Assistant Computer Science

Contact Address: Tel: 773-***-****

*** * **** ******, ***.2 *****@****.*****.***

Jacob D Abernethy

Philadelphia PA 19146 www.seas.upenn.edu/~jaber

Academic Research Position

CURRENTLY

Simons Postdoctoral Fellow at UPenn, focus: Machine Learning, advised by Professor

Michael Kearns; position supported by Simon s Foundation (October 2011 - Present)

Graduate

EDUCATION

(PhD - Computer Science) UC Berkeley, Dept. of Computer Science (Fall 2006 - Fall 2011).

Primary focus is algorithms in Machine Learning, Sequential Decision Making, Online Algo-

rithms, Game Theory.

(Masters - Computer Science) Toyota Technological Institute at the University of Chicago

(Fall 2006 - Spring 2006)

Undergraduate

(Bachelor of Science - Mathematics) Mass. Institute of Technology (Jan. 2001-June 2002)

Budapest Semesters in Mathematics (Spring 2000)

University of Massachusetts at Amherst (Fall 98 - Fall 99 and Fall 2000)

Publications

RESEARCH

SUBMITTED: Blackwell Approachability and No-Regret Learning are Equivalent,

Abernethy, J., Bartlett, P., Hazan, E., In Submission to Mathematics of Operations Research

SUBMITTED: Large-Scale Bandit Problems and KWIK Learning Abernethy, J.,

Amin, K., Draief, M., Kearns, M., Submitted to Proceedings of the International Conference on

Machine Learning 2013

Minimax Option Pricing Meets Black-Scholes in the Limit, Abernethy, J., Frongillo,

R., Wibisono, A., Proceedings of the Symposium on the Theory of Computation 2012

A Characterization of Scoring Rules for Linear Properties Abernethy, J. and Frongillo,

R., Proceedings of the Conference on Learning Theory 2012

E cient Market Making via Convex Optimization, and a Connection to Online

Learning, Abernethy, J., Chen, Y., Wortman Vaughan, J., Accepted for Publication 6/2012,

ACM Transactions on Economics and Computation

Interior Point Methods for Full-Information and Bandit Online Learning Abernethy,

J., Hazan, E., Rakhlin, A., Published 06/2012, IEEE Transactions on Information Theory

A Collaborative Mechanism for Crowdsourcing Prediction Problems, Abernethy, J.

and Frongillo, R., Proceedings of Neural Information Processing Systems 2011

Invited for full oral presentation

An Optimization-Based Framework for Automated Market-Making, Abernethy, J.,

Chen, Y., Wortman Vaughan, J., Proceedings of Electronic Commerce 2011

Blackwell Approachability and Low-Regret Learning are Equivalent, Abernethy, J.,

Bartlett, P., Hazan, E., Conference on Learning Theory 2011

Repeated Games against Budgeted Adversaries, Abernethy, J., Warmuth, M., Neural

Information Processing Systems, December 2010

A Regularization Approach to Metrical Task Systems, Abernethy, J., Bartlett, P.,

Buchbinder, N., Stanton, I., Algorithmic Learning Theory, October 2010

A Stochastic View of Optimal Regret through Minimax Duality, Abernethy, J.,

Agarwal, A., Bartlett, P., Rakhlin, A., Conference on Learning Theory, 2009

Beating the Adaptive Bandit with High Probability, Abernethy, J., Rakhlin, A., Con-

ference on Learning Theory, 2009

Graph Regularization Methods for Web Spam Detection, Abernethy, J., Chappelle,

O., Castillo, C., Machine Learning Journal, November 2009

A New Approach to Collaborative Filtering: Operator Estimation with Spectral

Regularization, Abernethy, J., Bach, F., Evgeniou, T., Vert, J.P., Journal of Machine Learning

Research, March 2009

Optimal Strategies from Random Walks Abernethy, J., Yellin, J., Warmuth, M., Con-

ference on Learning Theory, 2008

Competing in the Dark: An E cient Algorithm for Bandit Linear Optimization

Abernethy, J., Hazan, E., Rakhlin, A., Conference on Learning Theory, 2008

Best Student Paper Award and 2008 Pat Goldberg Memorial Best Paper Award from IBM

Optimal Strategies and Minimax Lower Bounds for Online Convex Games Aber-

nethy, J., Bartlett, P., Tewari, A., Rakhlin, A. Conference on Learning Theory, 2008

WITCH: Web Spam Identi cation Through Content and Hyperlinks, Abernethy, J.,

Chapelle, O., Castillo, C., WWW2008, Workshop on Adversarial Information Retrieval

Eliciting Consumer Preferences using Robust Adaptive Choice Questionnaires,

Abernethy, J., Evgeniou, T., Toubia, O., Vert, J.P., IEEE Transactions on Knowledge and Data

Discovery, 2007.

Multitask Learning with Expert Advice. Abernethy, J., Bartlett, P., Rakhlin, A., Con-

ference on Learning Theory, 2007.

Online Discovery of Similarity Mappings. Rakhlin, A., Abernethy, J., Bartlett, P.,

International Conference on Machine Learning, 2007.

The Binning Algorithm and Continuous Experts., Abernethy, J., Langford, J., War-

muth, M., Conference on Learning Theory 2006.

OPEN PROBLEM: Does an E cient Calibrated Forecasting Strategy Exist?, Aber-

nethy, J., Mannor, S., Conference on Learning Theory 2011.

OPEN PROBLEM: Can We Learn to Gamble E ciently?, Abernethy, J., Conference

on Learning Theory 2010.

OPEN PROBLEM: Minimax Games with Bandits, Abernethy, J., Warmuth, M., Con-

ference on Learning Theory 2009.

