Logo: University of Southern California

Events Calendar



Select a calendar:



Filter March Events by Event Type:



Events for March 25, 2022

  • Repeating EventGrammar Tutorials

    Fri, Mar 25, 2022 @ 10:00 AM - 12:00 PM

    Viterbi School of Engineering Student Affairs

    Workshops & Infosessions


    INDIVIDUAL GRAMMAR TUTORING FOR VITERBI UNDERGRADUATE AND GRADUATE STUDENTS

    Meet one-on-one with Viterbi faculty, build your grammar skills, and take your writing to the next level!

    Viterbi faculty from the Engineering in Society Program (formerly the Engineering Writing Program) will help you identify and correct recurring grammatical errors in your academic writing, cover letters, resumes, articles, presentations, and dissertations.
    Bring your work, and let's work together to clarify your great ideas!

    Contact helenhch@usc.edu with questions.




    Location: Zoom

    Audiences: Graduate and Undergraduate Students

    View All Dates

    Contact: Helen Choi

    OutlookiCal
  • CS Colloquium: Chuang Gan (MIT-IBM Watson AI Lab) - Neuro-Symbolic AI for Machine Intelligence

    Fri, Mar 25, 2022 @ 10:00 AM - 11:00 AM

    Computer Science

    Conferences, Lectures, & Seminars


    Speaker: Chuang Gan, MIT-IBM Watson AI Lab

    Talk Title: Neuro-Symbolic AI for Machine Intelligence

    Series: CS Colloquium

    Abstract: Machine intelligence is characterized by the ability to understand and reason about the world around us. While deep learning has excelled at pattern recognition tasks such as image classification and object recognition, it falls short of deriving the true understanding necessary for complex reasoning and physical interaction. In this talk, I will introduce a framework, neuro-symbolic AI, to reduce the gap between machine and human intelligence in terms of data efficiency, flexibility, and generalization. Our approach combines the ability of neural networks to extract patterns from data, symbolic programs to represent and reason from prior knowledge, and physics engines for inference and planning. Together, they form the basis of enabling machines to effectively reason about underlying objects and their associated dynamics as well as master new skills efficiently and flexibly.

    This lecture satisfies requirements for CSCI 591: Research Colloquium



    Biography: Chuang Gan is a principal research staff member at MIT-IBM Watson AI Lab. He is also a visiting research scientist at MIT, working closely with Prof. Antonio Torralba and Prof. Josh Tenenbaum. Before that, he completed his Ph.D. with the highest honor at Tsinghua University, supervised by Prof. Andrew Chi-Chih Yao. His research interests sit at the intersection of computer vision, machine learning, and cognitive science. His research works have been recognized by Microsoft Fellowship, Baidu Fellowship, and media coverage from BBC, WIRED, Forbes, and MIT Tech Review. He has served as an area chair of CVPR, ICCV, ECCV, ICML, ICLR, NeurIPS, ACL, and an associate editor of IEEE Transactions on Image Processing.

    Host: Ram Nevatia

    Location: online only

    Audiences: By invitation only.

    Contact: Assistant to CS chair

    OutlookiCal
  • ECE Seminar: Distributed Systems: Rigorous Theoretical Foundations Unlock Promising Gains

    Fri, Mar 25, 2022 @ 10:00 AM - 11:00 AM

    Ming Hsieh Department of Electrical and Computer Engineering

    Conferences, Lectures, & Seminars


    Speaker: Dr. Mohammad Ali Maddah-Ali, Research Scientist, Department of Electrical Engineering, Stanford University

    Talk Title: Distributed Systems: Rigorous Theoretical Foundations Unlock Promising Gains

    Abstract: Over the last twenty years, we have witnessed several revolutionary technologies, from communication networks to learning platforms to blockchains, that have profoundly changed our daily lives. Often, these platforms are modeled, designed, and operated based on intuition and folk wisdom. In this talk, we challenge some of those common beliefs. We show that by meticulously elaborating the key performance bottlenecks from first principles, we can propose counterintuitive solutions grounded in rigorous analysis that unlock considerable scaling gains in several areas:

    1) In wireless communications, the delay in acquiring channel information is a significant bottleneck in supporting multiple users at a time. Contrary to popular belief, we demonstrate that even completely outdated channel information can be used for interference management and enabling simultaneous communications, thus alleviating the bottleneck of channel training.

