Select a calendar:
Filter January Events by Event Type:
SUNMONTUEWEDTHUFRISAT
Events for the 4th week of January
-
Viterbi - Resume & Cover Letter 1.0: What Recruiters Look For
Tue, Jan 21, 2025 @ 12:00 PM - 01:00 PM
Viterbi School of Engineering Career Connections
Workshops & Infosessions
This event is for Viterbi engineering students only. Non-Viterbi students will not be admitted. Please register on Handshake.
Wondering how to write a strong resume? Join Viterbi Career Connections for "Resume & Cover Letter 1.0: What Recruiters Look For," an introductory workshop designed to help you get started on crafting an effective resume and cover letter to ensure you stand out by reflecting on your greatest strengths, skills, and accomplishments. Find out what employers look for, review tips and samples, and explore different resume formats. Have your resume draft ready to take notes and make modifications in real-time!
In this informative session, you will:
Understand the fundamental components of a resume and cover letter, including formatting, structure, and content.
Identify key elements recruiters look for in resumes and cover letters.
Learn how to structure and write a cover letter that effectively introduces yourself and complements your resume.
Understand the best practices for formatting and structuring your documents to ensure they are Applicant Tracking Systems (ATS) and AI-friendly.
Learn how AI comes into play and how to utilize it effectively.
Walk away with the next steps to make your resume application-ready.
Location: Ronald Tutor Hall of Engineering (RTH) - 211
Audiences: All Viterbi
Contact: RTH 218 Viterbi Career Connections
Event Link: https://usc.joinhandshake.com/
-
Viterbi - Resume 2.0: How to Turn Applications into Interviews
Tue, Jan 21, 2025 @ 01:00 PM - 02:00 PM
Viterbi School of Engineering Career Connections
Workshops & Infosessions
This event is for Viterbi engineering majors only. Non-Viterbi students will not be admitted. Please register through Handshake.
Wondering what resumes get picked for interviews? Join Viterbi Career Connections for "Resume 2.0: How to Turn Applications into Interviews," a workshop designed to provide you with strategies and techniques to make your resume stand out. By the end of the session, you should feel more confident in creating a resume that captures recruiters' attention and conveys your qualifications for the positions you are applying for.
In this informative and interactive session, you will:
Understand the importance of tailoring your resume and cover letter to specific job descriptions and how this directly impacts interview invitations.
Learn how to write a cover letter highlighting your suitability for the position.
Understand the importance of how to align your skills and experiences with the job description.
Learn how AI comes into play and how you can utilize it effectively.
Identify and incorporate industry-specific keywords that will help your resume get noticed.
Gain insights into what makes an application stand out versus what can detract from your candidacy.
Location: Ronald Tutor Hall of Engineering (RTH) - 211
Audiences: All Viterbi
Contact: RTH 218 Viterbi Career Connections
Event Link: https://usc.joinhandshake.com/
-
How Hollywood AI Helps Creators Go Global
Tue, Jan 21, 2025 @ 03:00 PM - 04:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Hao Li , CEO/Co-Founder of Pinscreen
Talk Title: How Hollywood AI Helps Creators Go Global
Abstract: In this talk I will talk about our journey on bringing digital humans to life in Hollywood, and how we use AI to democratize technologies for generating and manipulating photorealisitc human faces. I will showcase some use cases of our technologies developed at my startup Pinscreen, and how top film studios and streaming companies have adopted our approach for AI-driven visual effects, including face swapping, actor de-aging, and facial reenactment. One particular scalable application consists of dubbing movies from foreign language into English, where lips are perfectly matched to the translated content using AI. I will show how we are moving from Hollywood adoption to building a consumer product: pindub.ai for any Internet content creator. I will also showcase some our recent research efforts at MBZUAI with focus on more advanced generative AI capabilities and metaverse application such as telepresence using the currently most advanced one-shot head reenactment system.
This lecture satisfies requirements for CSCI 591: Research Colloquium.
