Abstract:

While the typical behaviors of stochastic systems are often deceptively oblivious to the tail distributions of the underlying uncertainties, the ways rare events arise are vastly different depending on whether the underlying tail distributions are light-tailed or heavy-tailed. In light-tailed settings, a system-wide rare event arises because every component of the system subtly deviates from its nominal behavior as if the entire system has conspired to provoke the rare event (conspiracy principle), whereas, in heavy-tailed settings, a system-wide rare event arises because a small number of components fail catastrophically (catastrophe principle). In the first part of this talk, I will introduce the recent developments in the theory of large deviations for heavy-tailed stochastic processes at the sample path level and rigorously characterize the catastrophe principle for such processes.

The empirical success of deep learning is often attributed to the mysterious ability of stochastic gradient descents (SGDs) to avoid sharp local minima in the loss landscape, as sharp minima are believed to lead to poor generalization. To unravel this mystery and potentially further enhance such capability of SGDs, it is imperative to go beyond the traditional local convergence analysis and obtain a comprehensive understanding of SGDs' global dynamics in complex non-convex loss landscapes. In the second part of this talk, I will characterize the global dynamics of SGDs building on the heavy-tailed large deviations and local stability framework developed in the first part. This leads to heavy-tailed counterparts of the classical Freidlin-Wentzell and Eyring-Kramers theories. Moreover, we reveal a fascinating phenomenon in deep learning: by injecting and then truncating heavy-tailed noises during the training phase, SGD can almost completely avoid sharp minima, resulting in improved generalization performance for test data.

Bio:

Chang-Han Rhee is an Assistant Professor in Industrial Engineering and Management Sciences at Northwestern University. Before joining Northwestern University, he was a postdoctoral researcher at Centrum Wiskunde & Informatica and Georgia Tech. He received his Ph.D. from Stanford University. His research interests include applied probability, stochastic simulation, experimental design, and the theoretical foundation of machine learning. His research has been recognized with the 2016 INFORMS Simulation Society Outstanding Publication Award, the 2012 Winter Simulation Conference Best Student Paper Award, the 2023 INFORMS George Nicholson Student Paper Competition (2nd place), and the 2013 INFORMS George Nicholson Student Paper Competition (finalist). Since 2022, his research has been supported by the NSF CAREER Award.

Artificial Neural Networks (ANNs) are composed of very simple elements that implement Boolean logic. The basic model, the McCulloch-Pitts (MP) neuron, was introduced in 1943. This triggered a long effort to combine MP neurons into feed-forward networks whose connectivity could be adjusted to implement desired input-output maps. In 1986, this effort led to the back-propagation algorithm that underlies all current applications to deep learning. But this is not the topic of this talk.

We will concentrate on Recurrent Neural Networks (RNNs) whose lateral connectivity generates N-dimensional dynamical systems. It was networks of this type that were studied by Hopfield in his 1982 and 1984 papers, and by Ackley, Hinton, and Sejnowski in their 1985 paper, and led to the 2024 Nobel Prize in Physics for Hopfield and Hinton.

The dynamics of these networks are controlled by a Hamiltonian that is fundamentally similar to that of a disordered Ising model: the Sherrington-Kirkpatrick (SK) spin glass. Statistical physicists working in models of the SK type used all the tools available to them – replicas, Langevin equations, path integrals – to analyze these novel models that exhibited novel behaviors: memory storage, memory retrieval, constraint satisfaction, generative models.

The fundamental idea behind the development of both layered and recurrent ANNs is that of ‘connectionism’: these networks store their long-term knowledge of the task they have been trained to perform as the strengths of the connections between simple ‘neural’ processing elements.

Sara Solla, Professor / Joint with Department of Neuroscience, Northwestern University

Host: Michelle Driscoll

Learn how traditional 3D graphics engineering has converged with Generative AI and Robotics using OpenUSD and NVIDIA Omniverse.

I’ll discuss my career path from NU, to grad school, to Oscar nominated and Emmy winning Visual Effects, to my current role at NVIDIA.

This speaker will present via Zoom, but an in-person session will be held for the presentation.

