Select a calendar:
Filter November Events by Event Type:
SUNMONTUEWEDTHUFRISAT
Events for November 30, 2023
-
PhD Thesis Defense - Yunhao Ge
Thu, Nov 30, 2023 @ 09:00 AM - 11:00 AM
Thomas Lord Department of Computer Science
University Calendar
PhD Thesis Defense - Yunhao Ge
Committee Members: Laurent Itti (chair), Yan Liu, Greg Ver Steeg, Nicolas Schweighofer
Title: Learning Controllable Data Generation for Scalable Model Training
Abstract: As machine learning models grow in complexity and power, the demands on training datasets surge correspondingly, necessitating both greater volume and enhanced quality. Harnessing real data, however, brings to the fore several challenges, including the hefty costs and sluggishness of human annotations—particularly in the fields of vision and robotics. Further obstacles include biases, spurious correlations, privacy concerns, and copyright constraints.In this talk, I will explore the potential of controllable automatic data generators as a solution to these data-related challenges. We will delve into harnessing learning techniques to control different data generation properties, culminating in photorealistic quality and significantly enhancing the training and performance of downstream models. Key insights include: ·
Methods to learn control over varying attributes, categories, distributions, and physical properties to bolster both 2D and 3D model training.
The transition of control from humans to downstream models, and how it paves the way for on-demand data generation, forging a symbiotic loop between the data generator and the downstream models.
A look ahead: The promise and challenges of generating intricate 3D and video data, underpinned by vision-language foundation models. We chart the frontier of controllable data generation and explore its vast potential in shaping the future of scalable model training.
Zoom Meeting ID: 222 662 0525Location: Hedco Neurosciences Building (HNB) - B15
Audiences: Everyone Is Invited
Contact: Melissa Ochoa
-
ECE Seminar: Safe Autonomous Systems through Neurosymbolic Reasoning
Thu, Nov 30, 2023 @ 10:00 AM - 11:00 AM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Dr. Jyotirmoy V. Deshmukh, Associate Professor, Thomas Lord Department of Computer Science, USC Viterbi School of Engineering
Talk Title: Safe Autonomous Systems through Neurosymbolic Reasoning
Abstract: Huge strides have made in the widespread adoption of autonomous and human-in-the-loop cyber-physical systems (CPS), partly fueled by dramatic improvements in learning-based techniques. An important aspect of many such CPS applications is that they are safety-critical; any undesirable behavior by such systems can cause serious harm to human lives or property. The formal methods community has been an advocate of using logic and automata as specifications for safety-critical CPSs, and the past few decades have seen significant strides in algorithms for their verification, testing, and automated synthesis. A new challenge now is the presence of learning-enabled components (LECs) in CPSs. In this talk, we will review some recent work on using logic and learning-based techniques to provide guarantees for CPS applications using LECs. Such techniques are neurosymbolic in nature; they rely on infusing symbolic knowledge in neural network-based learning algorithms, as well as using symbolic techniques to reason about such neural systems. We will discuss the applicability and scalability of these techniques to real-world systems, discussing some success stories, as well as lay out some of the challenge problems that would need to be solved.
Biography: Jyotirmoy V. Deshmukh (Jyo) is an Associate Professor in the Department of Computer Science at the University of Southern California, and the co-Director of the Center for Autonomy and AI. Before joining USC, Jyo worked as a Principal Research Engineer at Toyota R&D. He got his Ph.D. in Electrical and Computer Engineering from the University of Texas at Austin in 2010. He was the 2010-12 Computing Innovation Postdoctoral research Fellow at the University of Pennsylvania. He is the recipient of the 2021 NSF Career Award and the 2021 Amazon Research Award.
