Past events

The UMN Machine Learning Seminar Series brings together faculty, students, and local industrial partners who are interested in the theoretical, computational, and applied aspects of machine learning, to pose problems, exchange ideas, and foster collaborations. The talks are every Wednesday from 12 p.m. - 1 p.m. during the Spring 2022 semester.

This week's speaker, Qizhi He (Department of Civil, Environmental, and Geo- Engineering of UMN), will be giving a talk titled "Machine Learning Enhanced Computational Mechanics for Materials Modeling".

Abstract

It remains challenging in constitutive modeling and large-scale simulation due to the inherent complexities of materials such as inelasticity and heterogeneities. This talk will survey our recent research on developing hybrid computational methods by combining physics-based models, data-driven machine learning techniques, and mesh-free schemes to address various difficulties in computational mechanics arising from material characterization, modeling, and reduction.

First, I will introduce a physics-constrained data-driven modeling framework, which enables predictive physical simulation directly from material data without the employment of phenomenological constitutive models. A new approach inspired by manifold learning is formulated under the Galerkin meshfree methods for modeling nonlinear solids and biological tissues. Second, I will present how to use manifold learning together with sparse sampling to construct effective low-dimensional models of nonlinear mechanics systems. This hyper-reduction method has been applied to fast prediction of thermal fatigue behaviors of electronic packages. Lastly, I will discuss our recent work on developing physics-informed deep learning framework for discovering and identifying hidden constitutive models with a specific application to subsurface flow and transport in heterogeneous porous media. The goal of this study is to develop a computational framework that allows a seamless fusion of measurement data and the information from widely accepted physical laws.

Biography

Qizhi He is an Assistant Professor in the Department of Civil, Environmental, and Geo- Engineering at the University of Minnesota. He received his M.A. in applied mathematics and Ph.D. in structural engineering and computational science from UC San Diego, in 2016 and 2018, respectively. Afterwards, he was a Postdoctoral Research Associate at Pacific Northwest National Laboratory (PNNL), where he developed scientific machine learning methods for modeling flow and transport processes in porous media. His current research interests lie at the intersection of computational mechanics, materials modeling, and data-driven computing, with a focus on advancing data-driven machine learning enabled computational tools to predict mechanics of complex multiphysical processes and improve our fundamental understanding of multiscale materials and structures in engineered and natural systems.

Prospective students can RSVP for an information session to learn about the following graduate programs:

• Computer Science M.S.
• Computer Science MCS
• Computer Science Ph.D.
• Data Science M.S.
• Data Science Post-Baccalaureate Certificate

During the information session, we will go over the following:

• Requirements (general)
• Applying
• Prerequisite requirements
• What makes a strong applicant
• Funding
• Resources
• Common questions
• Questions from attendees

The last day to receive a 50% tuition refund for canceling full semester classes is Monday, February 7.

View the full academic schedule on One Stop.
 

The computer science colloquium takes place on Mondays and Fridays from 11:15 a.m. - 12:15 p.m.

This week's speaker, Lu Cheng (Arizona State University), will be giving a talk titled "Algorithmic Solutions for Socially Responsible AI".

Abstract

Artificial intelligence (AI) technologies have become increasingly pervasive, offering both promises and perils. In response, researchers and organizations have been working to publish principles for the responsible use of AI. To bridge from these principles to responsible AI practice, in this talk, I will introduce a framework of Socially Responsible AI (SRAI) that encompasses an interdisciplinary definition of AI's social responsibility, a pyramid that outlines four specific AI responsibilities in a hierarchy, and three human-centered operations -- to Protect, Inform, and Prevent. To materialize AI for good, for example, the protecting dimension aims to cover or shield humans from harm, injury, and negative impact; the informing dimension aims to deliver the facts or information in a timely way; the preventing dimension aims to prevent/mitigate the negative impact of AI algorithms. I will use representative tasks in social media mining to illustrate each of the three operations. In particular, I will introduce my recent work in session-based cyberbullying detection, causal understanding of disinformation dissemination, and unintended bias mitigation to elucidate how SRAI can protect and inform users, as well as mitigate societal harms. Finally, I will touch upon future work with themes focused on advancing AI algorithms and AI for social good. 

Biography

Lu Cheng is a fifth-year Ph.D. candidate in the School of Computing and Augmented Intelligence (SCAI) at  Arizona State University (ASU). Advised by Prof. Huan Liu, Lu's research focuses on bridging from conceptual AI principles to responsible AI practice using both statistical and causality-aware methods. Lu's work has appeared in and been invited to top venues for AI (e.g., AAAI, IJCAI), data mining (e.g., KDD, WWW, WSDM), and NLP (ACL). She is the web chair of WSDM'22 and senior program committee member of AAAI'22. Lu was the recipient of the 2021 ASU Engineering Dean's Dissertation Award, 2020 ASU Graduate Outstanding Research Award, 2021 ASU CIDSE Doctoral Fellowship, 2019 ASU Grace Hopper Celebration Scholarship, IBM Ph.D. Social Good Fellowship, and Visa Research Scholarship. 

The computer science colloquium takes place on Mondays and Fridays from 11:15 a.m. - 12:15 p.m.

This week's speaker, Jiaoyang Li (University of Southern California), will be giving a talk titled "Intelligent Planning for Large-Scale Multi-Agent Coordination".

Abstract

There is no doubt that robots will play a crucial role in the future and need to work as a team in increasingly more complex applications. Advances in robotics have laid the hardware foundations for building large-scale multi-robot systems, such as for mobile robots and drones. But how to coordinate robots intelligently is a difficult problem: Exact methods, such as integer programming and A*, do not scale as the joint-state space increases exponentially with the number of robots, while heuristic methods, such as reactive methods and end-to-end learning, may lead to deadlocks or traffic congestion. 