Biography: Hao Li is the CEO and Co-Founder of Pinscreen, a Los Angeles-based startup renowned for its pioneering generative AI technologies in visual effects, dubbing, and digital humans. He is also Professor at the Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) and the Director of the MBZUAI Metaverse Center. Previously, Hao was a Distinguished Fellow with the Computer Vision Group at UC Berkeley and an Associate Professor of Computer Science at the University of Southern California (USC), where he directed the USC Institute for Creative Technologies. He has also been a visiting professor at Weta Digital, a research lead at Industrial Light & Magic / Lucasfilm, and a postdoctoral fellow at Columbia and Princeton Universities. Hao’s research lies at the intersection of computer vision, computer graphics, and machine learning with a focus on photorealistic human digitization, real-time performance capture, immersive telepresence, and generative video synthesis and manipulation (deepfakes). He is widely recognized for his groundbreaking work in real-time 3D facial tracking, which has led to the technology behind Apple's Animoji, the digital recreation of Paul Walker in the movie Furious 7, and the first real-time deepfake face-swapping technology based on generative AI. His research on facial reenactment has been widely adopted in Hollywood, driving advancements in photorealistic digital doubles and seamless de-aging effects. Additionally, he is known for pioneering production-grade visual dubbing solutions, which have made possible the first fully AI-lip-synced foreign language films and TV shows dubbed into English. Hao and his company, Pinscreen, have earned credits in numerous high-profile motion pictures, including Fallout, Money Heist – Berlin, Slumberland, Free Guy, Blade Runner 2049, Valerian and the City of a Thousand Planets, Furious 7, and The Hobbit: The Battle of the Five Armies. His algorithms on deformable shape registration have not only advanced the field of data-driven facial animation and human body modeling, but also improved the radiation treatment for cancer patients all over the world through his collaboration with C-RAD on surface guided radiation therapy. Hao has received numerous accolades, including being named a Top 35 Innovator Under 35 by MIT Technology Review in 2013. he has been awarded the Google Faculty Award, the Okawa Foundation Research Grant, and the Andrew and Erna Viterbi Early Career Chair. In 2018, he won the Office of Naval Research (ONR) Young Investigator Award, and in 2019, he was named to the DARPA ISAT Study Group. He received the ACM Symposium on Computer Animation Best Paper Award in 2009, the ACM SIGGRAPH Real-Time Live! “Best in Show” Award in 2020, and an Epic Megagrant in 2021. Hao has been a speaker at the World Economic Forum in Davos (2020) and was featured in the inaugural season of Amazon’s documentary series re:MARS Luminaries in 2022. Hao earned his PhD from ETH Zurich and his MSc from the University of Karlsruhe (TH).
Host: Prof. Stefanos Nikolaidis
Location: Kaprielian Hall (KAP) - 267
Audiences: Everyone Is Invited
Contact: Thomas Lord Department of Computer Science
-
Epstein Institute, ISE 651 Seminar Class
Tue, Jan 21, 2025 @ 03:30 PM - 04:30 PM
Daniel J. Epstein Department of Industrial and Systems Engineering
Conferences, Lectures, & Seminars
Speaker: Dr. Salar Fattahi, Assistant Professor Industrial and Operations Engineering at the University of Michigan
Talk Title: Finding the needle in the haystack: How gradient descent converges to the correct solution in the face of many local minima
Host: Dr. Andres Gomez
More Information: FLYER ISE 651 Salar Fattahi 1.21.25.png
Location: Ethel Percy Andrus Gerontology Center (GER) - 206
Audiences: Everyone Is Invited
Contact: Casi Jones/ ISE
-
Viterbi - MyReaa Information Session
Tue, Jan 21, 2025 @ 06:30 PM - 07:30 PM
Viterbi School of Engineering Career Connections
Workshops & Infosessions
This event is for Viterbi engineering students only. Non-Viterbi students will not be admitted. Please register through Handshake.