Host: Dr. Richard M. Leahy, leahy@usc.edu
Webcast: https://usc.zoom.us/j/93509653910?pwd=QjVaQUhPOWVHVHFibXE3VjRkRXN4dz09Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248
WebCast Link: https://usc.zoom.us/j/93509653910?pwd=QjVaQUhPOWVHVHFibXE3VjRkRXN4dz09
Audiences: Everyone Is Invited
Contact: Mayumi Thrasher
-
NL Seminar
Thu, Nov 30, 2023 @ 11:00 AM - 12:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Kawin Ethayarajh, Stanford University
Talk Title: Machine Learning with Human Fault-Tolerance
Abstract: REMINDER: This talk will be a live presentation only, it will not be recorded. Meeting hosts only admit guests that they know to the Zoom meeting. Hence, you’re highly encouraged to use your USC account to sign into Zoom. If you’re an outside visitor, please provide your: Full Name, Title and Name of Workplace to (nlg-seminar-host(at)isi.edu) beforehand so we’ll be aware of your attendance. Also, let us know if you plan to attend in-person or virtually. More Info for NL Seminars can be found at: https://nlg.isi.edu/nl-seminar/ In machine learning, we have long recognized the need to build systems that can tolerate hardware faults and software faults. In this talk, I propose the need for a third kind of fault-tolerance: human fault-tolerance. The methods used to develop, evaluate, and deploy machine learning systems today assume that the humans that build and use them are rational actors making highly-informed decisions based on consistent preferences—this is far from true in practice. We can address the failures of these assumptions by drawing from economics, a field that has long been aware of how unfounded beliefs about human behavior can go wrong. Specifically, I will cover how we can develop theoretically grounded tools that discover human mistakes, design algorithms and methods for robustly eliciting and incorporating human feedback, and implement end-to-end platforms that make ML and NLP more transparent and reproducible. This line of work has led to the creation of datasets, models, and platforms that have been widely adopted by industry giants like Amazon, Google, and Meta.
Biography: Kawin Ethayarajh is a 5th year PhD student at Stanford University, where he works on bringing human fault-tolerance to machine learning. His research draws from economics to make machine learning and NLP more robust to the irrational, inconsistent, and uninformed human decisions made at every step. His work has been supported by a Facebook Fellowship and an NSERC PGS-D, and he has received an Outstanding Paper Award at ICML 2022. He co-created the Stanford Human Preferences dataset and the Dynaboard platform (behind Dynabench).
Host: Jon May and Justin Cho
More Info: https://nlg.isi.edu/nl-seminar/
Webcast: https://usc.zoom.us/j/99484520082Location: Information Science Institute (ISI) - Virtual and ISI-Conf Rm#689
WebCast Link: https://usc.zoom.us/j/99484520082
Audiences: Everyone Is Invited
Contact: Pete Zamar
Event Link: https://nlg.isi.edu/nl-seminar/
-
Quantum Science & Technology Seminar - Chaitali Joshi, Thursday, Nov. 30th at 2pm in EEB 248
Thu, Nov 30, 2023 @ 02:00 PM - 03:30 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Chaitali Joshi, Google, Santa Barbara
Talk Title: A chiral light-matter interface with superconducting qubits
Series: Quantum Science & Technology Seminar Series
Abstract: Noise Improving qubit connectivity in quantum networks is crucial for distributed information processing, and for reducing resource overheads in certain error correction protocols. While superconducting circuits have shown great promise for large-scale quantum processors, controlling the flow of light in complex qubit networks has remained a challenge. In this talk, I will discuss our recent work on realizing nonreciprocal light-matter interactions in the microwave domain using a transmon qubit strongly coupled to a 1D waveguide. By modulating the atom-waveguide coupling using magnetic fields, we gain control over the direction of photon emission from the qubit, with the ratio of forward-to-backward coupling rates exceeding 100. I will discuss applications of this platform, including photon-mediated gates between distant qubits and the preparation of many-body dark states in chiral atom arrays. In the second part, I will discuss our exploratory work on using disordered superconducting materials for nonlinear devices suitable for quantum links operating in the millimeter-wave frequency regime. Work based on: Phys. Rev. X 13, 021039 (2023), Phys. Rev. Applied 18, 064088 (2022)
Biography: Chaitali is currently a quantum research scientist at Google Santa Barbara. Previously, she was an IQIM/AWS Postdoctoral scholar in Electrical Engineering at Caltech, where she worked on waveguide quantum electrodynamics with superconducting qubits. She obtained her PhD from Cornell University in 2020, where she worked on nonlinear and integrated photonics for time-frequency manipulation of quantum states of light.
Host: Quntao Zhang, Wade Hsu, Mengjie Yu, Jonathan Habif & Eli Levenson-Falk
More Information: Chaitali Joshi Flyer.pdf
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248
Audiences: Everyone Is Invited
Contact: Marilyn Poplawski
-
VLP Meditation and Mandalas De-stress Event
Thu, Nov 30, 2023 @ 05:00 PM - 07:00 PM
Viterbi School of Engineering Student Affairs
Student Activity
Relax and unwind before the end of the semester and stressful finals with the Viterbi Learning Program (VLP) through light channeling, meditation, coloring mandalas, and hot chocolate! There are special snacks for the attendees :)
Location: Ronald Tutor Hall of Engineering (RTH) - 206
Audiences: Everyone Is Invited
Contact: Alex Bronz
Event Link: https://cglink.me/2nB/r393852
