Events

Remarkable model performance makes news headlines and compelling demos, but these advances rarely translate to a lasting impact on real-world users. A common anti-pattern is overlooking the dynamic, complex, and unexpected ways humans interact with AI, which in turn limits the adoption and usage of AI in practical contexts. To address this, Juho Kim will argue that human-AI interaction should be considered a first-class object in designing AI applications.

In this talk, Kim will present a few novel interactive systems that use AI to support complex real-life tasks. I discuss tensions and solutions in designing human-AI interaction, and critically reflect on my own research to share hard-earned design lessons. Factors such as user motivation, coordination between stakeholders, social dynamics, and user’s and AI’s adaptivity to each other often play a crucial role in determining the user experience of AI, even more so than model accuracy. My call to action is that we need to establish robust building blocks for “Interaction-Centric AI”—a systematic approach to designing and engineering human-AI interaction that complements and overcomes the limitations of model and data-centric views.

About Juho Kim
Juho Kim is an Associate Professor in the School of Computing at KAIST, affiliate faculty in the Kim Jaechul Graduate School of AI at KAIST, and a director of KIXLAB (the KAIST Interaction Lab) [kixlab.org]. His research in human-computer interaction and human-AI interaction focuses on building interactive and intelligent systems that support interaction at scale, with the goal of improving the ways people learn, collaborate, discuss, make decisions, and take action online. 

He earned his Ph.D. from MIT in 2015, M.S. from Stanford University in 2010, and B.S. from Seoul National University in 2008. In 2015-2016, he was a Visiting Assistant Professor and a Brown Fellow at Stanford University. He is a recipient of KAIST’s Songam Distinguished Research Award, Grand Prize in Creative Teaching, and Excellence in Teaching Award, as well as 14 paper awards from ACM CHI, ACM CSCW, ACM Learning at Scale, ACM IUI, ACM DIS, and AAAI HCOMP.

He is currently spending his sabbatical year at Ringle Inc., a startup building an online language tutoring platform, to transfer his research on automatically analyzing and diagnosing learners’ English proficiency into a real product.

Nessy Tania
Senior Principal Scientist
Quantitative Systems Pharmacology, Early Clinical Division
Pfizer Worldwide Research, Development, and Medical

In this talk, I will share some of my personal journey as a math biologist and applied mathematician who had pursued a tenure-track position in academia and is now working as a research scientist in the biopharma industry. I will discuss similarities and differences, rewards and challenges that I have encountered in both positions. On a more practical aspect, I will discuss how current trainees can prepare for a career in industry (specifically biopharma) and how to seek those opportunities.

I will also describe the emerging field of Quantitative Systems Pharmacology (QSP): its deep root in mathematical biology and how it is currently shaping the drug development process. Finally, I will share some of my own ongoing work as a QSP modeler who is supporting the Rare Disease Research Unit at Pfizer. As a key takeaway, I hope to share that there are multiple paths to success and a rewarding and stimulating career in applied mathematics.

DSI Collaboration Day's facilitate brainstorming for possible research topics and funding opportunities. They are intended for faculty and senior researchers working (or interested in working) in these areas and looking for collaboration. In Spring 2023, we anticipate opening a solicitation for seed funding, with the goal of providing initial support for the newly formed research teams. 

Meet and Greet (Lunch provided) — 12:00 p.m. CST
Presentations Begin — 12:30 p.m. to 2:30 p.m. CST

Speakers will be given about 5 minutes to present their topic and the group will then be provided time to respond via a Q&A format. The speaker schedule is as follows:

1. Ben Hackel, CEMS

2. Sapna Sarupria, Chemistry

3. Ellad Tadmor, AEM

4. Zhi-Li Zhang, CS&E

5. Lian Shen, ME

6. Ardeshir Ebtehaj, CEGE

7. Tian Cui, ME

8. Yiling Zhang, ISyE

Please email csedsi@umn.edu with any questions.

Robust and Efficient Neural Network Training: A New Bi-level Learning Paradigm

This talk will introduce Bi-level Machine Learning (BML), an emerging but overlooked topic rooted in bi-level optimization (BLO), to tackle the recent neural network training challenges in robust and efficient AI. In the first part, I will revisit adversarial training (AT)–a widely recognized training mechanism to gain adversarial robustness of deep neural networks–from a fresh BML viewpoint. Built upon that, I will introduce a new theoretically-grounded and computationally-efficient robust training algorithm termed Fast Bi-level AT (Fast-BAT), which can defend sign-based projected gradient descent (PGD) attacks without using any gradient sign method or explicit robust regularization.

In the second part, I will move to a sparse learning paradigm that aims at pruning large-scale neural networks for improved generalization and efficiency. As demonstrated by the Lottery Ticket Hypothesis (LTH), iterative magnitude pruning (IMP) is the predominant sparse learning method to successfully find ‘winning’ sparse sub-networks. Yet, the computation cost of IMP grows prohibitively as the sparsity ratio increases. I will show that BML provides a graceful algorithmic foundation for model pruning and helps us close the gap between pruning accuracy and efficiency. Please see the references and codes at OPTML Lab repository.

About Dr. Sijia Liu
Dr. Sijia Liu is currently an Assistant Professor at the Department of Computer Science and Engineering, Michigan State University, and an Affiliated Professor at the MIT-IBM Watson AI Lab, IBM Research. His research spans the areas of machine learning, optimization, computer vision, signal processing, and computational biology, with a recent focus on Trustworthy and Scalable ML. He received the Best Paper Runner-Up Award at the Conference on Uncertainty in Artificial Intelligence (UAI) in 2022 and the Best Student Paper Award at the 42nd IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) in  2017. He has published over 60 papers at top-tier ML/AI conferences and (co-)organized several tutorials and workshops on Trustworthy AI and Optimization for Deep Learning at KDD, AAAI, CVPR, and ICML, to name a few.