We will be hosting an informative session about MyReaa, our cutting-edge AI platform revolutionizing real estate analytics. In addition to showcasing how AI is transforming the industry, we will discuss the journey of starting a business, share key life experiences, and highlight valuable lessons from both business and AI perspectives. This session will be interactive, offering students a chance to connect with the founder, Shant Samuelian, and the co-founder to gain insights into innovation, entrepreneurship, and career development in the intersection of AI and real estate.
Degree Levels: Bachelors, Masters
Majors: Computer Science, Computer Science/Business Adminstration, Data Science, Analytics, AI in Business
Sponsorship: No
Location: Ronald Tutor Hall of Engineering (RTH) - 211
Audiences: All Viterbi
Contact: RTH 218 Viterbi Career Connections
Event Link: https://usc.joinhandshake.com/
-
CS Colloquium: Corey Baker (USC / ECE) - Patient Centered Systems for Remote Patient Monitoring
Wed, Jan 22, 2025 @ 10:00 AM - 11:00 AM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Corey Baker, USC / ECE
Talk Title: Patient Centered Systems for Remote Patient Monitoring
Abstract: Reliance on Internet connectivity is detrimental where modern networking technology is lacking, power outages are frequent, or network connectivity is expensive, sparse, or non-existent (i.e., underserved urban communities, rural areas, natural disasters). Though there has been much research conducted around 5G and 6G serving as the conduit for connecting any and everything; scalability issues are a major concern and real-world deployments have been limited. Realization of the limitations resulting from reliance on Internet and cellular connectivity are prevalent in mHealth applications where remote patient monitoring has improved the timeliness of clinical decision making, decreased the length of hospital stays, and reduced mortality rates everywhere in the nation except in medically underserved and rural communities in the US like Appalachian Kentucky, where chronic disease is approximately 20% more prevalent than other areas. As an alternative, deploying resilient networking technology can facilitate the flow of information in resource-deprived environments to disseminate non-emergency, but life saving data. In addition, leveraging opportunistic communication can supplement cellular networks to assist with keeping communication channels open during high-use and extreme situations. This talk will discuss the pragmatic applications of designing opportunistic systems for particular entities (patients, citizens, etc.); specifically applied to healthcare and increasing patient adherence, permitting any community to become smart and connected while simultaneously keeping network connectivity costs to a minimum.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Corey E. Baker holds the position of Assistant Professor in the Ming Hsieh Department of Electrical and Computer Engineering at the University of Southern California (USC). Prior to joining USC, Baker served as an Assistant Professor in the Department of Computer Science at the University of Kentucky (UK) as well as an Application and Framework Engineer in Health and Research Products at Apple Inc where he worked on medical and research frameworks such as CareKit and ResearchKit. In his current role, Baker directs the Network Reconnaissance (NetRecon) Lab, where his research focuses on developing full-stack systems for distributing, protecting, and authenticating data in opportunistic networking scenarios. These scenarios encompass rural remote patient monitoring, smart cities, and natural disasters, with the ultimate goal of enhancing the well-being of individuals. Baker’s research also involves evaluating the practical applications of opportunistic delay-tolerant networks (DTNs), software-defined networks (SDNs), and human-centered design in empowering device-to-device (D2D) social networks for crowd-sourcing information. By leveraging opportunistic communication, Baker seeks to provide complementary solutions to traditional networks, which often rely on centralized infrastructures such as the Internet.
Host: CS Department
Location: Olin Hall of Engineering (OHE) - 132
Audiences: Everyone Is Invited
Contact: CS Faculty Affairs
-
Computer Science General Faculty Meeting
Wed, Jan 22, 2025 @ 12:00 PM - 02:00 PM
Thomas Lord Department of Computer Science
Receptions & Special Events
Bi-Weekly regular faculty meeting for invited full-time Computer Science faculty and staff only. Event details emailed directly to attendees.
