News & EventsDepartment Events & Announcements
Events
-
Apr19
EVENT DETAILS
Abstract
Though the potential of artificial intelligence (AI) in healthcare warrants genuine enthusiasm, meaningful impact will require careful integration into clinical care. AI tools are susceptible to mistakes and rarely capable of capturing all of the nuances pertaining to a complex clinical situation. Thus, we propose approaches designed to augment, rather than replace, clinicians during clinical decision making. In this talk, I will highlight two related research directions in which we propose i) a transfer learning approach for mitigating potentially harmful shortcuts when making diagnoses and ii) a novel reinforcement learning approach for matching patients to treatments. In summary, there’s a critical need for machine learning in healthcare; however, the safe and meaningful adoption of these techniques will require collaboration between clinicians and AI.
Biography
Jenna Wiens is an Associate Professor of Computer Science and Engineering (CSE) and co-Director of Precision Health at the University of Michigan in Ann Arbor. Her primary research interests lie at the intersection of machine learning, data mining, and healthcare. Wiens received her PhD from MIT in 2014, was named Forbes 30 under 30 in Science and Healthcare in 2015, received an NSF CAREER Award in 2016, was named to the MIT Tech Review's list of Innovators Under 35 in 2017, and recently was awarded a Sloan Research Fellowship in Computer Science.
TIME Monday, April 19, 2021 at 12:00 PM - 1:00 PM
CONTACT Pamela Villalovoz pmv@northwestern.edu EMAIL
CALENDAR Department of Computer Science
-
Apr21
EVENT DETAILS
Abstract
Mental health problems such as anxiety disorders and depression have grown in the last decades, with a huge increase in the rate of growth for people under age 26 since the early years of smartphone adoption. Depression is on track to be the global #1 disease burden by 2030. To date, doctors do not understand what causes or precipitates depression, or how to forecast it; nor is there a reliable treatment, although most people will get better. Mobile and wearable devices are improving in their ability to model continuous physiology and activity/sleep/social data for detecting subtly changing patterns related to changing health, and providing early AI-generated forecasts of changes in health. What are some of these patterns telling us about neurological activity, mood, stress and brain health and how accurately? This talk will highlight some of the latest results from our studies at MIT, focusing on both scientific and ethical challenges in using AI to foster good mental health.
Biography
Dr. Rosalind Picard, MIT and Empatica. Dr. Picard is a professor, inventor, engineer, and scientist. She wrote the book Affective Computing that outlines how to give machines the skills of emotional intelligence, which inspired the growing field of Affective Computing. She co-founded two companies that commercialized inventions by her and her team at MIT: Empatica, providing the first AI-smartwatch recognized by the FDA for monitoring seizures, and Affectiva, providing emotion AI software. At MIT, she is a full professor teaching and directing research at the Media Lab, and serves as founding faculty chair for MindHandHeart, MIT’s campus-wide wellbeing initiative. She serves also as chief scientist and Chairman for Empatica, where working with many medical partners, they recently released the first medically approved (in the EU) non-invasive smartwatch running AI for forecasting a positive PCR test for a viral respiratory infection (Influenza H1N1, Rhinovirus, or COVID-19).
TIME Wednesday, April 21, 2021 at 12:00 PM - 1:00 PM
CONTACT Pamela Villalovoz pmv@northwestern.edu EMAIL
CALENDAR Department of Computer Science
-
Apr23
EVENT DETAILS
Abstract
In this talk, I’ll provide a brief overview of “Learning-augmented algorithms”: online algorithms that have access to predictions made by a machine learning oracle. The goal in this setting is to design an algorithm that can utilize good predictions to obtain near-optimal performance in practical instances, while simultaneously guaranteeing robust worst-case guarantees (even when the predictions are of a poor quality).
I’ll then present a specific application of this framework to the classical online learning problem where the online player receives a "hint'' regarding the upcoming cost vector before choosing the action for that round. Our goal is to design a learning algorithm that utilizes "good'' hints to obtain regret that grows only logarithmically with the number of time steps (thus circumventing well known lower bounds), while at the same time being resilient to bad hints. We develop a general framework for online learning with hints that obtains near-optimal regret bounds with respect to the quality of the hints. We also discuss extensions to settings where the algorithm has access to multiple, possibly conflicting, hints at each round and must also pay for switching actions between rounds.