Register Now

CDL is proud to welcome Dr. BalaKrishna Kolluru (Cerence, Inc) for his seminar "Generative AI: from chat2act"

Abstract:

Generative AI has some huge models: 1.76 trillion parameters, 300 billion and 2 billion; some of them transcend the traditional boundaries of exclusively text, speech and vision. The question to ask is how do we ride this crest to solve the real issues and the follow up with where is this innovation heading? In particular, I would like to discuss some pertinent problems that we encounter when productising research and discuss where the most impact could be felt.

Speaker Bio:

BalaKrishna Kolluru is a senior director of generative AI at Cerence, an automotive AI company. He has over 20 years of experience in the field of machine learning. In particular, his career spanned over large vocabulary speech recognition, embedded speech recognition, natural language processing, speech synthesis, computer vision, cybersecurity and signal processing. He obtained his PhD in Computer Science (speech processing) from University of Sheffield in 2006 and his fellowship at University of Manchester. He has a keen affinity towards productisation of research and therefore you see him associated with startups, incubators and investors. He is currently serving as an advisor for a handful of them (mostly 2-10 person companies). He relishes analogue photography and is deeply into numismatics. Bala is currently leading a generative AI project that aims to revolutionize the driving experience for next-gen cars.

Related Info

Register Now

This dissertation proposes methods for verifying that machine learning models are trustworthy, while keeping sensitive information (such as training data and model parameters) confidential from the verifier via cryptographic techniques. Several previous methods from academic literature are capable of producing models that are trustworthy (i.e. satisfying mathematical metrics of unbiasedness, reliability, privacy, etc.), but in practice violations of user trust from service providers are common. I argue that this disjuncture occurs in part because service providers are materially disincentivized from trustworthy behavior. This problem is compounded by the fact that most machine learning services are provided in a “black-box” model in order to protect intellectual property, which makes it difficult to assess model trustworthiness and thus hold service providers accountable for breaches in trust.

Here we take up the task of ensuring the use of trustworthy models in practice. We do so by designing zero-knowledge proof and secure multiparty computation protocols which verify trustworthiness via controlled releases of information about machine learning models to external parties, leaving the black box intact. Both methods do so by computing a specified set of operations on hidden data with provable correctness and confidentiality even in the presence of adversarial parties. We devote much attention to tailoring our protocols for concrete efficiency, as the computational
overhead of the cryptography we employ would otherwise impose a substantial practical barrier to the use of our methods. The main components of the work are as follows:
1. We design a zero-knowledge proof protocol that verifies the fairness of a decision tree model relative to a set of training data. We design our own “crypto-friendly” decision tree training algorithm, whose assumptions enable highly efficient zero-knowledge proofs of fairness.
2. We expand the scope of zero-knowledge proofs of fairness dramatically by designing a protocol which is modular to model type and training algorithm, and has improved security guarantees. We utilize an efficient probabilistic auditing protocol which enables practical scalability even for neural networks with tens of millions of parameters.
3. We provide a secure multiparty computation protocol that enables many parties to collaboratively train a machine learning model with verified confidentiality and robustness to unhelpful or poisonous training data. We formalize and exploit properties arising from the intersection of Byzantine robust aggregation algorithms and secure computation to make our methods concretely efficient.

Related Info

The increasing availability of large-scale datasets that trace the entirety of the scientific enterprise, have created an unprecedented opportunity to explore scientific production and reward. Parallel developments in data science, network science, and artificial intelligence offer us powerful tools and techniques to make sense of these millions of data points. Together, they tell a complex yet insightful story about how scientific careers unfold, how collaborations contribute to discovery, and how scientific progress emerges through a combination of multiple interconnected factors. These opportunities--and challenges that come with them--have fueled the emergence of multidisciplinary community of scientists that are united by their goals of understanding science and innovation. These practitioners of the science of science use the scientific methods to study themselves, examine projects that work as well as those that fail, quantify the patterns that characterize discovery and invention, and offer lessons to improve science as a whole. In this talk, I’ll highlight some examples of research in this area, hoping to illustrate the promise of science of science as well as it's limitations.

