Logo: University of Southern California

Events Calendar



Select a calendar:



Filter April Events by Event Type:



Events for April 20, 2023

  • NL Seminar - Modular Language Models

    Thu, Apr 20, 2023 @ 11:00 AM - 12:00 PM

    Information Sciences Institute

    Conferences, Lectures, & Seminars


    Speaker: Suchin Gururangan, University of Washington, Univ. Of Washington

    Talk Title: Modular Language Models

    Series: NL Seminar

    Abstract: REMINDER:

    Meeting hosts only admit guests that they know to the Zoom meeting. Hence, you are highly encouraged to use your USC account to sign into Zoom.

    If you are an outside visitor, please inform us at nlg DASH seminar DASH host AT isi DOT edu beforehand so we will be aware of your attendance and let you in.

    Conventional language models (LMs) are trained densely: all parameters are updated with respect to all data. We argue that dense training leads to a variety of well-documented issues with LMs, including their prohibitive training cost and unreliable downstream behavior. We then introduce a new class of LMs that are fundamentally modular, where components (or experts) of the LM are specialized to distinct domains in the training corpus, and experts are conditionally updated based on the domain of the incoming document. We show how modularity addresses the limitations of dense training by enabling LMs that are rapidly customizable (with the ability to mix, add, or remove experts after training), embarrassingly parallel (requiring no communication between experts), and sparse (needing only a few experts active at a time for inference). Key to our proposal is exploring what constitutes the domains to which experts specialize, as well as reflecting on the data sources used to train LMs. Our new techniques chart a path towards collaborative LM development, where anyone can contribute and maintain experts at very modest computational cost.

    Biography: Suchin Gururangan is a 3rd year PhD candidate at the University of Washington, advised by Noah A. Smith and Luke Zettlemoyer. He was previously a visiting researcher at Meta AI, a pre doctoral resident at the Allen Institute for AI, and spent several years in industry as a data scientist. His research interests span many areas of NLP, currently he works on modular, sparse language models that are efficient to customize and scale. His work has received awards at ACL 2020 and 2021, and he is supported by the Bloomberg Data Science PhD Fellowship.

    Host: Jon May and Justin Cho

    More Info: https://nlg.isi.edu/nl-seminar/

    Webcast: https://www.youtube.com/watch?v=lWlVRGgwRK4

    Location: Information Science Institute (ISI) - Virtual and ISI-Conf Rm#689

    WebCast Link: https://www.youtube.com/watch?v=lWlVRGgwRK4

    Audiences: Everyone Is Invited

    Contact: Pete Zamar

    Event Link: https://nlg.isi.edu/nl-seminar/

    OutlookiCal
  • CS Colloquium: Alvaro Velasquez (DARPA) - Neuro-Symbolic Transfer and Optimization

    Thu, Apr 20, 2023 @ 11:00 AM - 12:00 PM

    Thomas Lord Department of Computer Science

    Conferences, Lectures, & Seminars


    Speaker: Alvaro Velasquez, DARPA

    Talk Title: Neuro-Symbolic Transfer and Optimization

    Series: CS Colloquium

    Abstract: Neuro-symbolic artificial intelligence (NSAI) has experienced a renaissance and gained much traction in recent years as a potential \"third wave\" of AI to follow the tremendously successful second wave underpinned by statistical deep learning. NSAI seeks the integration of neural learning systems and formal symbolic reasoning for more efficient, robust, and explainable AI. This integration has been successful in classification and reinforcement learning, among other areas, but its application to transfer learning and combinatorial optimization remains largely unexplored. In this talk, we will cover recent advancements for the integration of symbolic structures in transferring knowledge between agents in the context of reinforcement learning and planning for sequential decision-making. We will also explore the concept of dataless neural networks as a framework for integrating combinatorial optimization problems and learning models. We conclude with a vision for these areas and the technical challenges that follow.

    This lecture satisfies requirements for CSCI 591: Research Colloquium

    Biography: Alvaro Velasquez is a program manager in the Innovation Information Office (I2O) of the Defense Advanced Research Projects Agency (DARPA), where he currently leads programs on neuro-symbolic AI. Before that, Alvaro oversaw the machine intelligence portfolio of investments for the Information Directorate of the Air Force Research Laboratory (AFRL). Alvaro received his PhD in Computer Science from the University of Central Florida in 2018 and is a recipient of the distinguished paper award from AAAI, best paper and patent awards from AFRL, the National Science Foundation Graduate Research Fellowship Program (NSF GRFP) award, the University of Central Florida 30 Under 30 award. He has authored over 60 papers and two patents and serves as Associate Editor of the IEEE Transactions on Artificial Intelligence. His research has been funded by the Air Force Office of Scientific Research.