    2) In content delivery networks, folk wisdom design is to maximize the likelihood of serving a request from the local cache (hit rate); thus, the performance is bottlenecked by the size of an individual cache. We propose a fundamentally new approach with a gain that scales with the sum of the cache sizes in the network, rather than an individual cache size.

    3) In distributed learning, we demonstrate that training with combined data samples (i.e., erasure-coded samples), rather than raw samples, can significantly improve the reliability and convergence rate. Moreover, we highlight the surprising role of approximation theory in circumventing a major bottleneck in designing practical coded training procedures.

    We conclude with promising directions for further investigation: in particular, the challenges in adding decentralized trust and accountability to these systems, to place control over them back in the hands of individuals rather than big corporations.

    Biography: Mohammad Ali Maddah-Ali received the B.Sc. degree from the Isfahan University of Technology, the M.Sc. degree from the University of Tehran, and the Ph.D. degree from the Department of Electrical and Computer Engineering, University of Waterloo, Canada. From 2008 to 2010, he was a Postdoctoral Fellow in the Department of Electrical Engineering and Computer Sciences, University of California at Berkeley. From 2010 to 2020, he was working at Bell Labs, Holmdel, NJ, as a communication network research scientist. He also worked as a faculty member at the Department of Electrical Engineering, Sharif University of Technology. Currently, he is a research scientist at the Department of Electrical Engineering, Stanford University.

    Dr. Maddah-Ali is a recipient of several awards including the IEEE International Conference on Communications (ICC) Best Paper Award in 2014, the IEEE Communications Society and IEEE Information Theory Society Joint Paper Award in 2015, and the IEEE Information Theory Society Paper Award in 2016. He is currently serving as an associate editor for the IEEE Transactions on Information Theory and a lead editor for The IEEE Journal on Selected Areas in Information Theory.

    Host: Dr. Keith Chugg, chugg@usc.edu

    Webcast: https://usc.zoom.us/j/98149159985?pwd=cWFsVnRkZXRKcTlWYllMcy9Rempmdz09

    Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248

    WebCast Link: https://usc.zoom.us/j/98149159985?pwd=cWFsVnRkZXRKcTlWYllMcy9Rempmdz09

    Audiences: Everyone Is Invited

    Contact: Mayumi Thrasher

    OutlookiCal
  • Advanced Manufacturing Seminar

    Fri, Mar 25, 2022 @ 10:00 AM - 11:30 PM

    Aerospace and Mechanical Engineering

    Conferences, Lectures, & Seminars


    Speaker: Dr. Mostafa Bedewy, University of Pittsburgh

    Talk Title: Manufacturing for the Future: Carbon-Based Flexible Neural Interfaces

    Abstract: Abstract: Nanocarbons like graphene, carbon nanotubes (CNTs), and nanofibers are promising for various applications including advanced electronic devices, novel energy systems, and next-generation healthcare diagnostics. This is owing to the excellent physical, chemical and electrochemical properties arising from the ordered atomic structure, the hierarchical nanoscale morphology, and tunable chemistry of nanocarbons. In particular, high surface area carbon electrodes for biosensors and neural interfaces have consistently been shown to have superior performance when compared to state-of-the-art metal electrodes. Nevertheless, major manufacturing challenges still hinder our ability to scalably produce nanocarbon-based electrodes with tailored morphology and surface chemistry, especially on flexible substrates. Unlike different transfer technique of CVD-grown nanocarbons, this talk will focus on a unique bottom-up approach for directly growing different types of graphenic nanocarbons on polymer films by laser irradiation. The speaker will show how this direct-write process, often referred to as laser-induced graphene (LIG), can be controlled to produce spatially-varying morphologies and chemical compositions of LIG electrodes, by leveraging gradients of laser fluence. Moreover, a method will be introduced to control the heteroatom doping of these LIG electrodes based on controlling the molecular structure of the polymer being lased. Finally, a demonstration of these functional LIG electrodes as electrochemical biosensors will be presented for the detection of the neurotransmitter dopamine with nanomolar sensitivity.

