AI@AU Talks

Artificial intelligence as a Catalyst for Innovation and Discovery

Dr. Charreau Bell, Dept. of Computer Science / Vanderbilt Data Science Institute, Vanderbilt University

Abstract:

In this talk, we explore artificial intelligence as a transformative force across a range of academic disciplines. The Data Science Institute (DSI) at Vanderbilt University employs an inclusive, collaborative framework for data science/AI projects and research, culminating in a diverse AI portfolio. Contextualized by these efforts, we will discuss the evolution of AI and the increasingly sophisticated solutions it offers to various disciplines. Additionally, we will examine the impact of AI's growing democratization and its role in lowering technical barriers, broadening the scope of who can innovate and problem-solve using these tools. This shift is paving the way for new scientific discoveries and expansive innovation as a more diverse set of minds come together to create solutions for positive societal impact.

Bio:

Charreau Bell, Ph.D., is a senior data scientist at the Data Science Institute (DSI) at Vanderbilt University (VU), the faculty director of the undergraduate data science minor, and an assistant professor of the practice of computer science at VU. In her current role, she leads several data science projects across a spectrum of disciplines using artificial intelligence and machine learning to enable discovery and create new pathways for innovation. She seeks to train and empower others of diverse skill levels and educational backgrounds confidently and responsibly use data science tools to reach their research goals and business objectives.

Prior to joining the DSI, she earned her Doctorate, Master’s, and Bachelor’s degrees in engineering from Vanderbilt University. Her Ph.D. focused on creating new algorithms for understanding the resting state behavior of the brain using functional brain imaging technology. For more details about the speaker, please click here: https://www.vanderbilt.edu/datascience/person/charreau-bell/

View Presentation

AI in the SEC

Dr. LeNá Powe McDonald, SEC AI Constortium, Southeastern Conference

This talk will be presented in ACLC Rm. 259A (Biggio Center) and will be livestreamed on Zoom @ https://auburn.zoom.us/j/81706339239, and recorded for later viewing at https://eng.auburn.edu/ai-au/forum.

Abstract:

The SEC’s Artificial Intelligence Consortium is at the forefront of identifying and providing tangible resources for the campuses that we serve. Artificial intelligence (AI) continues to impact all aspects of daily life, including competitive sports. As the first athletic conference to host a consortium aimed at addressing concerns and potential benefits of AI, the SEC is finding new and exciting ways to promote the academic programming and rigor associated with our member institutions. This presentation will provide an overview of the accomplishments, partnerships, goals, and future objectives associated with the SEC Artificial Intelligence Consortium. A review of how the group was established and the educational resources that are shared throughout the Conference as the result of the consortium will also be addressed.

Bio:

Dr. LeNá Powe McDonald currently serves as Associate Commissioner for Academic Relations for the SEC. In her role, she leads SEC activities meant to highlight the academic accomplishments and impact of SEC universities, including academically focused intercollegiate athletics opportunities.

Prior to the SEC, she worked for the University of Alabama where she served as Director of Internal Affairs and UA System Liaison. She previously worked in communications for UA's Division of Advancement and served as a faculty member in UA's Department of Communication Studies.

Dr. McDonald holds a Bachelor of Arts degree in Dance and Community Development, a Master of Arts degree in Communications Studies, and a Doctor of Philosophy in Higher Education Administration all from The University of Alabama.

She is an alumna of the Oxford Women's Leadership Symposium at the University of Oxford and was named a Birmingham Trailblazer by The Birmingham Times. She was also named Best Young Professional by Birmingham Magazine Best in Minority Business Awards and an honoree in Who's Who in Black Alabama. In 2023, Dr. McDonald was named a Top Alumni Under 40 by the College of Education at The University of Alabama and she is a member of the 2023-2024 Class of the Alabama Leadership Initiative.

She is an active member of several boards including the Board of Directors of the Capstone Education Society at UA and the Advisory Board of the O’Neal Comprehensive Cancer Center at UAB. She is a member of the Birmingham Chapter of The Links, Incorporated and Alpha Kappa Alpha Sorority, Incorporated.