Dashun Wang, Professor of Management & Organizations and Professor of Industrial Engineering & Management Sciences, Northwestern University

Host: John Marko

I will discuss how large cosmic surveys are providing new insights into our understanding of the dark universe, with emphasis on recent results from the Dark Energy Survey. In particular, I will discuss recent observational hints that dark energy may not be a cosmological constant but rather a dynamical phenomenon, consistent with models introduced about 30 years ago. These ideas will be tested in near-future surveys such as the Vera Rubin Observatory Legacy Survey of Space and Time. Along the way, I will illustrate how machine learning methods are transforming the way we analyze large surveys and end by describing how the new SkAI Institute for AI in Astronomy will drive innovations and discoveries in both AI and Astronomy.

Speaker: Joshua Frieman, University of Chicago

Host: Tjitske Starkenburg

Related Info

The rapid advancement in Large Language Models has led to a proliferation of model options. It can be daunting to sort through variations in model creators, versions, sizes, and resource requirements. Fear not – we have you covered. This workshop will walk you through the basics of how to choose a model for your project, including aspects such as what type of data you can give to the model, how many words and other characters you can input, and which tasks the model is optimized for.

Prerequisites: Some familiarity with Python and machine learning will be useful for this workshop, but you don’t need to be an expert in either.

The Artificial Intelligence for Research (AIR) workshop series provides practical advice and examples on how to effectively and securely utilize AI in your research. From writing code to training Large Language Models (LLMs), our Data Scientists will walk you through the tools and tips you'll need.

This workshop series complements our other data science services and support for researchers. All Northwestern researchers (student, staff, and faculty in all fields) can request free consultations with our Data Scientists and Statisticians to guide you through your data project, from considering the feasibility and choosing a method to troubleshooting the code and suggesting improvements. If you're looking for more extensive support, we partner on faculty-sponsored research through our project collaboration service.

Register Now

Speaker:

Guy Aridor, Assistant Professor of Marketing, Kellogg School of Management, Northwestern University

Title:

The Value of Belief Data in Online Recommendation Systems

Abstract:

Designing algorithmic recommendation systems on online platforms is simultaneously a data collection and an algorithmic problem, though most work has focused on the algorithmic aspects. In this talk, I’ll describe several recent papers that show the value of collecting data not just on consumption behavior, but also on pre-consumption attitudes — what consumers think about items they have not consumed. I’ll discuss how an economic model of user consumption choices in environments that incorporates these attitudes can rationalize empirical consumption patterns on a movie recommendation platform, MovieLens. We then test the assumptions of and the hypotheses generated by this model in a field experiment on MovieLens that collects such belief data in order to decompose the mechanisms that drive the effectiveness of recommendations. Finally, I’ll discuss a practical procedure and a resulting open-source dataset collected via this procedure that we implement on the MovieLens platform that allows for the collection of such data at scale that can enable the incorporation of such data into the design of recommendation systems.

Speaker Bio:

Guy Aridor is an Assistant Professor of Marketing at Northwestern Kellogg School of Management and a research affiliate at CESifo. His research employs tools from economics and quantitative marketing to investigate policy and antitrust issues in the digital economy, as well as the effects of new technologies on consumer behavior. He has a particular interest in consumer privacy, recommendation systems, and social media platforms. He holds a PhD in economics from Columbia University and a BA in pure/applied mathematics, computer science, and economics from Boston University.

Location:

In person: Chambers Hall, 600 Foster Street, Lower Level
Remote option: https://northwestern.zoom.us/j/99906995637
Passcode: NICO24

About the Speaker Series:

Wednesdays@NICO is a vibrant weekly seminar series focusing broadly on the topics of complex systems, data science and network science. It brings together attendees ranging from graduate students to senior faculty who span all of the schools across Northwestern, from applied math to sociology to biology and every discipline in-between. Please visit: https://bit.ly/WedatNICO for information on future speakers.

