an O(")-fraction of entries in each column of our embedding matrices are non-zero to achieve distortion 1+ "with high probability, while still achieving the
Jelani Nelson minilek@seas.harvard.edu July 21, 2016 Today the focus of the lecture will be on linear sketching. Up until this point, we have focused on streaming algorithms in the so-called insertion-only model. Speci cally, consider the scenario of a vector x2Rn with nlarge, and xstarting as the 0 vector. Then we see a sequence of updates (i ...