A PDE-Based Analysis of the Symmetric Two-Armed Bernoulli Bandit
Data Science Seminar

Vladimir Kobzar (Columbia University)

The multi-armed bandit is a classic sequential prediction problem. At each round, the predictor (player) selects a probability distribution from a finite collection of distributions (arms) with the goal of minimizing the difference (regret) between the player’s rewards sampled from the selected arms and the rewards of the arm with the highest expected reward. The player’s choice of the arm and the reward sampled from that arm are revealed to the player, and this prediction process is repeated until the final round.

Our work addresses a version of the two-armed bandit problem where the arms are distributed independently according to Bernoulli distributions and the sum of the means of the arms is one (the symmetric two-armed Bernoulli bandit). In a regime where the gap between these means goes to zero and the number of prediction periods approaches infinity, we obtain the leading order terms of the expected regret for this problem by associating it with a solution of a linear parabolic partial differential equation.

Our results improve upon the previously known results; specifically we explicitly compute the leading order term of the optimal regret in three different scaling regimes for the gap. Additionally, we obtain new non-asymptotic bounds for any given time horizon.

This is joint work with Robert Kohn available here.

Using cloud computing? You might benefit from data science!

Marc Light (Censys.io)

Organizations are doing more and more of their computing using cloud resources. Ideally, this computing would be performant, reliable, and efficient (PRE, as they say at Meta Infra). Drawing on 2.5 years of managing a team of 10 infrastructure data scientists at Meta, I describe classes of use cases and approaches to making data-driven decisions and generally improving PRE using ML/OR techniques. Clearly, large companies like Meta, Google, Amazon, Microsoft, etc. could receive a huge return-on-investment from such work. But increasingly, even medium-sized organizations in cybersecurity, energy analytics, healthcare, etc., could benefit from DS/ML/OR for the effective use of cloud computing.

DSI Collaboration Day's facilitate brainstorming for possible research topics and funding opportunities. They are intended for faculty and senior researchers working (or interested in working) in these areas and looking for collaboration. In Spring 2023, we anticipate opening a solicitation for seed funding, with the goal of providing initial support for the newly formed research teams. 

Meet and Greet (Lunch provided) — 12:00 p.m. CST
Presentations Begin — 12:30 p.m. to 2:30 p.m. CST

Speakers will be given about 5 minutes to present their topic and the group will then be provided time to respond via a Q&A format. The speaker schedule is as follows:

1. Ardeshir Ebtehaj — Satellite Remote Sensing of Water and Environment
2. Leif Olmanson (presented by Marvin Bauer) — Satellite-Derived Water Quality Data from an Automated High-Performance Computing Environment for NearReal-time Monitoring of Lake Water Quality
3. Marvin Bauer — Research and applications of satellite remote sensing to mapping, monitoring and analysis of land and water resources
4. Tian Cui — Low-Cost Ultra-Sensitive Micro Sensors for Water Pollution Detection
5. Lian Shen — High-Fidelity High-Resolution Simulations of Atmospheric and Oceanic Turbulent Flows and Water Waves
6. Aaron Hirsch — The Minnesota Geological Survey: A brief overview
7. Zhenong Jin — Title TBD
8. Yiling Zhang — Hierarchical Decision-Making via Uncertain Bilevel Optimization

Please email csedsi@umn.edu with any questions.

Three Uses of Semidefinite Programming in Approximation Theory

Simon Foucart (Texas A & M University)

In this talk, modern optimization techniques are publicized as fitting computational tools to attack several extremal problems from Approximation Theory which had reached their limitations based on purely analytical approaches.

Three such problems are showcased: the first problem---minimal projections---involves minimization over measures and exploits the moment method; the second problem---constrained approximation---involves minimization over polynomials and exploits the sum-of-squares method; and the third problem---optimal recovery from inaccurate observations---is highly relevant in Data Science and exploits the S-procedure. In each of these problems, one ends up having to solve semidefinite programs.

The Minnesota Natural Language Processing (NLP) Seminar is a venue for faculty, postdocs, students, and anyone else interested in theoretical, computational, and human-centric aspects of natural language processing to exchange ideas and foster collaboration. The talks are every other Friday from 2 - 3 p.m. during the fall 2022 semester.

This week's speaker, Muhao Chen (USC), will be giving a talk titled "Robust and Indirectly Supervised Information Extraction".

View the full event page on the CS&E event page

AI and Future of Healthcare

AI technology is permeating every aspect of healthcare. It is transforming how care is provided in acute, ambulatory and home settings. It is transforming how population health is being managed. AI in combination with other disciplines, such as genetics and virtual reality, is accelerating innovative treatments. Drug discovery and clinical trials are being transformed. This presentation will provide an overview of where AI is already being used and prognosis for the immediate future.

About V. “Juggy” Jagannathan, PhD
V. “Juggy” Jagannathan, PhD, researches artificial intelligence and computer science for 3M Health Information Systems, developing new natural language and deep learning technologies that automatically structure patient-physician conversations into clinical documents. As an AI researcher with over four decades in the field, Juggy’s background includes both industry technology development and academia. The company he founded in 1996 with technology from West Virginia University (WVU) was acquired by M*Modal in 2003 and he has been with the company ever since. He is also an adjunct professor teaching Computer Science and Natural Language Processing (NLP) at WVU. Juggy is a tennis nut, golf nut and a yoga nut. In short, a nut.