    Biography: Dr. Mostafa Bedewy leads the NanoProduct Lab at the University of Pittsburgh. His research interests include carbon nanomaterials, laser processing, nanomanufacturing and micromanufactuing, chemical vapor deposition (CVD), and biology-assisted manufacturing. Dr. Bedewy received the Frontiers of Materials Award from the Minerals, Metals and Materials Society (TMS) in 2022, Outstanding Young Investigator Award from the Institute of Industrial and Systems Engineers Manufacturing and Design (IISE M&D) Division in 2020, Outstanding Young Manufacturing Engineer Award from the Society of Manufacturing Engineers (SME) in 2018, the Ralph E. Powe Junior Faculty Enhancement Award from the Oak Ridge Associated Universities (ORAU) in 2017, the Robert A. Meyer Award from the American Carbon Society in 2016, and many other prestigious awards.

    Host: Center for Advanced Manufacturing

    More Info: https://usc.zoom.us/webinar/register/WN_OMywkH2iRSmzYMtYVM-frQ

    Webcast: https://usc.zoom.us/webinar/register/WN_OMywkH2iRSmzYMtYVM-frQ

    WebCast Link: https://usc.zoom.us/webinar/register/WN_OMywkH2iRSmzYMtYVM-frQ

    Audiences: Everyone Is Invited

    Contact: Tessa Yao

    OutlookiCal
  • PhD Defense - Chaoyang He

    Fri, Mar 25, 2022 @ 11:00 AM - 12:30 PM

    Computer Science

    University Calendar


    Time: 11AM - 12:30PM, March 25th, 2022

    Committee Members: Salman Avestimehr (Chair), Mahdi Soltanolkotabi, Murali Annavaram, Ram Nevatia, Xiang Ren

    Zoom Link: https://usc.zoom.us/my/usc.chaoyanghe

    Title: Federated and Distributed Machine Learning at Scale: From Systems to Algorithms to Applications

    Abstract:
    Federated learning (FL) is a machine learning paradigm that many clients (e.g. mobile/IoT devices or whole organizations) collaboratively train a model under the orchestration of a central server (e.g., service provider), while keeping the training data decentralized. It has shown huge potential in mitigating many of the systemic privacy risks, regulatory restrictions, and communication costs resulting from traditional, centralized machine learning and data science approaches in healthcare, finance, smart city, autonomous driving, and the Internet of things. Though it is promising, landing FL into trustworthy data-centric AI infrastructure faces many realistic challenges from learning algorithms (e.g., data heterogeneity, label deficiency) and distributed systems (resource constraints, system heterogeneity, security, privacy, etc.), requiring interdisciplinary research in machine learning, distributed systems, and security/privacy. Driven by this goal, this thesis focuses on scaling federated and distributed machine learning end-to-end, from algorithms to systems to applications.

    In the first part, we focus on the design of the distributed system for federated and distributed machine learning. We propose FedML, a widely adopted open-source library for federated learning, and PipeTransformer, which leverages automated elastic pipelining for efficient distributed training of Transformer models. FedML supports three computing paradigms: on-device training using a federation of edge devices, distributed training in the cloud that supports exchanging of auxiliary information beyond just gradients, and single-machine simulation of a federated learning algorithm. FedML also promotes diverse algorithmic research with flexible and generic API design and comprehensive reference baseline implementations (optimizer, models, and datasets). In PipeTransformer, we design an adaptive on the fly freeze algorithm that can identify and freeze some layers gradually during training, and an elastic pipelining system that can dynamically allocate resources to train the remaining active layers. More specifically, PipeTransformer automatically excludes frozen layers from the pipeline, packs active layers into fewer GPUs, and forks more replicas to increase data-parallel width.