OPEN PROBLEM: An E cient Bandit Algorithm for T -Regret in Online Mul-

ticlass Prediction?, Abernethy, J., Rakhlin, A., Conference on Learning Theory 2008.

Selection of Recent Talks, Tutorials, Workshops

Tutorial: Prediction, Belief, and Markets

1. Invited Presentation, International Conference on Machine Learning Edinburgh 6/2012

2. Invited Presentation, Conference on Knowledge and Data Discovery Beijing 8/2012

3. Guest Lecturer, UC Berkeley course on Social and Information Networks Berkeley 10/2012

Talk: Prediction Markets and Learning from Crowds

1. Invited Talk, NICTA Canberra Australia 11/2011

2. Invited Talk, Kellogg School of Management, Workshop on the Problem of Prediction

Chicago 12/2011

Talk: Minimax Option Pricing Meets Black-Scholes in the Limit

1. Invited Talk, UCSD Computer Science San Diego 4/2012

2. Invited Talk, Harvard EECS Cambridge MA 5/2012

3. Invited Talk, MIT EECS Cambridge MA 5/2012

4. Conference Talk, Symposium on the Theory of Computation New York 5/2012

5. Invited Talk, Princeton CS Princeton NJ 10/2012

Co-Organized Workshop: Markets, Mechanisms, and Multi-Agent Models . One-day work-

shop examining the interaction of Machine Learning and Economics. Located at the Interna-

tional Conference on Machine Learning, Edinburgh UK, 7/2012. More information at http:

//goo.gl/JaidT.

Short-term Positions

Tubemogul.com, (consultant) designing learning algorithms for video ad selection. (Summer

2010 to Spring 2011)

Microsoft Research New England, intern with Adam Kalai and Nina Balcan at MSR in

Cambridge MA (Summer 2009)

Yahoo! Research, intern with Olivier Chapelle at Yahoo! Research Santa Clara. (Summer

2007)

Convexus Advisors, (consultant) algorithms and online portfolio selection strategies research

for startup hedge fund. (Fall 2006 through Winter 2008)

Yahoo! Research, intern with John Langford at Yahoo! Research New York. (Summer 2006)

INSEAD, research assistant with Prof. Theodoros Evgeniou, in Fontainbleau, France. (Spring

2004)

L Ecole des Mines and INSEAD, research assistantship in collaborative ltering, machine

learning, and matrix approximation, Fontainebleau, France (Feb - March 06)

Brigham and Women s Hospital, research assistantship in bioinformatics with microarray

dataset analysis, Boston, MA (July 03 - Mar. 04)

Academic Teaching Positions

TEACHING

Co-Instructor, Graduate Computational Learning Theory (Spring 12) UC Berkeley

Teaching Assistant in Algorithms (Spring 09) UC Berkeley

Teaching Assistant in Multivariable Calculus (Fall 01, Spring 02) MIT

Teaching Assistant in undergraduate Abstract Algebra (Fall 00) UMass Amherst

Teaching Assistant in Di erential Calculus. UMass Amherst

Assistant Coach of the Mass. State MathCounts team (Spring 01) Framingham, MA

Assistant Coach of the Amherst MathCounts team (Fall 00) Amherst, MA

Management and Development

OTHER WORK

Chief Data Quality O cer, Captricity.com, developed algorithms and strategies for

processing paper forms through Mechanical Turk into structured data (May-October 2011)

Co-Founder and Designer of MLcomp.org Primary developer (with Percy Liang) of

MLcomp, a collaborative environment for objectively executing and comparing machine learning

programs. We hope to make MLcomp a central hub for experimental ML research (March 2008

to Present)

CSGSA President at UC Berkeley The Computer Science Graduate Student Association

is UC Berkeley s primary CS grad student organization. The CSGSA runs all social events, man-

ages relationships with industry partners, maintains several community programs, has roughly

30 o cers and serves over 250 students. (May 2008 to May 2010)

Communications Coordinator, MEET the Middle East Education through Technol-

ogy, an educational initiative aimed at teaching Computer Science and Entrepreneurship to

Israeli and Palestinian high school students in Jerusalem. More information can be found at

http://meet.csail.mit.edu (Jan. 03 - Dec. 06)

Juggling and Performing

ACTIVITIES

I have been a juggler since the age of 14, and a performer since the age of 16. Juggling has been

both a hobby, a source of exercise, and a part time career.

Racquetball

Men s Team Leading Player, UC Berkeley Racquetball 2008-2011.

Winner, Collegiate Nationals Div. 1 Men s Singles, Red bracket, April 2011 (http://goo.gl/qW4Jo)

Bicycling

Avid touring cyclist. Longest trip began in Boston and ended in San Francisco, totaling 4600

miles, 72 days, and 13 at tires. More information at http://kvand.no-ip.org/xcountry/

Academic

AWARDS

Yahoo! PhD Fellowship 2008 Winner, one of four students to receive full two years of

nancial support towards a PhD. (Spring 2008)

Winner of the Web Spam Challenge, Track II of a Machine Learning competition that

compared submitted algorithms for detecting Web Spam. (Summer 2007)

Winner of Key Technical Challenges Grant, a Yahoo!-funded award of $5000 in support

of academic research. (Spring 2007)

Performing

Invited to perform Comedy/Juggling routine as the opener for two well-known comedians, Sinbad

and later Dave Chappelle, during their appearances at the University of Massachusetts, Amherst.

Winner of the UMass College Comedy Bake-O, acquiring the title of Funniest Student at UMass,

Oct. 3rd, 2001

Winner of the national College Comedy Bake-O, acquiring the title of Funniest Student in

America.

Featured on a commercial played on the College TV Network.

Contact this candidate