Location: Ginsburg Hall (GCS) - 107
Audiences: Invited Faculty Only
Contact: Julia Mittenberg-Beirao
-
Photonics Seminar - Dr. Chenzi Guo, Wednesday, January 22nd at 1pm in EEB 248
Wed, Jan 22, 2025 @ 01:00 PM - 02:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Chenzi Guo, Deputy Head of Light Publishing Group
Talk Title: Journals eLight, Light: Science & Applications: for the Wellbeing of Optics Community
Series: Photonics Seminar Series
Abstract: Dr. Guo will talk about why eLight was launched, the significant visibility that eLight has received, and how eLight serves the community with high quality, speed and transparency. In additions, Dr. Guo will also talk about Light: Science & Applications, as well as Light: Advanced Manufacturing.
Biography: Dr. Chenzi Guo serves as the Deputy Head of the Light Publishing Group, which oversees nine scientific journals. She is the Director of eLight (elight.springeropen.com) and an Editor of Light: Science & Applications (nature.com/lsa). Dr. Guo is a committee member of the International Division of the China Editology Society of Science Periodicals and a fellow of the Youth Innovation Promotion Association of the Chinese Academy of Sciences. She has received many national and state-level awards and has co-authored over 30 publications, including articles in Nature, Light: Science & Applications, as well as three books.
Host: Mercedeh Khajavikhan and Demetri Christodoulides
More Information: ChenziGuo_flier_PDF.pdf
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248
Audiences: Everyone Is Invited
Contact: Marilyn Poplawski
-
PhD Dissertation Defense - Mehrnoosh Mirtaheri
Wed, Jan 22, 2025 @ 01:00 PM - 03:00 PM
Thomas Lord Department of Computer Science
University Calendar
Title: Toward Learning and Forecasting with Temporal Knowledge Graphs
Date and Time: Tuesday, January 22nd, 2025 - 1:00p - 3:00p
Location: SAL 213
Committee: Aram Galstyan (Chair), Emilio Ferrara (Tenured Faculty), Fred Morstatter, Antonio Ortega (External Faculty
Abstract: Temporal knowledge graphs (TKGs) model real-world relationships between entities over time, enabling insight extraction from unstructured data. While powerful for various applications, TKGs are inherently limited by incompleteness and noise, making their completion and forecasting crucial research areas.
This thesis tackles the key challenges in TKG forecasting: relation sparsity in large-scale graphs, continuous integration of new data while preserving existing knowledge, and entity evolution as new entities emerge and existing ones appear in novel contexts. Through novel methodological frameworks, this research demonstrates improved predictive accuracy, robustness to data sparsity, and adaptability to evolving data, validated through extensive evaluation on both standard benchmark and real-world datasets.
Zoom Link: https://usc.zoom.us/j/96220815599Location: Henry Salvatori Computer Science Center (SAL) - 213
Audiences: Everyone Is Invited
Contact: Mehrnoosh Mirtaheri
-
AME Seminar
Wed, Jan 22, 2025 @ 03:30 PM - 04:30 PM
Aerospace and Mechanical Engineering
Conferences, Lectures, & Seminars
Speaker: Placid M. Ferreira, University of Illinois
Talk Title: IONICS-BASED NANOMANUFACTURING
Abstract: Mechanics and transport at the micro- and nanoscale offer a rich set of controllable phenomena that can be exploited for the development of manufacturing processes compatible with these dimensional scales. Here, we exploit ionic transport in solids as the basis of highly controllable, efficient, high-resolution, high-throughput nanomanufacturing processes for producing metallic (specifically, silver and copper) nanostructures.
This talk will focus on the exploitation of the high room-temperature ionic conductivity in silver and copper-based superionic glasses as the basis of subtractive and additive nano-manufacturing processes such as superionic imprinting/stamping, roll patterning and direct writing. Taking a traditional manufacturing perspective, the talk will discuss tooling and tool materials; process characterization and rates; and tool wear with such processes. Applications and future directions towards photochemical patterning for ionics-based nanomanufacturing processes will also be discussed.