View Presentation

AI Policies and Initiatives: Progress and Opportunities at the Federal Level and in Tennessee

Dr. Lynne Parker, Director, AI Tennessee Initiative, The University of Tennessee

Abstract:

The national dialogue around AI policies and initiatives has skyrocketed over the last five years. As a Nation, we have gone from everyday citizens having no awareness of AI to a society in which everyone now has an opinion about AI. While numerous policies and initiatives are in place to address the opportunities and challenges presented by AI, much work remains to be done – both at the federal level and within individual states. Drawing from my experience leading AI policy at the White House and now steering the AI Tennessee Initiative, I’ll overview key actions taken nationally and in Tennessee. We’ll explore key challenges that remain – particularly those that require broad interdisciplinary expertise. With this talk, I hope to inspire collaboration towards unlocking the power of AI, ultimately ensuring broad participation in its benefits.

Bio:

Dr. Lynne Parker is Associate Vice Chancellor at the University of Tennessee, Knoxville (UT) and Director of the AI Tennessee Initiative, which aims to position Tennessee as a national and global leader in the data-intensive knowledge economy. Previously, she served as Founding Director of the National Artificial Intelligence Initiative Office in the White House Office of Science and Technology Policy and as Deputy Chief Technology Officer of the United States. She previously served for two years at the National Science Foundation as Division Director for Information and Intelligent Systems. In these roles across three Administrations, she led the development of numerous landmark national AI policies bolstering research, governance, education and workforce training, international engagement, and the Federal use of AI.

Dr. Parker joined the UT faculty in 2002 and is an expert on distributed and intelligent robot systems, human-robot interaction, and AI. She has held numerous other leadership positions besides those noted above, including at UT (Interim Dean of the Tickle College of Engineering) and Oak Ridge National Laboratory (Distinguished R&D Staff Member and Group Leader). She is a Fellow of AAAI, AAAS, and IEEE, and a Distinguished Member of ACM. She received her PhD in computer science from MIT. For more details about the speaker, please click here: https://research.utk.edu/lynne-e-parker/

View Presentation

Human-Centered Spatiotemporal Modeling and Simulation in The Era of Infrastructure

Dr. Pan He, Dept. of Computer Science & Software Engineering, Auburn University

Abstract:

With the advent of increasingly affordable and innovative sensors, there has been a surge in interest surrounding the development of infrastructure applications that have a substantial impact on society and influence people's lives. However, the complexity arises from the fact that obtaining sensory data from real-world infrastructure is time-consuming and expensive, posing a major challenge in efficiently processing and understanding this data. Effectively connecting various spatiotemporal information perceived in these applications adds another layer of intricacy. In this talk, I am thrilled to share insights from my recent research on spatiotemporal visual and point cloud modeling. Specifically, I will delve into the realm of understanding motion and correspondence within dynamic point cloud sequences, requiring minimal human annotations. Moreover, I will showcase the integration of human intelligence into AI systems, emphasizing its role in enhancing model prediction and facilitating data annotation. The presentation will culminate in a discussion on ongoing research endeavors, including the construction of a high-fidelity 3D simulation environment for traffic scenarios. This environment will not only support auto-generate diverse traffic scenarios but also incorporate sensor data for reducing synthetic-to-real domain gap. This research focuses on systematically evaluating and improving services and safety for all road users, contributing to topics such as traffic signal control, transportation decarbonization, and traffic safety measures, thereby potentially reducing traffic congestion and incidents and improving overall quality of life. The research anticipates the rise of connected and automated vehicles, fostering research in multi-agent and collaborative perception.

Bio:

Dr. Pan He is an Assistant Professor in the Department of Computer Science and Software Engineering at Auburn University, Alabama. He mainly works in computer vision, machine learning, and smart infrastructure by developing fundamental AI techniques and integrating them to support downstream applications in next-generation infrastructure systems. He has authored or co-authored one book, one technical report for the Florida Department of Transportation and over 25 peer-reviewed journals and conference articles. The audiences can find more information on his website located at https://panhe.org/ He obtained his PhD in the Department of Computer and Information Science and Engineering, University of Florida (UFL) in August 2023. He received a B.E. degree with honors from the Department of Software Engineering at Sichuan University in June 2015. From 2014-2016, he worked as a Research Assistant in the Multimedia Lab at the Chinese University of Hong Kong and Multimedia Research Center, Shenzhen Institute of Advanced Technology, during which he received the Dean’s Award for Innovation and Creativity. At UFL, he worked as a research assistant at the CISE MALT lab after a one year ECE master program at the NSF Center for Big Learning. He received the International Center Outstanding Achievement Award (2021) and Gartner Group Graduate Fellowships (2020, 2021, and 2022). In addition, he received the Doctoral Consortium Awards at International Conference on Computer Vision (ICCV) 2021 and Conference on Computer Vision and Pattern Recognition (CVPR) 2022. He has been serving as an associate editor for IEEE Transactions on Neural Networks and Learning Systems (TNNLS) since 2024. For more details about the speaker, please click here: https://panhe.org/

