noise-injection-08.09

Challenge: Can we protect the privacy of individual consumers by injecting irrelevant or misdirected noise into the Internet data stream?


Background
Today, online data trails provide large volumes of private, real-time consumer information to companies doing business on the Internet.  Search histories, GPS locations, browsing behaviors or social media content postings, allow companies like Google, Amazon and Facebook to mine data streams to gain insights and details that consumers may consider private (and may incorrectly assume is undiscoverable).  Can we protect the privacy of individual consumers by injecting irrelevant or misdirected noise into the Internet data stream?


Kick Off Date: Wednesday, September 20th 4:30pm - 6pm (TBC)


The Collider Challenge
“Noise Injection” seeks to explore methods by which an individual’s data streams can be obfuscated or rendered statistically invalid via the injection of irrelevant data into an existing data stream.  The project serves to explore the methods by which the data on which machines are trained can be engineered to invalidate the training.  This serves the purpose of both building a mechanism of delivering some amount of Internet privacy to the individual, as well as providing an understanding of how data engineering attacks can be executed so that methods can in turn be developed to defeat malicious attacks by bad-actors.

 

Outcomes
Discover and define mechanisms by which Internet activity signatures can be statistically flooded.  A practical implementation would be to build a browser extension which randomly generates product search activity on Amazon in a way that mimics human browsing patterns, or a browser extension that generates random (but plausible) Google searches. Also, comprehensively discover and define the best practices through which Internet activity signatures can be simply erased: cookie cleaning, browser history cleaning, switching off GPS, etc.

 

Lead Advisor
Shomit Ghose - Managing Director & Partner at Onset Ventures - Shomit is a longtime Silicon Valley entrepreneur with deep experience in software start-ups, both as a venture capitalist / board member, and as an operating executive. Multiple successful IPOs as an operating exec, and multiple successful exits-by-acquisition as both an operating exec and as a board member. Started entrepreneurial life as a UC Berkeley-trained software engineer, and currently a managing director and partner at ONSET Ventures."

 

Eligibility
Graduate and undergraduate students (Juniors & Seniors) in data science, computer science, statistics, IEOR, math, and economics are encouraged to apply. This is a team project and teams will be pre-formed by the project leaders from the pool of eligible students.  

Mandatory Dates: (TBC)
Kick-off: Wednesday, September 20th 4:30pm - 6pm
Midpoint Check in: Wednesday, October 18th 4:30pm - 6pm
Final Presentation: Wednesday, November 15th 4:30pm - 6pm

How to Apply 
Interested students must submit a resume and one paragraph statement of interest to scet.collider@berkeley.edu in ONE pdf file. Please write Noise Injection in the subject line of the email and use Lastname_Firstname when labeling your document. We will let you know if you have been accepted by September 14th.
Deadline to apply is Sunday, September 10th.

Students can take this Collider for 2 units credit:
INDENG 190C Advanced Topics: Innovation Collider
Class # 39423
Academic Units: 2
Graded or Pass/No Pass

Note: If you registered for INDENG 190C and are not eligible or cannot participate in the project, you will be required to drop the course.