Related Info

Wednesday / CS Seminar
November 13th / 3:00 PM
Hybrid / Mudd 3514

Speaker
Adrian Perrig, ETH Zurich

Talk Title
Reaching Escape Velocity for Layer-3 Innovation: Deployability of a Next-generation Internet Architecture

Abstract

It appears nearly impossible to deploy a new Internet architecture
that innovates at Layer 3 of the networking stack, as the obstacles
seem insurmountable: billions of deployed devices, legacy network
infrastructure with hardware-based packet processing with long
replacement cycles, operating systems of a sprawling complexity, and a
diverse application landscape with millions of developers. As a new
Internet architecture seemingly needs support by all of these
stakeholders, fundamental innovation at the network layer appears
hopeless.

We identify dependency loops as a core barrier to the deployment of a
next-generation Internet architecture. We propose to break the
dependency loops with a virtuous cycle: the availability of
applications using the NGN will result in increasing amount of
traffic, encouraging more NSPs to deploy the NGN, fueling user demand,
inviting more applications to deploy. We postulate that 1 million
users with access to the NGN connectivity suffice to set the virtuous
cycle in motion.

The aim of this talk is to imbue hope for the deployment of a
next-generation Internet architectures. With the expanding real-world
deployment of the SCION secure network architecture, we show how a
next-generation education network can be established and connected to
the commercial network. Applications running on hosts in these
networks can immediately make use of the next-generation
infrastructure thanks to a bootstrapping service, even without OS
support. To provide sufficient incentives to applications to build in
SCION support, we present a path towards reaching 1 million hosts in
SCIONabled networks. On the path toward this vision, 12 R&D
institutions on 5 continents are now connected with native SCION
connectivity ("BGP free"), reaching an estimated 250'000 users /
hosts. We present several applications and use cases that can be used
across these institutions.

Biography
Adrian Perrig is a Professor at the Department of Computer Science at
ETH Zürich, Switzerland, where he leads the network security group. He
is also a Distinguished Fellow at CyLab, and an Adjunct Professor of
Electrical and Computer Engineering at Carnegie Mellon University.
From 2002 to 2012, he was a Professor of Electrical and Computer
Engineering, Engineering and Public Policy, and Computer Science
(courtesy) at Carnegie Mellon University. From 2007 to 2012, he served
as the technical director for Carnegie Mellon's Cybersecurity
Laboratory (CyLab). He earned his MS and PhD degrees in Computer
Science from Carnegie Mellon University, and spent three years during
his PhD at the University of California at Berkeley. He received his
BSc degree in Computer Engineering from EPFL. He is a recipient of the
ACM SIGSAC Outstanding Innovation Award. Adrian is an ACM and IEEE
Fellow. Adrian's research revolves around building secure systems --
in particular his group is working on the SCION secure Internet
architecture.

---
Zoom Link
Panopto Link
DEI Minute:

Quantum Computation and Statistics

Yazhen Wang, Department of Statistics, University of Wisconsin-Madison

Abstract: Quantum computation and quantum information are of great current interest across various fields, including computer science, mathematics and statistics, physical sciences and engineering. As the theory of quantum physics is fundamentally stochastic, quantum computation and quantum information are inherently infused with elements of randomness and uncertainty. Consequently, quantum algorithms are random in nature. This highlights the important role for statistics to play in the realm of quantum computation, which in turn offers great potential to revolutionize computational statistics. In this talk, I will provide an overview of quantum computation and statistics, covering the fundamental concepts and exploring quantum advantage along with the role of statistics and the implications for statistics.

Related Info

Speaker: Philip Dames, Associate Professor of Mechanical Engineering, Temple University

Presentation Title: Developing and Deploying Situational Awareness in Autonomous
Robotic Systems

Date and Time: Friday, November 15 at 12:00 PM CT

Location: Tech IEMS C211 and Zoom

Zoom Link: https://tinyurl.com/CRBSeminar
• NU-authenticated attendees will be automatically admitted. Others, please email amy.nedoss@northwestern.edu to be admitted from the waiting room.

Abstract:
Robotic systems must possess sufficient situational awareness in order to successfully operate in complex and dynamic real-world environments. In this talk, I will first describe how multi-target tracking (MTT) algorithms can provide mobile robots with this awareness. Next, I will discuss two key applications of MTT to mobile robotics. The first problem is distributed target search and tracking. To solve this, we develop a distributed MTT framework that scales to large teams and task assignment strategies that automatically balance the workload across a team. The second problem is autonomous navigation through dynamic social spaces filled with people. To solve this, we develop a novel neural network-based controller that takes as its input the target tracks from an MTT, unlike previous approaches which only rely on raw sensor data.