Biography: Placid M. Ferreira is the Tungchao Julia Lu Professor of Mechanical Science and Engineering at Illinois. From 2003 to 2009, he was the director of the Center for Chemical-Electrical-Mechanical Manufacturing Systems (Nano-CEMMS), an NSF-sponsored Nanoscale Science and Engineering Center after which he served as the Head of the Department of Mechanical Science and Engineering at Illinois until August 2015. He currently leads the Center for Networked Intelligent Components and Environments, a translational research collaboration between the University of Illinois and Foxconn Interconnect Technology.
Professor Ferreira graduated with a PhD in Industrial Engineering from Purdue University in 1987, M.Tech (Mechanical) from IIT Bombay, 1982 and B.E. (Mechanical) for University of Bombay in 1980. He has been on the mechanical engineering faculty at Illinois since 1987, serving as the associate head for graduate programs and research from 1999 to 2002. Professor Ferreira's research and teaching interests are in precision manufacturing and includes computer-controlled machines, nano-manufacturing and metrology. Professor Ferreira received NSF's Presidential Young Investigator Award in 1990, SME's Outstanding Young Investigator Award in 1991, University of Illinois' University Scholar Award in 1994, ASME’s Ennor Award for Manufacturing Technology in 2014. He is also a Fellow of ASME, SME and AAAS. He has served on the editorial board of a number of manufacturing-related journals
Host: AME Department
More Info: https://ame.usc.edu/seminars/
Webcast: https://usc.zoom.us/j/96060458816?pwd=8LmoG2q6vBCQubqqWpcizd2F1bxqsH.1Location: James H. Zumberge Hall Of Science (ZHS) - 252
WebCast Link: https://usc.zoom.us/j/96060458816?pwd=8LmoG2q6vBCQubqqWpcizd2F1bxqsH.1
Audiences: Everyone Is Invited
Contact: Tessa Yao
Event Link: https://ame.usc.edu/seminars/
-
NL Seminar-Inductive Biases for Data- and Parameter-Efficient Transfer Learning
Thu, Jan 23, 2025 @ 11:00 AM - 12:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Mozhdeh Gheini, USC/ISI, USC/ISI
Talk Title: Inductive Biases for Data- and Parameter-Efficient Transfer Learning
Abstract: THIS TALK WILL "NOT" BE RECORDED, PLEASE WATCH LIVE OR ATTEND IN PERSON. https://usc.zoom.us/j/95381979100?pwd=yKkC6snFuqRddSnRCEwnVWvtP9ZdCX.1 Meeting ID: 953 8197 9100Passcode: 911377 Data- and resource-intensive pre-training and fine-tuning applied upon Transformer-based models is the dominant paradigm at the forefront of rapid advancements in natural language processing, human language technologies, and most notably, large language models. Such reliance on massive amounts of data, computation, and energy, while effective and impressive from a performance-only perspective, can hinder open, nonexclusive, and sustainable development of these technologies. In this talk, we present how certain inductive biases can be devised to adjust current natural language methods under resource-constrained scenarios and provide insights into why the proposed inductive biases are successful in such cases. Specifically, we discuss four research directions on data and parameter efficiency of fine-tuning and transfer learning in natural language processing: (1) a universal regimen that creates a single pre-trained checkpoint suitable for machine translation transfer to practically any language pair and eliminates the need for ad hoc pre-training; (2) an architecture-guided parameter-efficient fine-tuning method that performs competitively with full fine-tuning while exclusively updating cross-attention parameters; (3) an analysis of Mega, a recently introduced augmentation of the Transformer architecture to incorporate explicit recency bias, through the lens of transfer learning; and (4) a meta-learning algorithm to prime pre-trained models for specific fine-tuning strategies. Combined with ablations that show how they are effective and analyses that demonstrate their generalizability, these directions are meant to serve as tools for resource-efficient transfer learning for natural language processing.
Biography: Mozhdeh "Mo" Gheini is a PhD candidate at the University of Southern California advised by Jonathan May. Her PhD focus has been on investigating different inductive biases to build data- and parameter-efficient methods for transfer learning for natural language processing tasks like machine translation and beyond. She has also spent three summers interning with Apple, where she will be joining again in February as Machine Learning Research Engineer.