View Presentation

Women Led Research in Computational Intelligence

Dr. Alice Smith, Joe W. Forehand, Jr. Distinguished Professor, Dept. of Industrial Systems Engineering, Auburn University

Abstract:

This talk will give a current overview of some of the exciting and impactful research endeavors in computational intelligence by women led investigative teams from around the world. The work is taken from the 2022 landmark volume Women in Computational Intelligence: Key Advances and Perspectives on Emerging Topics, https://link.springer.com/book/10.1007/978-3-030-79092-9. The 34 authors of this book are nearly all women and represent thirteen countries across five continents. All chapters are authored by IEEE Women in Computational Intelligence members except the first chapter which is a professional biography of computer pioneer Admiral Grace Hopper by Jill S. Tietjen, an esteemed author on women and technology.

The book is structured into four main sections of Intelligence, Learning, Modeling, and Optimization. The primary technical methods include artificial neural networks, evolutionary and swarm computation, and fuzzy logic and systems. The wealth of applications can be seen throughout the nineteen chapters within this volume. These include natural language processing, intelligent tutoring, autonomous systems, digital pathology, intrusion detection, and energy management. The talk will highlight a sampling of these research chapters, explaining the importance and novelty of the work described.

A unique part of this book is the biographies of the authors which include information concerning their beginnings and advancement in computational intelligence research along with advice for those considering this field and its possibilities. The talk will also give some short videos from these authors on their experiences with computational intelligence and their career advice to those getting started. The aim of this talk to celebrate the contributions of women in CI and to inspire future generations of CI scholars through a lens of diversity and inclusion.

Bio:

Dr. Alice Smith is the Joe W. Forehand, Jr. Distinguished Professor of the Industrial and Systems Engineering Department at Auburn University, where she served as Department Chair from 1999 to 2011. She also has a joint appointment with the Department of Computer Science and Software Engineering. Previously, she was on the faculty of the Department of Industrial Engineering at the University of Pittsburgh from 1991 to 1999, which she joined after industrial experience with Southwestern Bell Corporation. Dr. Smith has degrees from Rice University, Saint Louis University, and Missouri University of Science and Technology.

Dr. Smith’s research focus is analysis, modeling, and optimization of complex systems with emphasis on computation inspired by natural systems. She holds one U.S. patent and several international patents and has authored more than 200 publications which have garnered over 17,000 citations and an H-Index of 49 (Google Scholar). She is the editor of Women in Computational Intelligence: Key Advances and Perspectives on Emerging Topics ( https://link.springer.com/book/10.1007/978-3-030-79092-9) and Women in Industrial and Systems Engineering: Key Advances and Perspectives on Emerging Topics ( https://www.springer.com/us/book/9783030118655#aboutBook). Several of her papers are among the most highly cited in their respective journals including the most cited paper of Reliability Engineering & System Safety and the 3rd most cited paper of IEEE Transactions of Reliability. She won the E. L. Grant Best Paper Awards in 1999 and in 2006, and the William A.J. Golomski Best Paper Award in 2002. Dr. Smith is the Editor in Chief of INFORMS Journal on Computing and an Area Editor of Computers & Operations Research. For more details about the speaker, please click here: https://www.eng.auburn.edu/~aesmith/

View Presentation

Charting the Course - AI's Role in the Animation Industry

Dr. Jamy Wheless

Abstract:

The talk offers an insightful exploration into how artificial intelligence is revolutionizing animation, broadening creative possibilities, enhancing production efficiency, and fostering new career opportunities. It begins with an introduction to the various AI tools already in use within the industry, such as machine learning, motion capture, character rigging, rendering optimization, and facial animation. The presentation then delves into the animation filmmaking process, highlighting the ways in which AI aids in development and pre-production through conceptualization, storyboarding, character design, script development, and audience feedback prediction. It also covers the production phase, emphasizing the role of AI in character modeling, environment creation, motion capture techniques, and camera layout. The talk also introduces AI tools which promise to disrupt traditional animation processes further.