Bio:
Philip Dames is an Associate Professor of Mechanical Engineering at Temple University, where he directs the Temple Robotics and Artificial Intelligence Lab (TRAIL). Prior to joining Temple, he was a Postdoctoral Researcher in Electrical and Systems Engineering at the University of Pennsylvania. He received his PhD Mechanical Engineering and Applied Mechanics from the University of Pennsylvania in 2015 and his BS and MS degrees in Mechanical Engineering from Northwestern University in 2010. He is the recipient of an NSF CAREER award. His research aims to improve robots’ ability to operate in complex, real-world environments to address societal needs.

Members of the NITMB community are invited to join us for Research-In-Progress meetings, an informal venue for members of the NITMB to discuss ongoing and/or planned research. Ben Kuznets-Speck is a postdoctoral research fellow in the Goyal Lab at Northwestern University. Kuznets-Speck leverages and develops theory, simulation and machine learning techniques to work at the interface of statistical physics, biology and evolution. Learn more about Ben Kuznets-Speck’s research and engage in discussion with the NITMB community. Research-In-Progress talks take place on Fridays at 2pm at the NITMB office (875 N Michigan Ave., Suite 4010). Snacks and coffee will follow.

Related Info

Join us for an insightful seminar featuring Sergey Shebalov about the resurgence of Artificial Intelligence. We will delve into real-world use cases, the challenges and opportunities in AI adoption, and the impact of AI on the travel industry.

Lecture Summary: Artificial Intelligence is experiencing resurrection over the last several years. The number of use cases already implemented in practice and the amount of investment allocated to development and adoption of this technology indicates that this time it's here to stay. While several challenges still remain and some new ones are anticipated the path towards Al becoming as common as computer, smartphone or internet is clear. We will discuss this journey on the example of the travel industry. We'll consider several main applications of Al that already provided significant benefits, a typical process these applications go through from an idea to a full-scale adoption, and the skills required to support that process from the new generation of leaders, scientists, engineers and practitioners.

Sergey Shebalov is a VP of Data Science and Head of Research at Sabre. He leads the Sabre Labs team responsible for development and implementation of complex Al decision support systems. Sergey holds PhD in mathematics from University of Illinois and has two decades of experience in the travel industry IT. His area of expertise is intelligent retailing,

resource optimization and adoption of Al in practice.

Location: Krebs Classroom, McCormick Education Center, 2311 Campus Dr

EnCORE and IDEAL TRIPODS Institutes Collaboration website here

Monday, November 18 and Tuesday, November 19, 2024

The NSF TRIPODS Institutes—The Institute for Emerging CORE Methods for Data Science (EnCORE) at UCSD and The Institute for Data, Econometrics, Algorithms, and Learning (IDEAL) at Northwestern University—are co-hosting a workshop titled "Foundations of Fairness and Accountability." The event will take place from November 18-19, 2024 in a hybrid format at the Department of Computer Science at Northwestern University, Evanston, Illinois. We plan to follow up with a second workshop at early Spring at the EnCORE institute at the University of California San Diego.

The workshop will feature a blend of talks and interactive discussions, focusing on key topics related to fairness and accountability. The main areas of exploration include:

1. Fairness in Resource Allocation
2. Fairness in Clustering & Ranking
3. Fairness in Prediction
4. Socio-Technical Aspects of Fairness & Accountability
5. Accountability in Experiment Design, with an emphasis on Replicability
6. Applications of these concepts across fields like Law, AI, and Biology.

Related Info

Register Now

Monday / CS Seminar
November 18th / 12:00 PM
Hybrid / Mudd 3514

Speaker
Kyros Kutulakos, University of Toronto

Talk Title
The Ultimate Video Camera

Abstract
Over the past decade, advances in image sensor technologies have transformed the 2D and 3D imaging capabilities of our smartphones, cars, robots, drones, and scientific instruments. As these technologies continue to evolve, what new capabilities might they unlock? I will discuss one possible point of convergence---the ultimate video camera---which is enabled by emerging single-photon image sensors and photon-processing algorithms. We will explore the extreme imaging capabilities of this camera within the broader historical context of high-speed and low-light imaging systems, highlighting its potential to capture the physical world in entirely new ways.

