03:32
A12 Differentially Private Survival Function Estimation
-
Video
-
Table of contents
-
Video
- Description
- Transcript
- Discussion
About the talk
00:14 Problem Statement
00:30 Our Proposal
00:56 Algorithms
02:55 Results
03:24 Conclusion
About speakers
Ke Wang received Ph.D from Georgia Institute of Technology. He is currently a professor at School of Computing Science, Simon Fraser University. Ke Wang's research interests include database technology, data mining and knowledge discovery, with emphasis on massive datasets, graph and network data, and data privacy. He is particularly interested in combining the strengths of database, statistics, machine learning and optimization to provide actionable solutions to real life problems and industrial applications. Ke Wang has published in database, information retrieval, and data mining conferences, including SIGMOD, SIGIR, PODS, VLDB, ICDE, EDBT, SIGKDD, SDM, ICDM, WWW, AAAI, and CIKM. He has 30+ papers that receive 100+ citations each. He co-authored a book "Introduction to Privacy-Preserving Data Publishing: Concepts and Techniques", Data Mining and Knowledge Discovery Series, Chapman & Hall/CRC, August 2010. He is currently an associate editor of the ACM TKDD journal and he was an associate editor of the IEEE TKDE journal, an editorial board member for Journal of Data Mining and Knowledge Discovery. He has been the general co-chair for the SIAM Conference on Data Mining 2015 and 2016, and the PC co-chair for SIAM Conference on Data Mining 2008 and the PC co-chair for IEEE International Conference on Intelligence and Security Informatics (ISI) 2010.
View the profileI am a PhD student at Simon Fraser University’s database and data mining lab. I am supervised by Prof. Ke Wang. My main research interests are in Machine Learning (especially Deep Learning), Statistics, and the intersection of both with their real world applications. My current research is primarily focused on privacy preserving machine learning and machine learning in healthcare.
View the profilethis is lovely Pandora and I'll be presenting us steady different sleep by Poets survival function estimation, as be no cap on Mars, method is one of the most often used methods for the Survivor function estimation in clinical studies and given the nature of the patient-level data, and Trust to protect the privacy of the In order to do so. Our method being simple and straightforward can be easily extended to us to meet many of those such as private. In summary restart with a partial Matrix M that has number address for the first time point. And the number of events for every other time. Then
using that partial Matrix, we add noise to an unlawful distribution with the sensitivity ask and our privacy promise. No, the M Prime that is a noisy version of em is the currency pilot when using M Prime we complete the number address riches in a similar way as people do if you would have no noise. When using the noisy number 11 and the noisy number at risk, we estimate our survival function, which is now. The sensitivity for a scenario for the partial matrix-m is too because we only have the data for the very first time Point
phone number address and for every time point for number of events so adding or removing a single individual can only change the account where it goes to. That is either an address for the first time point or having intimate. The differential privacy of our algorithm one can be. Similarly, proved using the sensitivity and the fact that the only use no return for a front restoration. And as we've talked briefly before the conference center around the country
and the hypothyroid and our method can be extended to estimate the competing, this Kim rid of incident and the Nelson Allen decimated. 9 real-world data sets, and we can see, but our method provides good utility. And as expected horse, noise level increases for tight privacy. and we observe similar trends when looking at the estimates of the median survival along with the conference, For the complete set of results, please have a look at our paper. In the end, we would like to acknowledge and
Buy this talk
Ticket
Interested in topic “Artificial Intelligence and Machine Learning”?
You might be interested in videos from this event
Similar talks
Buy this video
Conference Cast
With ConferenceCast.tv, you get access to our library of the world's best conference talks.