Bio:

Jamy Wheless is an animation director, producer, adjunct professor, and chief operating officer of Ignite Animation Studios and co-founder of Lightstream Animation Studios.

With over 30 years of experience in film production, he began his career at George Lucas’s Industrial Light & Magic where he was a team leader on over a dozen films and commercials.

He was an animation lead on the Academy Award VFX winner Pirates of the Caribbean and was also responsible for the character development and animated performances for Yoda on the Star Wars prequels as well as Ang Lee’s film, Hulk. He was animation director on the fantasy film, The King’s Daughter, and co-director and producer of the Oscar-nominated and award-winning short film The Pig on the Hill.

He recently directed and produced another short film titled Andy: “A Dog’s Tale”, which earned a place on the longlist for an Oscar nomination. He collaborated with Jean Schulz, chairman of the board of Charles M. Schultz Creative Associates, (known for bringing us Peanuts), and Canine Companions. Jamy and his wife Amy have four adult children and live in Petaluma, California.

View Presentation

Alabama to the World: Pioneering AI for Rural Resilience

Dr. Jackey Gong, Dept. of Mechanical Engineering / Sensor-Accelerated Intelligent Learning Laboratory (SAIL)

Abstract:

At the forefront of integrating AI to enhance rural resilience, the state of Alabama stands poised to become a leader in this innovative field. This talk will spotlight Alabama's unique capabilities in rural health, water management, and transportation as a foundation for pioneering AI solutions. The newly founded Alabama Center for the Advancement of Artificial Intelligence focuses on developing sustainable, AI-driven solutions to address critical challenges in these sectors, with benefits extending beyond Alabama's rural communities to have a global impact. Additionally, the talk will emphasize the crucial role of collaborative efforts among academic researchers, local communities, and policymakers in driving these advancements in AI.

Bio:

Dr. Jiaqi (Jackey) Gong serves as the Inaugural Director of the Alabama Center for the Advancement of Artificial Intelligence. With a prolific academic record, he has authored over 70 peer-reviewed papers in the field of AI. Dr. Gong's extensive interdisciplinary research expertise shines in the development of advanced AI-based systems, particularly attuned to the social and ethical contexts within rural resilience across health, education, and environmental sectors.

At UA, Dr. Gong's research has been recognized and supported by significant funding, totaling around $3M in his share, primarily from the NSF and NIH. His work focuses on Human-Centered AI, integrating human-centered computing with AI principles to enhance human capabilities within a socially and ethically aware framework. This innovative approach has aided his efforts to develop AI systems that effectively monitor, model, and modify human behavior. These systems aim to elevate clinical benefits and significantly bolster rural resilience. Dr. Gong's contributions to AI are pivotal in bridging technological advancement with human-centric applications, underscoring his commitment to leveraging AI for societal betterment, especially in rural communities. For more details about the speaker, please click here: https://eng.ua.edu/eng-directory/dr-jiaqi-jackey-gong/

View Presentation

Democratizing AI with Information Assurance and Knowledge Grounding

Dr. Santu Karmaker

Abstract:

Even though Artificial Intelligence (AI) has existed for a while, its broad accessibility is a recent development, exemplified by ChatGPT. This is because, until recently, we lacked the capability to engage with AI in natural conversations and experience responses akin to human-like interactions. While such broad accessibility provides a great opportunity to democratize AI across general people, it comes with several key risks and challenges, e.g., Information Assurance, Contextual Understanding, Biases, Ethical Concerns, Cyber Threats, and Data Poisoning, to name a few. This talk will focus exclusively on two important challenges related to the democratization of AI, i.e., Information Assurance (in the first part) and Contextual Understanding (in the second part). The first part of the talk will introduce a general framework for Information Assurance in the context of multi-perspective narrative understanding and demonstrate how zero-shot deep-learning approaches can be devised to measure and provide such Assurance. The second part will demonstrate how Contextual Understanding can be enhanced in Conversational-AI systems for ill-defined and previously unseen complex tasks and lay the roadmap for a natural dialog-based intelligent agent, a “Virtual Interactive Data Scientist (VIDS)”. VIDS can be considered the future SIRI or ALEXA that can assist users in solving real-world data science problems through natural conversation. Finally, the talk will highlight some interesting future directions in line with the democratization of AI and associated challenges.