Biography
Kyros is a Professor of Computer Science at the University of Toronto and an expert in computational imaging and computer vision. His research over the past decade has focused on combining programmable sensors, light sources, optics and algorithms to create cameras with unique capabilities---from seeing through scatter and looking around corners to capturing surfaces with complex material properties robustly in 3D. He is currently leading efforts to harness the full potential of technologies such as single-photon cameras and programmable-pixel image sensors, for applications in extreme computer vision and scientific imaging. Kyros is the recipient of an Alfred P. Sloan Fellowship, an NSF CAREER Award, and eight paper prizes at the computer vision field's top conferences, including the best paper award at ICCV 2023 and CVPR 2019.

Research/Interest Areas: Computer vision, computational imaging
---
Zoom: TBA
Panopto: TBA
DEI Minute: TBA

Speaker

Corrina Schlombs, History, Rochester Institute of Technology

Title

"Data Entry, Labor, and Gender: Office Automation in Capitalist and Socialist Economies"

Abstract

In 1949, MIT mathematician Norbert Wiener warned US labor leader Walther Reuther that, in the US capitalist economy, automation technologies would cause massive unemployment. But a closer look at labor changes from computing technologies reveals a more complex picture: electronic computing also required new manual routine labor for data entry. Data occurred on paper, such as checks, insurance contracts or phone notes, and before it could be processed by a computer, it needed to be transferred into a computer-legible format—often punch cards or tape. In my talk, I examine mid-twentieth century office automation in capitalist and socialist economies, with a focus on the East German financial sector. In an economy promising full employment and lacking sufficient numbers of workers, officials promoted automation technologies with the goal of releasing workers. However, computing technologies were implemented in ways that heavily drew on women’s labor for data entry. Investigating how questions of technological change, employment, labor, and identity played out in different economic contexts, the talk calls technological promises into question at a time when artificial intelligence technologies are (again) expected to uproot the balance between human and machine labor.

Biography

Dr. Schlombs’s research focuses on technology and capitalism in transatlantic relations. In her current book project, she investigates transatlantic transfers of productivity culture and technology in the two decades before and after World War II. Productivity, a statistical measure of output per worker, came to encapsulate the American economic system, and transatlantic debates about productivity called into question the notion of the capitalist West during the Cold War conflict.

This workshop covers the basics of SQL syntax required to query databases, particularly basic commands such as SELECT, LIMIT, OFFSET, WHERE, BETWEEN, IN, IS NULL, ORDER BY, DISTINCT, GROUP BY, HAVING, and CASE WHEN.

Prerequisites: None. No installations are required.

This workshop complements our other data science services and support for researchers. All Northwestern researchers (student, staff, and faculty in all fields) can request free consultations with our Data Scientists and Statisticians to guide you through your data project, from considering the feasibility and choosing a method to troubleshooting the code and suggesting improvements. If you're looking for more extensive support, we partner on faculty-sponsored research through our project collaboration service.

Register Now

Today’s AI models can do a lot of things. But how did they become so powerful—and in what ways might they reshape the economy? In this complimentary webinar from Kellogg Executive Education and Kellogg Insight, Sergio Rebelo will take us on a journey into the past and present of AI. A leading macroeconomist, he will also offer his perspective on how AI stands to impact society, jobs, and the broader economy.

Speaker bio: Sergio Rebelo is the MUFG Bank Distinguished Professor of International Finance at Kellogg, where he has published widely in leading economics journals. He has served as a consultant to the World Bank, the International Monetary Fund, the Board of Governors of the Federal Reserve System, and the European Central Bank. He’s also an award-winning teacher.

Related Info

Register Now

This workshop covers intermediate SQL syntax required to query databases, including joins, UNION, EXCEPT, INTERSECT, subqueries, common table expressions, and window functions.

