University of New Mexico Department of Computer Science

04/26/2019

UPDATE: UNM Computer Science Department Colloquium Series

Wednesday, May 1st, 2019
Centennial Engineering Center 1041
2:00-3:00 PM

Guest: Sara Del Valle, Los Alamos National Lab

"Real-time Heterogeneous Data to Guide Infectious Disease Forecasting Models"

Colloquium Abstract: Globalization has created health problems that can no longer be adequately analyzed and mitigated using traditional data analysis techniques and data sources. Whether the goal is to monitor the rapid spread of infectious diseases, societal changes initiated by climate change, or the emergence of a new deadly virus being subtly observed by citizenry, there is a near-real-time digital signature associated with the disruptive event. In this presentation, I will describe an approach that combines heterogeneous data streams, such as Internet and satellite imagery, with mathematical models to monitor and forecast infectious diseases. Our goal is to improve decision support by assimilating real-time information into predictive models to reduce global disease burden.

Speaker Bio: Sara Del Valle is a scientist and deputy group leader in the Information Systems and Modeling Group at Los Alamos National Laboratory. She has a Ph.D. in Applied Mathematics and works on developing, integrating, and analyzing mathematical, computational, and statistical models for the spread of infectious diseases such as smallpox, anthrax, influenza, malaria, HIV, hepatitis C, MERS-CoV, zika, Chikungunya, dengue, and Ebola. In addition, she has modeled the potential effects of mass casualties on the Healthcare and Public Health Sector including resource allocation and dependencies on other infrastructures. Most recently, she has been investigating the role of heterogeneous data streams such as satellite imagery, social media, and climate on detecting, monitoring, and real-time forecasting of infectious diseases.



----
Colloquium faculty contact: Prof. Trilce Estrada, [email protected]
Contact the faculty contact indicated above with questions.

03/27/2019

Posters from the Computer Science Student Conference.

03/26/2019

CSGSA, Women in Computing, and ACM Student Chapter invite you to the 15th CS Student Conference. This is a conference entirely organized by students in our department and it showcases research talks (both mature and in-progress), posters, and demos. We will have the visit of two seasoned researchers for a special talk (Joseph Izraelevitz @ 9) and a keynote (Christopher Moore @ 2). Free breakfast, lunch, and ice cream will be provided through the day. Undergrads are especially encouraged to attend the CS Festival at 3:15.

All are welcome!

=============================================

Agenda (Wednesday, March 27th):

8:30 Breakfast and Registration @ Stamm Commons
CEC 1044

9:00 Special Talk: Joseph Izraelevitz - Practical and Formal Infrastructure for Nonvolatile Memory @ Stamm Commons CEC 1044

10:00 to 12:10 Lightning Talk and Research Sessions @ Stamm Commons CEC 1044

12:10 Lunch (free) @ Centennial Courtyard

1:00 Poster Session @ Stamm Commons CEC 1044

2:00 Keynote: Professor Christopher Moore @ Centennial Auditorium CEC 1041

3:15 CS Festival (Demos, Games, & Ice Cream Social) @ Stamm Commons CEC 1044

https://sc2019.cs.unm.edu/program.html
=============================================

Special Talk: Joseph Izraelevitz - Faculty Candidate
Stamm Commons CEC 1044
9:00-10:00 AM

Title: Practical and Formal Infrastructure for Nonvolatile Memory

For decades, programmers have interacted with persistent storage via a well-defined block-based API, namely, that of the file system. However, it is expected that byte-addressable nonvolatile memory (NVM) will soon be commonplace, potentially transforming main memory into a storage device. This transition forces fundamental changes in how to reason about and manage persistent storage. The possibility that programmers may wish for data in main memory to survive program runs and even system crashes suggests the need to reassess design decisions at all levels of the system stack.

This talk will discuss how system architecture and formal foundations are changing to meet the demands of NVM at all scales. At a formal level, we explore what it means for a program to be "correct" on a machine with NVM. At the hardware level, extensions of memory consistency control the ordering and timing of writes into persistence. At the library level, data structures are transformed to be consistent not only in the face of concurrent access, but also in the wake of crashes. Finally, at the compiler and language level, extensions to transactional memory give programmers the ability to ensure that complex changes to persistent state are not only isolated and consistent, but also failure atomic and durable. We conclude by discussing the second-order effects of these changes and how we expect the software ecosystem will mutate after NVM has become deeply integrated into the stack.

Bio. Dr. Joseph (Joe) Izraelevitz is a postdoctoral researcher in the Nonvolatile Systems Lab at the University of California, San Diego. He previously completed a postdoctoral position at IMDEA Software Institute in Madrid, Spain. He completed his doctoral degree in March 2018 under Prof. Michael L. Scott at the University of Rochester in Rochester, NY. With a background in shared memory programming, his doctoral research explored the impact of new nonvolatile memory technologies on both practical systems infrastructure and formal program reasoning. His research interests include distributed computing theory, shared memory synchronization, and parallelism in general. Prior to his graduate studies, Joe served as an active duty US Army officer for three years, including a year-long deployment to Afghanistan as a staff officer. He received his undergraduate degree from Washington University in St. Louis in 2009.