Bio:

Dr. Shubhra Kanti Karmaker (“Santu”) is an Assistant Professor in the Department of Computer Science and Software Engineering at Auburn University, Alabama. With a broad interest in the academic field of Artificial Intelligence and Data Science, his primary research focus lies at the intersection of Natural Language Processing (NLP) and Information Retrieval (IR). More specifically, his research is primarily driven by the following broad research question: “How can we make AI and Data Science more accessible to the general people while ensuring Information Integrity and Knowledge Grounding?” Dr. Karmaker is currently leading an NSF-funded project ($700,854) as PI to conduct research leveraging LLMs to design intelligent support for simulation-based science learning in schools as part of the NSF program Research on Emerging Technologies for Teaching and Learning. He also leads a single PI 3-year AFOSR-funded project ($542,485) to analyze, understand, extract commonalities and differences from, and summarize descriptions of topics/events by different people (or agents) with different perspectives/ biases. Dr. Karmaker has previously received the Short-Term Innovative Research (STIR) Award ($60,000) as the PI by the Army Research Office for conducting basic scientific research on natural language understanding. His research team at Auburn CSSE has recently become Champion in the “Food for Thought” NLP challenge ($50,000 award) hosted by Coleridge Initiative in collaboration with the USDA. Dr. Karmaker completed his Ph.D. in Computer Science from the University of Illinois Urbana Champaign (UIUC) and was then a Postdoctoral Research Associate in the Laboratory for Information and Decision Systems (LIDS) at the Massachusetts Institute of Technology (MIT). During his Ph.D., he also worked as a summer research intern at Microsoft Research, Yahoo Research, and @WalmartLabs. For more details about the speaker, please click here.

Click here for the Presentation

Natural Language Processing for Privacy, Empowerment, and Social Good

Dr. Shomir Wilson, College of Information Sciences & Technology, Penn State University

Abstract:

Internet users care about privacy, but persistent gaps exist in their understanding of how data about them is collected and used. I will describe a trajectory of research to automatically extract information from the text of websites' and apps' posted privacy policies and to present it to consumers, regulators, and privacy researchers in ways that better respond to their needs. A recent effort in this trajectory is PrivaSeer (https://privaseer.ist.psu.edu/) a search engine and corpus that together make a collection of over 1M website privacy policies available and explorable for privacy stakeholders. I will also describe my other projects to apply natural language processing (NLP) to problems in privacy and fairness, with a general strategy of using NLP for empowerment and social good.

Bio:

Shomir Wilson is an Assistant Professor in the College of Information Sciences and Technology at the Pennsylvania State University, where he leads the Human Language Technologies Lab. His research interests span natural language processing, privacy, and computational social science. He is particularly interested in breaking down technology's "walls of text", situations where a human reader is expected to consume a large quantity of text to take action while lacking time or expertise to properly understand it. He holds over $2M in active grants from the National Science Foundation and the National Institutes of Health, covering research on usable privacy, legal text, and fairness in law enforcement. Prior to becoming faculty he held postdoctoral positions in Carnegie Mellon University's School of Computer Science and the University of Edinburgh's School of Informatics. He received his Ph.D. in Computer Science from the University of Maryland in 2011. For more details about the speaker, please click here.

Click here for the Presentation

Deep Convolutional Neural Network Enabled Automated Face Recognition and its Unforeseen Consequences

Dr. Michael C. King, Dept. of Computer Engineering & Sciences, Florida Tech University

Abstract:

Face recognition continues to grow in popularity as a key enabler of biometric authentication. This technology is becoming an integral part of everyday lives as it is used to access smartphones and even board aircraft—with frictionless travel in mind. While independent tests confirm impressive gains in the overall accuracy of the technology, concerns remain as it relates to the measured disparity in performance relative to various demographic groups and the sources of large data repositories needed for development. Further, there have been calls to completely ban the technology due to concerns over how it is used by the federal government, law enforcement, and even private/commercial entities. This talk will discuss salient aspects of face recognition technology development and the unforeseen consequences of its use (e.g., wrongful arrests and denial of services.)