Prerequisites: Basic SQL at the level of the Introduction to SQL workshop. No installations are required.

This workshop complements our other data science services and support for researchers. All Northwestern researchers (student, staff, and faculty in all fields) can request free consultations with our Data Scientists and Statisticians to guide you through your data project, from considering the feasibility and choosing a method to troubleshooting the code and suggesting improvements. If you're looking for more extensive support, we partner on faculty-sponsored research through our project collaboration service.

Register Now

QUALIFY FOR INNOVATIVE AND TECHNICAL JOBS AT TOP COMPANIES

With more companies using data, the demand for data scientists continues to soar. Register for our Master of Science in Machine Learning and Data Science online information session to learn how you can take the next step in your career as an effective, knowledgeable leader in a rapidly growing field.

Learn more or register

Related Info

Register Now

Soon to be released artificial intelligence programs are being developed to assist with patients’ examinations. AI has been developed to integrate radiographs, facial images, and intra-oral scanning into an avatar. The radiographs and images are combined to create a patient-like 3D avatar with enhanced diagnostics information and automated analysis, including but not limited to three cephalometric, Boltons, Schwarz Index, endodontic, caries, and periodontal (with inflammation index) analyses. Additionally, an airway analysis is also completed, along with automated nutritional and home hygiene instructions digitally delivered.

Most importantly, the digital Total Health history is completed before the visit and includes all pertinent information necessary to personalize the patient interview. Using headpieces worn by patients and dental personnel, the AI not only records the patient’s responses for informed consent and charting but also suggests additional information needed for the dental professional. The total health history questionnaire includes but is not limited to age, sex, ethnicity, zip code, weight, height (automatic BMI determination), A1C, frequency of medical/dental visits, sugar and alcohol exposures, smoking, exercise regimen, number of steps/stairs daily, medications, medical history, dental history, bruxism and airway questions, fatigue, snoring, and many other factors. This is analyzed in real time, along with the patient’s responses. Using a learning tree protocol, the AI suggests pertinent paths of inquiries, combining the avatar information with the responses and then listing the needed salivary testing. Because saliva is a liquid biopsy for total body health, the salivary sample may be submitted for specific laboratory testing of the patient’s most significant health risks: cardiovascular, cancer,

Featuring:
Mark L Cannon, DDS, MS
Professor Emeritus
Feinberg School of Medicine

Related Info

Register Now

Beyond 2D: Collective Cell Migration and Topological Data Analysis

ABSTRACT:

Epithelial cell migration is shaped by extracellular matrix architecture during development and disease, but the extracellular matrix is also dynamically remodeled by epithelial cells. Here, I present recent results from my group on the so-called epithelial-mesenchymal transition, whereby cells weaken cell-cell adhesions and gain a pro-invasive phenotype. First, we analyze the disorganization and dissemination of multicellular clusters cultured in 3D matrix, which exhibit both collective and individual invasion phenotypes with spatially non-uniform traction signatures. Second, we elucidate how larger multicellular spheroids transition from coordinated circumferential orbiting towards radial matrix invasion. Finally, we describe the use of topological barcodes for machine learning of tissue architecture based on spatial connectivity (i.e. persistent homology). We envision new insights into the dynamic reciprocity that emerges between cells and their physical microenvironment via mechanobiology and AI / ML approaches.

BIO:

Ian Wong is Associate Professor of Engineering, and of Pathology / Laboratory Medicine at Brown University. He engineers new miniaturized technologies based on biomaterials and microfluidics to investigate cancer cell invasion, drug resistance, and heterogeneity. He is also interested in the unconventional fabrication of bio and nano materials using self-assembly and 3D printing. He did his graduate work on the directed self-assembly of biomolecular materials with Nick Melosh, receiving a Ph.D. in Materials Science and Engineering from Stanford University. His postdoctoral training was with Mehmet Toner and Daniel Irimia at the Center for Engineering in Medicine at Massachusetts General Hospital. He has been recognized with an NSF Graduate Research Fellowship, a Damon Runyon Cancer Research Fellowship, the Brown University Pierrepont Award for Outstanding Advising and the Fain Engineering Faculty Research Award.