Host: Jonathan May and Katy Felkner
More Info: https://www.isi.edu/events/5379/inductive-biases-for-data-and-parameter-efficient-transfer-learning/
Webcast: https://usc.zoom.us/j/95381979100?pwd=yKkC6snFuqRddSnRCEwnVWvtP9ZdCX.1Location: Information Science Institute (ISI) - Virtual and ISI-Conf Rm#689
WebCast Link: https://usc.zoom.us/j/95381979100?pwd=yKkC6snFuqRddSnRCEwnVWvtP9ZdCX.1
Audiences: Everyone Is Invited
Contact: Pete Zamar
Event Link: https://www.isi.edu/events/5379/inductive-biases-for-data-and-parameter-efficient-transfer-learning/
-
Alfred E.Mann Department of Biomedical Engineering - Seminar series
Fri, Jan 24, 2025 @ 11:00 AM - 12:00 PM
Alfred E. Mann Department of Biomedical Engineering
Conferences, Lectures, & Seminars
Speaker: Wyatt Shields, Ph.D., Assistant Professor in the Department of Chemical and Biological Engineering University of Colorado Boulder
Talk Title: Programmable Microrobots for Biomedicine
Abstract: Colloidal particles are often used as building blocks for generating hierarchical structures with useful capabilities at small scales. However, the capabilities of such structures often depend on the physical properties of the particles. My research group is interested in broadening the complexity of microparticle designs, giving rise to distinctive behaviors outside of equilibrium. Inspired by microorganisms, we fabricate and synthesize microparticles that are highly dissipative, bestowing the fascinating and occasionally useful capability of harvesting energy from their environment and locally dissipating it to perform specific functions, such as self-propel or reconfigure (e.g., latch, crawl, contort). In my seminar, I will highlight our efforts to engender symmetry-breaking principles into microparticles for directed motion using energy from external acoustic, electric, and magnetic fields. I will describe how particle systems can be intelligently designed to actuate in prescribed ways. Building on basic principles, I will share how these dissipative systems can be used in functional assays for biomedicine. I will discuss how active particles can in some cases enhance the transport of drugs through biological barriers, facilitate the sensitive detection of biomolecules for disease identification, and cooperate with immune cells to enhance the performance of cell-based immunotherapies. Overall, I hope to convey how active and responsive microparticles show promise as a powerful and potentially disruptive tool for next-generation biomedicine.
Biography: Wyatt Shields joins us an Assistant Professor in the Department of Chemical and Biological Engineering at the University of Colorado Boulder. He received his B.S. from the University of Virginia in 2011 and Ph.D. from Duke University in 2016. He performed a brief postdoc at NC State on active matter and a second postdoc at Harvard University on cell-based immunotherapies. He started his research group at CU Boulder in 2020 and has gained national recognition for his work with awards such as the Packard Fellowship in Science and Engineering, the NSF CAREER award, the ONR young investigator award, the Pew Biomedical Scholars award, the NIH MIRA, and most recently the Camille Dreyfus Teacher-Scholar Award. His group focuses on developing field-responsive and active particles as vehicles for next-generation biosensing, drug delivery, and immunoengineering.
Host: Eunji Chung
Location: Ronald Tutor Hall of Engineering (RTH) - 109
Audiences: Everyone Is Invited
Contact: Carla Stanard
-
CA DREAMS - Technical Seminar Series
.jpg)
Fri, Jan 24, 2025 @ 12:00 PM - 01:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Umesh K. Mishra, Dean, Richard A Auhll Dean of the College of Engineering at the University of California Santa Barbara
Talk Title: A Brief History and the Promise of Gallium Nitride (GaN) Electronics; the Next Wave After GaN Photonics
Series: CA DREAMS - Technical Seminar Series
Abstract: In this talk, we will recount the development of GaN electronics over its history of nearly 40 years to its widepsread deployment today in commercial and DoD systems.