Keynote: Professor Christopher Moore
Centennial Auditorium CEC 1041
2:00-3:00 PM

Cristopher Moore received his B.A. in Physics, Mathematics, and Integrated Science from Northwestern University and his Ph.D. in Physics from Cornell. From 2000 to 2012 he was a professor at the University of New Mexico with joint appointments in Computer Science and Physics. Since then, Moore has been a resident professor at the Santa Fe Institute and has held visiting positions at École Normale Superieure, École Polytechnique, Université Paris 7, the Niels Bohr Institute, Northeastern University, and the University of Michigan.
He has published over 150 papers at the boundary between physics and computer science, randing from quantum computing, to phase transitions in NP-complete problems, to the theory of social networks and efficient algorithms for analyzing their structure. He is an elected Fellow of the American Physical Society, the American Mathematical Society, and the American Association for the Advancement of Science. With Stephan Mertens, he is the author of The Nature of Computation from Oxford University Press.

03/25/2019

Congratulations to Melanie Moses for becoming a Women in Technology Award winner.

Two professors from The University of New Mexico School of Engineering will be among nine honored by the New Mexico Technology Council as Women in Technology Award winners this month.

03/18/2019

Young alumni, we'd really love to see you here. Sign up at the link below by March 14 for this fun event!

http://evite.me/bDKupqVKHp

03/18/2019

The University of New Mexico School of Engineering placed in the top 100 in the U.S. News & World Report 2020 Best Graduate Schools rankings for engineering, released today. The School of Engineering ranks No. 85 this year. Last year, the School placed No. 83, rising from No. 99 the previous year.

03/18/2019

03/07/2019

UNM Computer Science Department
Friday, March 8th, 2019
Centennial Engineering Center 1026
2:00-3:00 PM

Ferdinando Fioretto
Georgia Institute of Technology

Privacy in the field: Protecting Sensitive Data for AI Applications

Abstract: Advances in artificial intelligence and data science have allowed the development of products that leverage individuals' data to provide valuable services. However, the use of this massive quantity of personal information raises fundamental privacy concerns. Differential Privacy (DP) has emerged as the de-facto standard to addresses the sensitivity of such information and can be used to release privacy-preserving datasets.
Despite its large theoretical value, when these private datasets are used as inputs to complex machine learning or optimization tasks, they may produce results that are fundamentally different from those obtained on the original data.

In this talk, I will focus on the problem of releasing privacy-preserving data for complex data analysis tasks. I will introduce the notion of Constrained-Based Differential Privacy (CBDP) which allow us to cast the data release problem to an optimization problem whose goal is to preserve the salient features of the original dataset. Finally, I will discuss two applications of CBDP for large socio-technical systems related to the optimization of operations in transportation systems and energy networks.

Bio: Ferdinando Fioretto is a postdoctoral researcher at the Georgia Institute of Technology. His research focuses on artificial intelligence, data privacy, and multiagent coordination. Ferdinando has published in several top-ranked AI journals and conferences. He has organized workshops, special tracks, and gave tutorials at AAAI, AAMAS, and CP, and has served the program committee of various AI conferences, including AAAI, IJCAI, AAMAS, and CP. He is the recipient of a best student paper award (CMSB, 2013), a most visionary paper award (AAMAS workshop series, 2017), and a best AI dissertation award (AI*IA, 2017).

03/07/2019

03/05/2019

UNM Computer Science Department Colloquium Series
Wednesday, March 6th, 2019
Centennial Engineering Center 1041
2:00-3:00 PM

Yifan Sun
University of British Columbia

Title: Optimization for machine learning

Abstract:
In the recent few years, huge advances have been made in machine learning, which has transformed many fields such as computer vision, speech processing, and even games. A key "secret sauce" in the success of these models is the ability of certain architectures to learn good representations of complex data; that is, preprocessing the data in a way to make the optimization either more robust or more efficiently solvable.

We investigate two instances where encoding structure in the feature variable facilitates optimization. In the first case, we look at word vectors as language-modeling mathematical constructs, and their use in data mining applications. Then, we investigate the curious ability of proximal methods to quickly identify sparsity patterns in an optimization variable, which greatly facilitates feature selection. These two examples illustrate the diversity of machine learning optimization problems, but also highlight the prominence of underlying themes in structured representations.

Bio: Yifan Sun got her PhD in Electrical Engineering from UCLA in 2015. She worked at Technicolor Research in Palo Alto, California, for two years post graduate schools, focusing on machine learning projects. She is now a postdoctorate at the University of British Columbia in Vancouver. Her research interests are convex optimization, semidefinite optimization, first-order and stochastic methods, and machine learning interpretability.