Master’s Thesis Presentation • Machine Learning • Private Distribution Learning with Public DataExport this event to calendar

Friday, June 30, 2023 — 3:00 PM to 4:00 PM EDT

Please note: This master’s thesis presentation will take place in DC 3317.

Alex Bie, Master’s candidate
David R. Cheriton School of Computer Science

Supervisors: Professors Gautam Kamath, Shai Ben-David

We study the problem of private distribution learning with access to public data. In this setup, a learner is given both public and private samples drawn from an unknown distribution p belonging to a class Q, and has the task of outputting an estimate of p while adhering to privacy constraints (here, pure differential privacy) only with respect to the private samples.

Our setting is motivated by the privacy-utility tradeoff: algorithms satisfying the mathematical definition of differential privacy offer provable privacy guarantees for the data they operate on, however, owing to such a constraint, exhibit degraded accuracy. In particular, there are classes Q where learning is possible when privacy is not a concern, but for which any algorithm satisfying the constraint of pure differential privacy will fail on.

We show that in several scenarios, we can use a small amount of public data to evade such impossibility results. Additionally, we complement these positive results with an analysis of how much public data is necessary to see such improvements. Our main result is that to learn the class of all Gaussians in R^d under pure differential privacy, d+1 public samples suffice while d public samples are necessary.

Location 
DC - William G. Davis Computer Research Centre
DC 3317
200 University Avenue West

Waterloo, ON N2L 3G1
Canada
Event tags 

S M T W T F S
30
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
1
2
3
  1. 2024 (184)
    1. September (1)
    2. August (5)
    3. July (20)
    4. June (17)
    5. May (23)
    6. April (41)
    7. March (27)
    8. February (25)
    9. January (25)
  2. 2023 (296)
    1. December (20)
    2. November (28)
    3. October (15)
    4. September (25)
    5. August (30)
    6. July (30)
    7. June (22)
    8. May (23)
    9. April (32)
    10. March (31)
    11. February (18)
    12. January (22)
  3. 2022 (245)
  4. 2021 (210)
  5. 2020 (217)
  6. 2019 (255)
  7. 2018 (217)
  8. 2017 (36)
  9. 2016 (21)
  10. 2015 (36)
  11. 2014 (33)
  12. 2013 (23)
  13. 2012 (4)
  14. 2011 (1)
  15. 2010 (1)
  16. 2009 (1)
  17. 2008 (1)