Related Info

Sara Solla is a Professor of Neuroscience and of Physics and Astronomy at Northwestern University. Solla’s research interests lie in the application of statistical mechanics to the analysis of complex systems. Her research has led her to the study of neural networks, and she has used spin-glass models to describe associative memory, worked on a statistical description of supervised learning, investigated the emergence of generalization abilities in adaptive systems, and studied the dynamics of incremental learning algorithms.

Learn more about Sara Solla's research

The NSF-Simons National Institute for Theory and Mathematics in Biology Seminar Series aims to bring together a mix of mathematicians and biologists to foster discussion and collaboration between the two fields. The seminar series will take place on Fridays from 10am - 11am at the NITMB offices in the John Hancock Center in downtown Chicago. There will be both an in-person and virtual component.

Structure-driven design of reinforcement learning algorithms: a tale of two estimators

Wenlong Mou, Assistant Professor of Statistical Sciences, University of Toronto

Abstract: Reinforcement learning (RL) offers a flexible framework for sequential decision-making in uncertain environments, and its success heavily depends on efficiently learning value functions. Over the years, a diverse range of RL algorithms has been proposed, but at their core, two foundational principles stand out: to solve the Bellman fixed-point equations (known as ``bootstrapping methods''), or to average the rollout rewards. Despite their success, finding the optimal trade-off between these principles in practical applications remains elusive. Current theoretical guarantees -- either worst-case or asymptotic -- often fall short of providing actionable insights.

In this talk, I will discuss recent advances in methods that optimally reconcile bootstrapping and rollout for policy evaluation. The bulk of this talk will focus on a new class of estimators that strikes an optimal balance between temporal difference learning and Monte Carlo methods. Through the statistical lens, I will highlight why the local structure of the underlying Markov chain determines the fundamental complexity for estimation, and how our estimator adapts to these structures. Extending this perspective to continuous-time RL, I will also explore how the elliptic structure of diffusion processes provides key insights for making algorithmic choices.

Related Info

The era of data-driven discovery is producing a wave of new science in high energy astrophysics, and the Chandra Source Catalog (CSC) is an effective tool that enables it. The treasure trove found in CSC datasets has propelled population studies, the search for high energy transients, and the characterization of accretion in luminous X-ray systems. But the CSC has also become a valuable tool in machine learning studies, as it provides an exquisite training set that relates the basic units of X-ray data -single photon detections from a source- to astrophysically relevant measurables such as the spectral parameters of acreeting binaries or the timescales of exotic cosmic explosions. In this talk, I will present an overview of how the X-ray community is using machine learning in combination with the CSC to produce new representations of X-ray datasets that improve of fundamental tasks of X-ray astronomy, such as the classification of sources, the inference of physical parameters, and the discovery of anomalies of astrophysical relevance. I will also provide a perspective of how the CSC -and Chandra data in general- represent a legacy dataset as we enter the era of foundation AI.

Rafael Martinez Galarza, Staff Astrophysicist, Harvard-Smithsonian Center for Astrophysics

Host: Tim Kovachy

NOVEMBER MEETING: Tuesday, November 26, 2024 at 5:20pm (US Central)

LOCATION:
In person: Chambers Hall, Lower Level
600 Foster Steet, Evanston Campus

AGENDA:
TBA

SPEAKERS:
TBA

DATA SCIENCE NIGHTS are monthly talks on data science techniques or applications, organized by Northwestern University graduate students and scholars. Aspiring, beginning, and advanced data scientists are welcome! For more information: http://bit.ly/nico-dsn

Related Info

Drawing on the Northwestern Engineering whole-brain philosophy and leadership in cognitive science, the Master of Science in Artificial Intelligence program would like to invite you to learn more at our upcoming webinar.

Take this opportunity to join Dr. Kristian Hammond, Professor of Computer Science and director of the MSAI program, as he discusses the complexities of this field, and how this newly offered program at Northwestern Engineering will prepare students for a career in artificial intelligence. At the end of the presentation, we will offer an open Q&A where you will be able to have your specific questions answered. You are also welcome to email your questions to us ahead of the session (msai@northwestern.edu).

Related Info

Register Now