Bio:

Dr. Michael King joined the Florida Institute of Technology’s Harris Institute for Assured Information as a Research Scientist in 2015 and holds a joint appointment as an Associate Professor in the Department of Computer Engineering and Sciences. Before joining academia, Dr. King served for more than 10 years as a scientific research/program management professional in the United States Intelligence Community. While in government, Dr. King created, directed, and managed research portfolios covering various topics related to biometrics and identity, including advanced exploitation algorithm development, advanced sensors and acquisition systems, and computational imaging. He crafted and led the Intelligence Advanced Research Projects Activity’s (IARPA) Biometric Exploitation Science and Technology (BEST) Program to transition technology deliverables successfully to several Government organizations. Recognized as an expert in biometrics and identity intelligence, he has been invited to brief the Director of National Intelligence, Congressional staffers and science advisers, the Defense Science Board, and the Intelligence Science Board. He also served as Intelligence Community Department Lead to the White House Office of Science and Technology Policy's National Science and Technology Council Subcommittee on Biometrics and Identity Management (2005 – 2012). Dr. King received his Ph.D. in Electrical Engineering from North Carolina Agricultural and Technical State University in 2001 and has research interests in biometrics, cyber identity, and machine learning. For more details about the speaker , please click here.

Click here for the Presentation

Network Medicine in the Age of AI

Dr. Zongliang Yue, Dept. of Health Outcomes Research and Policy, Auburn University

Abstract:

Network-based approaches offer a robust means of investigating, modeling, and characterizing molecular mechanisms and gene modules in genotype-to-phenotype associations. Over the past decade, significant advancements in disease classification, drug repositioning, and treatment response prediction have been achieved through the application of systems biology. The emergence of multi-omics techniques, single-cell techniques, and artificial intelligence presents unprecedented opportunities, propelling us into a new era of network-based drug discovery. This exciting paradigm involves integrating multi-omics data into network analytics to identify specific biological interactions and regulations tailored to personalized medicine using artificial intelligence. The integrative multi-omics network-based analysis raises fundamental questions: How can we effectively eliminate noise and amplify signals in multi-omics data analytics? How can we discern biological conditional interactions and regulations using multi-omics data? And, most importantly, how can we enhance the chances of success in precision medicine through the utilization of multi-omics data? In this presentation, I will showcase the use of multi-omics data to integrate network biology, empowering personalized medicine initiatives. Additionally, I hope to ignite the common interest of the AI@AU community, inspiring us to revisit and rethink the critical questions as we develop cutting-edge technologies to elevate health outcomes and overall quality of life. 

Bio:

I am an assistant research professor, currently working in the Health Outcomes Research and Policy department at Harrison College of Pharmacy at Auburn University, and the research committee board member at AI@AU. My research has mainly focused on Gene-set, Network, Pathway Analysis (GNPA), disease novel biomarker discovery, and drug repositioning. My work has also been directly implemented in discovering the molecular mechanisms of several cancers and critical diagnostic biomarkers to improve human health. In the coming years, I intend to extend my research on complex disease Omics data analysis, which contributes to cancer progression risk assessment, screening, differential diagnosis, determination of prognosis, prediction of response to treatment, and monitoring of progression of cancers. The challenge in this field is genetic heterogeneity, which needs a complete understanding of the biomolecular and genetic mechanisms before making any personalized treatment plan. The specific cohorts monitored in drug susceptibility should be taken care of with personalized medicine. Therefore, I will contribute to complex network-based modeling by providing cohort-specific network-based modules. I also intend to extend my prior experiments to drug repositioning for personalized medicine to save more lives or tremendously increase the survival rate in specific cohorts. Working at Harrison College of Pharmacy allows me to engage in exciting interdisciplinary projects focused on developing more effective and safer therapies for complex diseases. For more details about the speaker, please click here.  

Click here for the Presentation

Multimodal Generative LLMs: Unification, Interpretability, Evaluation

Dr. Mohit Bansal, Dept. of Computer Science, University of North Carolina at Chapel Hill

Abstract:

In this talk, I will present our journey of large-scale multimodal pretrained (generative) models across various modalities (text, images, videos, audio, layouts, etc.) and enhancing important aspects such as unification, efficiency, interpretability, and evaluation. We will start by discussing early cross-modal vision-and-language pretraining models (LXMERT). We will then look at early unified models (VL-T5) to combine several multimodal tasks (such as visual QA, referring expression comprehension, visual entailment, visual commonsense reasoning, captioning, and multimodal translation) by treating all tasks as text generation. We will also look at recent advanced unified models (with joint objectives and architecture, as well as newer unified modalities during encoding and decoding) such as textless video-audio transformers (TVLT), vision-text-layout transformers for universal document processing (UDOP), and composable any-to-any text-audio-image-video multimodal generation (CoDi). Second, we will discuss interpretable and controllable multimodal generation via LLM-based planning and programming, such as layout-controllable image generation via visual programming (VPGen), consistent multi-scene video generation via LLM-guided planning (VideoDirectorGPT), and open-domain, open-platform diagram generation (DiagrammerGPT). I will conclude with important evaluation aspects of multimodal generation models, based on fine-grained skill and social bias evaluation (DALL-Eval), as well as interpretable and explainable visual programs (VPEval).