In this talk, I will introduce intelligent planning algorithms for solving this challenge with a focus on one fundamental problem: letting a large team of agents navigate without collisions in congested environments while minimizing their travel times. I will present principled planning algorithms that can efficiently coordinate hundreds of agents while providing rigorous guarantees on completeness and even optimality. I will also present techniques to apply these algorithms to real-world problems with robustness guarantees, such as warehouse robot coordination, multi-robot motion planning, railway planning, and airport surface operation.

Biography

Jiaoyang Li is a Ph.D. candidate in the Department of Computer Science at the University of Southern California (USC). Her research lies in the intersection of artificial intelligence, robotics, and optimization, with a focus on coordinating large teams of autonomous agents to accomplish collaborative tasks intelligently. She has published her research at top venues for AI planning and multi-agent systems (such as AAAI, IJCAI, ICAPS, and AAMAS) and received press coverage for it. She has also received a Technology Commercialization Award from the USC Stevens Center for Innovation in 2018, an Outstanding Student Paper Award from ICAPS’20, a Best System Demonstration Award from ICAPS’21, and a Best Research Assistant Award from USC in 2021. She recently led a team that won the NeurIPS’20 Flatland Challenge on railway planning and was selected to participate in Rising Stars in EECS 2021. More information can be found on her webpage https://jiaoyangli.me/

The UMN Machine Learning Seminar Series brings together faculty, students, and local industrial partners who are interested in the theoretical, computational, and applied aspects of machine learning, to pose problems, exchange ideas, and foster collaborations. The talks are every Wednesday from 12 p.m. - 1 p.m. during the Spring 2022 semester.

This week's speaker,  Dr. Alberto Bietti (NYU Center for Data Science), will be giving a talk titled "Benefits of Convolutional Models".

Abstract

Many supervised learning problems involve high-dimensional data such as images, text, or graphs. In order to make efficient use of data, it is often useful to leverage priors in the problem at hand, such as invariance to certain transformations or stability to small deformations. Empirically, deep convolutional architectures have been very successful on such problems, raising the question of how they are able to capture the structure of these problems for efficient learning.
I study this question from a theoretical perspective using kernel methods, in particular convolutional kernels, which are constructed following similar architectural principles, and provide good empirical performance on standard vision benchmarks such as Cifar10. I will present three contributions that highlight the benefits of (deep) convolutional architectures in terms of stability to deformations and sample complexity.

Biography

Alberto Bietti received his PhD in 2019 from Inria Grenoble, where he worked under the supervision of Julien Mairal. He was a postdoc at Inria Paris in 2020 and is currently a Faculty Fellow at the NYU Center for Data Science. His main research focus is on the theoretical foundations of deep learning, particularly through the lens of kernel methods.

The last day to apply for spring undergraduate graduation is Tuesday, February 1.

View the full academic schedule on One Stop.
 

The last day to cancel full semester classes and not receive a "W" is Monday, January 31. This is also the last day to receive a 75% tuition refund for canceling full semester classes.

In addition, this is the last day to add classes without college approval and to change grade basis (A-F or S/N) for full semester classes.

View the full academic schedule on One Stop.
 

The computer science colloquium takes place on Mondays from 11:15 a.m. - 12:15 p.m.

This week's speaker, University of Minnesota's Department of Computer Science Chair and Professor, Mats Heimdahl, will be giving a talk titled "Requirements, models, properties, and analysis: How do they all fit together?"

Abstract

Getting the system requirements right in a development project is crucial for success. One highly promising approach to rigorous requirements capture and definition is modeling of the requirements in formal notations. In such Model-Based Requirements Engineering, an initial set of natural language requirements forms the basis for an initial behavioral model of the intended system behavior as well as the basis for an initial formalization of the natural language requirements into formal requirements properties. Formal verification now allow formal verification techniques to be used to analyze the set of requirements properties as well as the behavioral models. For example, the set of requirements properties can be checked for consistency and the behavioral model can be verified against the formalized requirements properties. The
results from this analysis can then be used in an iterative requirements validation process where the analysis results serve as a basis for the modification, refinement, and extension of the set of requirements and/or the behavioral models to bring them in conformance with the truly desired (or notional) system requirements.

Over 20 years of research (and many case studies and development efforts), we have accumulated a wealth of experiences in model-based development and requirements engineering that provide the foundation for the research covered in this talk. The goal of the presentation is threefold. First, we aim to clarify the often confused relationships between natural language requirements, formal behavioral models, and formally captured requirements. Second, we will describe how one can use verification techniques to ensure that (1) the model indeed behaves as stated in the requirements,(2) that the requirements are the “true” requirements needed to meet the system objectives, and (3) that the analysis results can be used to better understand the characteristics of the system being developed. Finally, we want to spend some time opportunities for future research.

Biography

Mats Heimdahl is the Department Head and a Professor of Computer Science & Engineering at the University of Minnesota. He earned an M.S. in Computer Science and Engineering from the Royal Institute of Technology (KTH) in Stockholm, Sweden and a Ph.D. in Information and Computer Science from the University of California at Irvine.

His research interests are in software engineering, safety critical systems, software safety, testing, requirements engineering, formal specification languages, and automated analysis of specifications.

Paritosh Desai (Google Inc.)

Registration is required to access the Zoom webinar.

While there are many commonalities between academic research and roles in the industry for applied math professionals, there are also important differences. These differences are material in shaping career outcomes in the industry and we try to elaborate on them by focusing on two broad themes for people with academic research backgrounds. First, we will look at the common patterns related to applied AI/ML problems across multiple industries and specific challenges around them. Second, we will discuss emergent requirements for success in the industry setting. We will share principles and anecdotes related to data, software engineering practices, and empirical research based upon industry experiences.