First Week
David Evans — 20 Jan 2018

This message was also sent out by email.

Since not everyone has joined the slack yet, I’m sending this out by email, but please make sure to join https://secprivml.slack.com soon. I will use that for future communications.

I have grouped the 18 full participants in the class into three teams of six:

Team Bus:

Anant Kharkar
Ashley Gao
Atallah Hezbor
Joshua Holtzman
Mainuddin Ahmad Jonas
Weilin Xu

Team Gibbon:

Aditi Narvekar
Austin Chen
Ethan Lowman
Guy “Jack” Verrier
Jialei Fu
Jin Ding

Team Panda:

Bargav Jayaraman
Christopher Geier
Faysal Hossain Shezan
Nathaniel Grevatt
Nishant Jha
Tanmoy Sen

The teams will probably adjust a bit over the next week as some people who registered for the class didn’t submit a survey, and others may be joining the class late, etc., but otherwise we will plan to keep with these teams through spring break and possible re-arrange things after that. You can, of course, rename your team and develop your own adversarial (but tasteful) team icon.

Each week, one team will be responsible for leading the class, one team for blogging (writing a summary of the class), and one team for food. Everyone in the class is expected to do the preparation for each meeting, which will usually involve reading a few papers (but may involve other activities, at the design of the leading team). See https://secml.github.io/teams/ for the description of team responsibilities. For the first week, Team Bus will be the leading team, Team Gibbon will be the feeding team, and Team Panda will be the blogging team. I have created a slack channel for each team, and you should have an invitation to join it. The immediate tasks for each team are:

Cheers,

— Dave

Starting Seminar
David Evans — 15 Jan 2018

Our first seminar meeting will be Friday, January 26 (not this Friday, which would normally be the first class day, since I will be getting back from California too late to meet this week). Meetings will be Fridays in Rice 032, 9:30am-noon.

Since we’re missing the normal first meeting, I want to do as much of the organizational stuff this week to be able to have a substantive first meeting next week. This means we will group students into teams, have a team assigned to lead the first class, and announced topic and readings by next Tuesday (Jan 23).

To enable this, everyone interested in participating in the class in any way should:

  1. Submit this form by Wednesday (Jan 24): https://goo.gl/forms/FqSmUNHYBPbaKsg72

  2. Read the course syllabus, and follow the directions there to join the seminar slack group.

I will work out the teams based on the form submissions, and set up the schedule for leading classes/blogging. I will have office hours on Monday (Jan 22) at 2:30pm. People from the first presenting team will be expect to meet with me then (not necessarily everyone but at least some of the team), and anyone with questions about the seminar is also welcome to come by. (If this time works for the class, we’ll keep this as my regular office hours time for the semester.)

Sorry for missing the first class, but I think we’ll be able to manage most of what we would have done the first day electronically, and be able to have an effective first meeting on Jan 26.

Syllabus Posted
David Evans — 31 Dec 2017

The Syllabus is now posted.

Welcome
David Evans — 31 Dec 2017

This graduate-level special topics course will be offered in Spring 2018. Meetings will be Fridays, 9:30-noon in Rice Hall 032. More information will be posted here soon, but the seminar format will be roughly similar to what we used to TLSeminar last Spring.

This seminar will focus on understanding the risks adversaries pose to machine learning systems, and how to design more robust machine learning systems to mitigate those risks.

The seminar is open to ambitious undergraduate students (with instructor permission), and to graduate students interested in research in adversarial machine learning, privacy-preserving machine learning, fairness and transparency in machine learning, and other related topics. Previous background in machine learning and security is beneficial, but not required so long as you are willing and able to learn some foundational materials on your own.

For more information, contact David Evans.