    In the second part, we propose a series of algorithms to scale up federated learning by breaking many aforementioned constraints, such as FedGKT, an edge-cloud collaborative training for resource-constrained clients, FedNAS, a method towards automation on invisible data via neural architecture search, SpreadGNN, effective training on decentralized topology, SSFL, tackling label deficiency via personalized self-supervision, and LightSecAgg, the lightweight and versatile secure aggregation. Most algorithms are compatible with each other. Specially, we unified all implementations under the FedML framework. Therefore, under the complex constraints of the real world, the orchestration of these algorithms has the potential to greatly enhance the scalability of federated learning.

    Finally, we further propose FedML Ecosystem, which is a family of open research libraries to facilitate federated learning research in diverse application domains. FedNLP (Natural Language Processing), FedCV (Computer Vision), FedGraphNN (Graph Neural Networks), and FedIoT (Internet of Things). Compared with TFF and LEAF, FedNLP and FedCV greatly enrich the diversity of data sets and learning tasks. FedNLP supports various popular task formulations in the NLP domain, such as text classification, sequence tagging, question answering, seq2seq generation, and language modeling. FedCV can help researchers evaluate the three most representative tasks: image classification, image segmentation, and object detection. Moreover, FedGraphNN is the first FL research platform for analyzing graph-structured data using Graph Neural Networks in a distributed computing manner, filling the gap between federated learning and the data mining field. Going beyond traditional AI applications, FedIoT further extends FL to perform in wireless communication (e.g., 5G) and mobile computing (e.g., embedded IoT devices such as Raspberry PI, smartphones running on Android OS).

    WebCast Link: https://usc.zoom.us/my/usc.chaoyanghe

    Audiences: Everyone Is Invited

    Contact: Lizsl De Leon

    OutlookiCal
  • Astani Civil and Environmental Engineering Seminar

    Fri, Mar 25, 2022 @ 12:30 PM - 01:30 PM

    Sonny Astani Department of Civil and Environmental Engineering

    Conferences, Lectures, & Seminars


    Speaker: Dr. Kristopher McNeill, Department of Environmental Systems Science, ETH Zurich

    Talk Title: An Environmental Chemist View of Biodegradable Plastics

    Abstract: Contamination of the environment with plastic is a long-recognized problem, but in recent years, there has been a remarkable increase in public attention and outcry regarding plastic pollution. The low cost and durability of plastic materials, which make them desirable for many applications, are the same factors that contribute to their accumulation. The low cost lowers the barrier to short- term and single use applications and the durability means that, once in the environment, these materials are highly persistent. On this latter point, there is a growing interest and market for non- persistent, biodegradable plastic materials, which could help the problem of accumulation of plastic in the environment. This presentation will focus on several key questions about these alternative biodegradable materials: How do we know that a material is really biodegrading instead of just breaking down into microplastic? How does the receiving environment affect biodegradability? Are there applications where biodegradable plastics are viable alternatives to conventional plastics? What are the challenges that we face from an environmental chemistry perspective?

    Biography: Prof. Kris McNeill received his B.A. in Chemistry from Reed College (Portland, Oregon) in 1992 and his Ph.D. in Chemistry from the University of California, Berkeley in 1997. At Berkeley, he was co-advised by Professors Robert Bergman and Richard Andersen. Following his PhD, he switched his research focus from organometallic chemistry to environmental chemistry. He was a postdoctoral researcher at MIT from 1997 to 1999 with Prof. Philip Gschwend in the department of Civil and Environmental Engineering. McNeill began his independent career as a faculty member at the University of Minnesota in the Department of Chemistry, holding ranks of Assistant Professor (2000-2006) and Associate Professor (2007-2009). In 2009, Kris McNeill joined the faculty of ETH Zurich, where he continues to apply physical organic chemistry to the study of environmental processes.

    Host: Dr. Daniel McCurry

    More Info: https://usc.zoom.us/j/93390473354 Meeting ID: 933 9047 3354 Passcode: 527888

    Location: Ray R. Irani Hall (RRI) - 421

    Audiences: Everyone Is Invited

    Contact: Evangeline Reyes

    OutlookiCal