    Host: Jyo Deshmukh

    Location: Olin Hall of Engineering (OHE) - 132

    Audiences: Everyone Is Invited

    Contact: Assistant to CS chair

    OutlookiCal
  • PhD Thesis Defense - Naghmeh Zamani

    Thu, Apr 20, 2023 @ 11:00 AM - 01:00 PM

    Thomas Lord Department of Computer Science

    University Calendar


    PhD Thesis Defense - Naghmeh Zamani

    Title: Perception and Haptic Interface Design for Rendering Hardness and Stiffness

    Committee: Heather Culbertson, Jernej Barbic, Somil Bansal

    Abstract: In this talk, I will discuss the challenge of accurately rendering the sensations of hardness and stiffness in haptic applications, which is a critical problem for applications such as medical simulation that require accurate virtual hardness and stiffness replication. The first part of the talk will present a set of experiments to investigate human tactile perception sensitivity in tool-mediated systems. The second part will explore a new method for rendering hard objects using an encountered-type haptic display and augmented reality. The talk will evaluate how changing the hardness of the end-effector affects the user\'s perception of the interaction and proposes a dynamic end-effector for a more accurate and realistic simulation of hardness and stiffness. Furthermore, I will discuss the investigation of the underlying events on the skin during the interaction between a bare finger and the environment. The results suggest that the spectral content of vibration feedback is important mechanical information for surface hardness discrimination and natural material identification. The talk will provide insights and solutions to improve the accuracy and realism of haptic simulations for applications that require the perception of hardness and stiffness in virtual objects

    Location: Charles Lee Powell Hall (PHE) - 325

    Audiences: Everyone Is Invited

    Contact: Melissa Ochoa

    OutlookiCal
  • MHI ISSS Seminar - Prof. Yahya Tousi, Thursday, April 20th at 2pm in RTH 211

    Thu, Apr 20, 2023 @ 02:00 PM - 02:00 PM

    Ming Hsieh Department of Electrical and Computer Engineering

    Conferences, Lectures, & Seminars


    Speaker: Prof. Yahya Tousi, University of Minnesota, Twin Cities

    Talk Title: Toward energy-efficient and scalable mm-wave systems

    Series: Integrated Systems

    Abstract: The end of device scaling is the dawning of a new era in integrated circuit design. Today, there is a growing demand for energy-efficient systems in multi-sensor electric vehicles, UAVs, and distributed wireless pico-cells. This is while, the intrinsic performance of analog building blocks no longer scales with technology nodes. In this talk I will argue that in the absence of device-level scaling, rethinking the frontend architecture by modernizing the traditional hierarchical design can open the door to substantial improvements in hardware efficiency and scalability. I will present two examples to support this claim.
    In the first work we rethink digital processing in phase modulated radars by replacing it with a more efficient mixed analog processing scheme. The new system demonstrates more than an order of magnitude improvement in energy efficiency compared to traditional radar sensors. In the second work I introduce a nearest element phase monitoring architecture that overcomes the scalability challenges in traditional LO distribution schemes. Based on this new approach and for the first time, we implement a mm-wave phased array radiator with seamless multi-chip scalability. These two examples demonstrate how combining architectural and circuit-level innovations in this new era can lead to efficient and scalable mm-wave and THz systems.

    Biography: Yahya Tousi received his Ph.D. degree in 2012 from the Department of Electrical and Computer Engineering at Cornell University, Ithaca, NY. In 2014 he joined the IBM T. J. Watson Research Center at Yorktown Heights, NY to develop the next generation of mm-wave phased array transceivers for wireless communication systems, and since 2017 he has been with the ECE Department at the University of Minnesota, Twin Cities. His current research interests are in high performance integrated circuits and novel architectures for mm-wave and terahertz systems with applications in communication, sensing, and healthcare. Dr. Tousi is the co-recipient of ISSCC Lewis Award for Outstanding Paper, and the Journal of Solid-State Best Paper Award both in 2017, the DARPA Young Faculty Award in 2020 and the DARPA Director\'s Fellowship Award in 2022.

    Host: MHI - ISSS, Hashemi, Chen and Sideris

    More Info: Zoom Link/Code: Meeting ID: 950 2226 0136, Passcode: 325523

    More Information: FLYER_Tousi.pdf

    Location: Ronald Tutor Hall of Engineering (RTH) - 211

    Audiences: Everyone Is Invited

    Contact: Marilyn Poplawski

    Event Link: Zoom Link/Code: Meeting ID: 950 2226 0136, Passcode: 325523

    OutlookiCal