Biography
Manish is a Research Scientist at Google Research, Mountain View. He obtained his PhD from the University of Maryland, College Park under the supervision of Samir Khuller. He is broadly interested in the design and analysis of approximation algorithms as well as machine learning.TIME Friday, April 23, 2021 at 2:00 PM - 3:00 PM
CONTACT Pamela Villalovoz pmv@northwestern.edu EMAIL
CALENDAR Department of Computer Science
-
Apr28
EVENT DETAILS
Abstract
Solver-aided tools have automated the verification and synthesis of practical programs in many domains, from high-performance computing to executable biology. These tools work by reducing verification and synthesis tasks to satisfiability queries, which involves compiling programs to logical constraints. Developing an effective symbolic compiler is challenging, however, and until recently, it took years of expert work to create a solver-aided tool for a new domain.
In this talk, I will present Rosette, a programming language for rapid creation of solver-aided tools. To build a new tool, you write an interpreter for the tool's input language, and Rosette lifts this interpreter into a symbolic compiler. This is made possible by Rosette's symbolic virtual machine, which can translate both a language implementation and programs in that language to efficient constraints. Since its first public release in 2014, Rosette has enabled a wide range of programmers, from professional developers to high school students, to create dozens of new verification and synthesis tools. Example applications include verifying radiation therapy software in current clinical use, synthesizing GPU kernels, and verifying and synthesizing just-in-time compilers that are part of the Linux operating system. This talk will provide a brief introduction to Rosette and describe recent applications.
Biography
Emina Torlak is an Associate Professor in the Paul G. Allen School of Computer Science & Engineering at the University of Washington, working on new languages and tools for computer-aided design, verification, and synthesis of software. She received her Bachelors (2003), Masters (2004), and Ph.D. (2009) degrees from MIT, and subsequently worked at IBM Research, LogicBlox, and as a research scientist at U.C. Berkeley. Emina is the creator of the Kodkod solver, which has been used in over 70 academic and industrial tools for software engineering. Her work on the Rosette system integrates solvers programming languages, enabling programmers to create their own solver-aided tools for all kinds of systems, from radiotherapy machines to automated algebra tutors. Emina is a recipient of the NSF CAREER Award (2017), Sloan Research Fellowship (2016), and the AITO Dahl-Nygaard Junior Prize (2016).
TIME Wednesday, April 28, 2021 at 12:00 PM - 1:00 PM
CONTACT Pamela Villalovoz pmv@northwestern.edu EMAIL
CALENDAR Department of Computer Science
-
Apr28
TIME Wednesday, April 28, 2021 at 12:30 PM - 1:00 PM
CONTACT Northwestern Engineering Events northwestern-engineering-events@northwestern.edu EMAIL
CALENDAR McCormick School of Engineering and Applied Science
-
May10
EVENT DETAILS
TIME Monday, May 10, 2021 at 2:00 PM - 3:00 PM
CONTACT Northwestern Engineering Events northwestern-engineering-events@northwestern.edu EMAIL
CALENDAR McCormick School of Engineering and Applied Science
-
May12
EVENT DETAILS
Speaker: Joseph Holtgreive, assistant dean for undergraduate engineering
TIME Wednesday, May 12, 2021 at 12:00 PM - 12:30 PM
CONTACT Northwestern Engineering Events northwestern-engineering-events@northwestern.edu EMAIL
CALENDAR McCormick School of Engineering and Applied Science
-
May31
EVENT DETAILS
Memorial Day (no classes)
TIME Monday, May 31, 2021
CONTACT Office of the Registrar nu-registrar@northwestern.edu EMAIL
CALENDAR University Academic Calendar
-
Jun7
EVENT DETAILS
Spring examinations begin
TIME Monday, June 7, 2021
CONTACT Office of the Registrar nu-registrar@northwestern.edu EMAIL
CALENDAR University Academic Calendar
-
Jun12
EVENT DETAILS
Spring examinations end
TIME Saturday, June 12, 2021
CONTACT Office of the Registrar nu-registrar@northwestern.edu EMAIL
CALENDAR University Academic Calendar
-
Jun14
EVENT DETAILS
Commencement (Tentative)
TIME Monday, June 14, 2021
CONTACT Office of the Registrar nu-registrar@northwestern.edu EMAIL
CALENDAR University Academic Calendar