Главная
Study mode:
on
1
Intro
2
Finding frequent items
3
Harder problem: change detection
4
Turnstile streaming algonthms
5
Bounds attained for lg-heavy hitters
6
BPTree
7
Reduction to finding super-heavy items (BLIW'16)
8
Final reduction
9
Example application of core lemma
10
Basic idea to make use core lemma
11
Comparison of bounds
12
Our contribution
13
Our solution
14
Main data structure
15
Stitching chunks as paths
16
Stitching chunks the right way
17
Using expanders for stitching
18
Answering HH queries
19
Local Differential Privacy
20
Things to optimize
21
Utility of meta approach By independence
22
Experiments
23
Code release
24
Tradeoff
Description:
Explore the evolution and advancements in frequent items algorithms over the past four decades in this 46-minute lecture by Jelani Nelson for the International Mathematical Union. Delve into various aspects of the field, including change detection, turnstile streaming algorithms, and bounds for lg-heavy hitters. Examine the BPTree structure, reductions to finding super-heavy items, and the application of core lemmas. Discover the speaker's contributions, including their novel solution and main data structure. Learn about techniques such as stitching chunks as paths, using expanders for stitching, and answering HH queries. Investigate Local Differential Privacy, optimization strategies, and the utility of meta approaches. Gain insights from experimental results and explore the tradeoffs involved in frequent items algorithms.

Jelani Nelson- Forty Years of Frequent Items

International Mathematical Union
Add to list