Bio:

Dr. Mohit Bansal is the John R. & Louise S. Parker Professor and the Director of the MURGe-Lab (UNC-NLP Group) in the Computer Science department at UNC Chapel Hill. He received his PhD from UC Berkeley in 2013 and his BTech from IIT Kanpur in 2008. His research expertise is in natural language processing and multimodal machine learning, with a particular focus on multimodal generative models, grounded and embodied semantics, language generation and Q&A/dialogue, and interpretable and generalizable deep learning. He is a recipient of IIT Kanpur Young Alumnus Award, DARPA Director's Fellowship, NSF CAREER Award, Google Focused Research Award, Microsoft Investigator Fellowship, Army Young Investigator Award (YIP), DARPA Young Faculty Award (YFA), and outstanding paper awards at ACL, CVPR, EACL, COLING, and CoNLL. He has been a keynote speaker for the AACL 2023 and INLG 2022 conferences. His service includes ACL Executive Committee, ACM Doctoral Dissertation Award Committee, CoNLL Program Co-Chair, ACL Americas Sponsorship Co-Chair, and Associate/Action Editor for TACL, CL, IEEE/ACM TASLP, and CSL journals. For more details about the speaker, please click here.

Click here for the Presentation

Digital Literacy and Citizenship: AI Practices for the 21st Century Workforce

Dr. Leslie A. Cordie, Dept. of Educational Foundations, Leadership and Technology, Auburn University

Abstract:

Workplaces are changing. It is estimated that in less than 20 years, more than 90% of all jobs will require people to work with digital technologies. All technological skills, both advanced and basic, will see a substantial growth in demand. Advanced technologies require people who understand how they work and can innovate, develop, and adapt to them. In a world of AI-generated content and automated decision-making, individuals who can critically assess information and make informed judgments will be highly valued. To better serve students, colleges need to understand the evolving world of work and hiring in this AI-intense environment. Thus, we must assist higher education students in developing the digital skills that employers want now, as well as the confidence and deeper-seated capabilities that will enable them to progress in their careers and adapt to emerging technologies. Encouraging a growth mindset and adaptability will prepare learners for the ongoing changes in the workforce. A focus on lifelong learning, while employing a ‘skills first’ approach in higher ed can help meet the future challenges due to the constant evolution of technology.

Bio:

Dr. Leslie Cordie is an Associate Professor in the EFLT (Educational Foundations, Leadership and Technology) Department , and Affiliate Faculty with University Writing. She holds a bachelor’s degree in nursing from the University of Wisconsin-Milwaukee, an MBA from the University of Texas at Austin, and a PhD in Adult Education and Technical Communication from Colorado State University. Her specialties include instructional and curricula design, professional development, and blended learning. Cordie has over 25 years of experience working and consulting in academic, business, and military environments. Dr. Cordie is a Fulbright Scholar with active international collaborations in the West Indies, Ireland, the UK, and Asia. For more details about the speaker, please click here.  

Click here for the Presentation

The Neural Computations Underlying Human Social Interaction Recognition

Dr. Leyla Isik, Dept. of Cognitive Science, Johns Hopkins University

Abstract:

Humans perceive the world in rich social detail. We effortlessly recognize not only objects and people in our environment, but also social interactions between people. The ability to perceive and understand social interactions is critical for functioning in our social world. We recently identified a brain region that selectively represents others’ social interactions in the posterior superior temporal sulcus (pSTS) in a manner that is distinct from other visual and social processes, like face recognition and theory of mind. However, it is unclear how social interactions are processed in the real world where they co-vary with many other sensory and social features. In the first part of my talk, I will discuss new work using naturalistic fMRI paradigms and machine learning analyses to understand how humans process social interactions in real-world settings. We find that features of a social interaction are extracted hierarchically along the STS, with strong selectivity for communicative interactions, even after controlling for the effects of other co-varying perceptual and social information. In the second part of my talk, I will discuss the computational implications of social interaction selectivity in the brain, and present a novel graph neural network model, SocialGNN, that instantiates these insights. SocialGNN reproduces human social interaction judgements in both controlled and natural videos using only visual information, without any explicit model of agents’ minds or the physical world, but requires relational, graph structure and processing to do so. Together, this work suggests that social interaction recognition is a core human ability that relies on specialized, structured visual representations.

