Upcoming Events

Fair Resource Allocation: From Theory to Practice

CS-ECON Seminar Series
Speaker Name
Alexsandros Psomas
Location
Gross 304B
Date and Time
-

We study the problem of fairly allocating a set of indivisible items among $n$ agents. Typically, the literature has focused on one-shot algorithms. In this talk we depart from this paradigm and allow items to arrive online. When an item arrives we must immediately and irrevocably allocate it to an agent. A paradigmatic example is that of food banks: food donations arrive, and must be delivered to nonprofit organizations such as food pantries and soup kitchens.

On the Convergence of Policy Gradients for the Linear Quadratic Regulator

Duke Computer Science Colloquium
Speaker Name
Maryam Fazel
Location
LSRC D106
Date and Time
-

Policy gradient methods for reinforcement learning and continuous control are popular in practice and have helped recent advances in robotic navigation and in game playing. However, they lack theoretical guarantees even for the simplest case of linear dynamics and a quadratic cost, the Linear Quadratic Regulator (LQR) problem. A difficulty is that unlike the classical approaches to LQR, these methods must solve a nonconvex optimization problem to find the optimal control policy.

The Resurgence of Software Performance Engineering

Duke Computer Science Distinguished Lecture Series
Speaker Name
Charles E. Leiserson
Location
LSRC D106
Date and Time
-

Today, most application developers write code without much regard for how quickly it will run.  Moreover, once the code is written, it is rare for it to be reengineered to run faster.  But two technology trends of historic proportions are instigating a resurgence in software performance engineering, the art of making code run fast.  The first is the emergence of cloud computing, where the economics of renting computation, as opposed to buying it, heightens the utility of application speed.

Approximate Inference for High-Dimensional Latent Variable Models

Ph. D. Defense
Speaker Name
Zilong Tan
Location
3232 French Family Science Center
Date and Time
-

Latent variable models are widely used in applications ranging from natural language processing to recommender systems. Exact inference using maximum likelihood for these models is generally NP-hard, and computationally prohibitive on big and/or high-dimensional data. This has motivated the development of approximate inference methods that balance between computational complexity and statistical efficiency. Understanding the computational and statistical tradeoff is important for analyzing approximate inference approaches as well as designing new ones.

Controlled Experiments: An indispensable tool for modern developers

Duke Computer Science/Electrical Computer Engineering Colloquium
Speaker Name
Jennifer Beckmann
Location
Fitzpatrick Center Schiciano Auditorium Side B, room 1466
Date and Time
-

Controlled experiments, or A/B tests, are the gold standard for optimizing websites. From Amazon's checkout flow to Google's search results, A/B tests can help companies improve workflows to download software, sign-up for subscriptions, click on content, or make a purchase. But, controlled experiments have roots far beyond optimizing websites for conversion. In its simplest form, controlled experiments help establish a causal relationship of a change for a target audience.

Neural Data Structures and their Interrelationship with Algorithms: Insights from how the Brain Links Visual and Auditory Space

Duke Computer Science Colloquium
Speaker Name
Jennifer Groh
Location
LSRC D106
Date and Time
-
The brain confronts a sensor fusion problem in that the auditory and visual systems work in concert to provide information about the environment. I will talk about the computational challenges intrinsic to this problem and how the brain solves them. In particular, the brain uses different “data structures” for visual and auditory information, with digital-like maps for visual spatial location and analog-like meters for auditory spatial information. I will emphasize recent discoveries concerning the algorithms for linking visual and auditory space.