Upcoming Events

On the Recent Progress on the Inapproximability of High Dimensional Clustering and the Johnson-Coverage Hypothesis

Algorithms Seminar
Speaker Name
Vincent Cohen-Addad
Location
LSRC D344
Date and Time
-
k-median, k-means, and k-minsum are amongst the three most popular objectives for clustering algorithms. Despite intensive effort, a complete understanding of the approximability of these objectives remains a major open problem. In this paper, we significantly improve upon the hardness of approximation factors known in literature.

Toward a New Theoretical Foundation for Machine Learning: The Lessons of Deep Learning

Duke Computer Science Colloquium
Speaker Name
Mikhail Belkin
Location
LSRC D106
Date and Time
-
In this talk I will discuss first steps toward a new theoretical foundations for modern ML. Furthermore, I will show how classical and modern ML models can be unified within a single "double descent" risk curve, which subsumes the classical U-shaped generalization curve and spans both parametric and non-parametric models. Finally, I will comment on the nature of the inductive biases, directions of future theoretical analyses and important implications for optimization.

A Variational Perspective on Optimization, Sampling, and Games for Machine Learning

Duke Computer Science Colloquium
Speaker Name
Andre Wibisono
Location
LSRC D106
Date and Time
-

Machine learning has been successful and is prevalent in everyday life, shaping many aspects of modern society. Nevertheless, many fundamental questions remain, and it is important to develop a proper theoretical understanding of machine learning to guide its future development. In this talk I will discuss the fundamental properties of optimization, sampling, and game dynamics for machine learning.

Active Learning Methods for Model Personalization in Mobile Health

Miscellaneous Talk
Speaker Name
Benjamin Marlin
Location
Bryan Research Room 103
Date and Time
-

In mobile health (mHealth), collecting momentary activity or behavioral state labels often comes at a significantly higher cost or level of user burden than collecting unlabeled data using passive sensing. This observation motivates the idea of attempting to optimize the collection of labeled data to minimize cost or burden when developing personalized models. In this talk, I will present on-going research on the problem of developing active learning methods for use in mobile health that leverage the affordances of this domain while respecting its unique constraints.

TechConnect 2020

inDuke Event
Location
Penn Pavilion, Duke University
Date and Time
-

brings students and employers together for networking, education and connections. Twice each year, employers connect with Engineering and Computer Science students in an open, dynamic networking environment. Students come prepared with resumes to meet industry and tech representatives and learn about employment opportunities available, the characteristics employers seek, and also a realistic and insightful view of the job market and career paths for students interested in engineering and technical careers.

Putting Ethical AI to the Vote

Triangle Computer Science Distinguished Lecturer Series
Speaker Name
Ariel Procaccia
Location
LSRC D106
Date and Time
-

I will present the 'virtual democracy' framework for the design of ethical AI. In a nutshell, the framework consists of three steps: first, collect preferences from voters on example dilemmas; second, learn models of their preferences, which generalize to any (previously unseen) dilemma; and third, at runtime, predict the voters' preferences on the current dilemma, and aggregate these virtual 'votes' using a voting rule to reach a decision.