Biography: Umesh K. Mishra is the Richard A Auhll Dean of the College of Engineering at the University of California Santa Barbara and the Donald W. Whittier Distinguished Professor of Electrical and Computer Engineering at UC Santa Barbara. He received his B.Tech from the Indian Institute of Technology in Kanpur, India, his M.S from Lehigh University in Bethlehem, PA., and his Ph.D. in 1984 from Cornell University in Ithaca, NY. He has supervised 81 Ph.D theses to completion with 15 of them being women and 69 of them in the field of Gallium Nitride (GaN) materials and devices. 11 of his students are members of the faculty in prestigious universities, with 5 of them being women. His students have founded/co-founded 10 companies. He co-founded the first start-up in the world to commercialize RF GaN transistors and LEDs in 1996 (Nitres) which was acquired by CREE (now Wolfspeed) in 2000. Umesh co-founded Transphorm in 2007 which was honored as a Technology Pioneer at the World Economic Forum, 2013, to commercialize GaN-on-Si transistors for power conversion. Transphorm was acquired by Renesas in 2024. He has over 1000 papers (>70,000 citations; h-index 130) and over 100 patents. Umesh received several awards including the IEEE Jun-Ichi Nishizawa Medal for his contributions to the development and commercialization of GaN electronics. He is a Fellow of the IEEE, an International Fellow of the Japanese Society of Applied Physics, Fellow of the National Academy of Inventors, a Member of the National Academy of Engineering and a Distinguished Alumnus of IIT Kanpur.
Host: Dr. Steve Crago
More Info: https://www.isi.edu/events/5356/a-brief-history-and-the-promise-of-gallium-nitride-gan-electronics-the-next-wave-after-gan-photonics/
Webcast: https://usc.zoom.us/j/97017422125?pwd=Dbrt8MNMrmBV3xalKQJcAiNsggFJjJ.1&from=addonWebCast Link: https://usc.zoom.us/j/97017422125?pwd=Dbrt8MNMrmBV3xalKQJcAiNsggFJjJ.1&from=addon
Audiences: Everyone Is Invited
Contact: Amy Kasmir
-
PhD Dissertation Defense - Mozhdeh Gheini
Fri, Jan 24, 2025 @ 04:00 PM - 05:00 PM
Thomas Lord Department of Computer Science
University Calendar
Title: Inductive Biases for Data- and Parameter-Efficient Transfer Learning
Date and Time: Fri, Jan 24, 2025 @ 10:00 AM - 12:00 PM
Location: Salvatori Computer Science Center (SAL) - 213 and https://usc.zoom.us/j/6564802162
Committee Members: Jonathan May (Chair), Emilio Ferrara, Xuezhe Ma, Khalil Iskarous
Abstract: Data- and resource-intensive pre-training and fine-tuning applied upon Transformer-based models is the dominant paradigm at the forefront of rapid advancements in natural language processing, human language technologies, and most notably, large language models. Such reliance on massive amounts of data, computation, and energy, while effective and impressive from a performance-only perspective, can hinder open, nonexclusive, and sustainable development of these technologies. In this talk, we present how certain inductive biases can be devised to adjust current natural language methods under resource-constrained scenarios and provide insights into why the proposed inductive biases are successful in such cases.
Specifically, we discuss four research directions on data and parameter efficiency of fine-tuning and transfer learning in natural language processing: (1) a universal regimen that creates a single pre-trained checkpoint suitable for machine translation transfer to practically any language pair and eliminates the need for ad hoc pre-training; (2) an architecture-guided parameter-efficient fine-tuning method that performs competitively with full fine-tuning while exclusively updating cross-attention parameters; (3) an analysis of MEGA, a recently introduced augmentation of the Transformer architecture to incorporate explicit recency bias, through the lens of transfer learning; and (4) a meta-learning algorithm to prime pre-trained models for specific fine-tuning strategies.
Combined with ablations that show how they are effective and analyses that demonstrate their generalizability, these directions are meant to serve as tools for resource-efficient transfer learning for natural language processing.Location: Henry Salvatori Computer Science Center (SAL) - 213
Audiences: Everyone Is Invited
Contact: Mozhdeh Gheini
.jpg)