Bio:

Leyla Isik is the Clare Boothe Luce Assistant Professor of Cognitive Science at Johns Hopkins University. Her research aims to understand humans’ vast visual and social abilities. In just a fraction of a second, humans not only detect the objects and people in their environment, using a combination of human neuroimaging, intracranial recordings, machine learning, and behavioral techniques. Before joining JHU, Leyla was a postdoctoral researcher at MIT and Harvard in the Center for Brains, Minds, and Machines, and she completed her PhD at MIT. For more details about the speaker, please click here.

Click here for the Presentation

Engaging Students with ChatGPT in an Intro to MIS Class & Supporting A.I. in Education at Auburn University

Dr. Asim Ali

Abstract:

   Dr. Asim Ali, Executive Director of the Biggio Center and instructor in the Harbert College’s Department of Business Analytics & Information Systems, will discuss his modification to an essay assignment to use ChatGPT in his Intro to MIS class taught to 120 business students. He will share the benefits and challenges associated with implementing the redefined assignment. Dr. Ali will also share the Biggio Center’s strategy for building the campus community's capacity for A.I.’s impact in education. Topics will include a brief overview of Biggio Center’s Teaching with AI at AU Canvas course, an upcoming speaker series, and an open discussion about curricular enhancement opportunities.

Click here for the Presentation

Going to the Dogs: The DCSITE Program – Advancing Mobile Sensor Technology

Dr. F.F. 'Skip' Bartol

Abstract:

   They are man’s best friend and man’s best defense. They are smart, durable, adaptable, mobile, rapidly programmable, hypersensitive sensor platforms with ability to trace an odor to its source. With extreme odor sensitivity in the part per trillion range ( e.g. 1 second/31,500 years) they can learn new targets, discriminate complex odor profiles in real-time, and search large areas and populations for targets in environmental extremes. They are DOGS! Dogs have lived and worked with and for mankind for millennia. Our partners and companions since before the invention of agriculture, dogs have been by our sides through times of peace and war for centuries. This presentation will introduce and review aspects of the transdisciplinary, AU/AUCVM Canine Performance Sciences (CPS) Detection Canine Sciences, Innovation, Technology and Education (DCSITE) program with the objective of inviting discussion regarding opportunities for engagement with AI@AU.

Turbocharging Agri-Forestry Sciences and Production Through Imaging, Robotics, and Artificial Intelligence

Dr. Yin Bao

Abstract:

   The recent technological convergence of sensors, imaging, robotics, and artificial intelligence (AI) enables practical and novel applications for turbocharging scientific discoveries and production systems in agriculture and forestry. Such automation technologies not only reduce the dependency on a declining farm workforce, but also improve sustainability in multiple aspects via data-driven decision making. This seminar will present several such case studies in the field of phenomics and precision farming, covering high-throughput phenotyping of plant characteristics in field and controlled environments, AI-powered low-cost precision livestock monitoring, and robotic forest nursery inventory. Looking into the future, I envision several novel AI-based sensing and robotic systems to bridge the gap between high-end instrumentation systems for plant/animal research and accessible technologies for crop/livestock production. Interdisciplinary research between biologists, geneticists, engineers, and computer scientists is critical to the success of such endeavors to help feed and fuel a growing global population.

Click here for the Presentation

The Role of Artificial Intelligence on Digital Agriculture: Study Cases

Dr. Brenda Ortiz ,  Dr. Mailson Oliveira

Abstract:

   Digital technologies, artificial intelligence (AI), and decision support systems play a key role in agriculture as they support the implementation of site-specific crop management practices to increase productivity and protect the environment. Today, large volumes of data from crop, soils, and the environment are collected in near real-time from farm machinery, proximal and remote crop and soil sensors, as well as weather sensors. Various examples of how these data and AI are integrated to support farmers’ crop management decisions will be highlighted. Specific AI applications to row crops agriculture discussed will cover aspects of irrigation management, crop yield prediction, and crop maturity forecast.

Click here for the Presentation

AI and Academic Libraries: Chasing a Moving Target

Dr. Aaron Trehub ,  Dr. Ali Krzton

Abstract:

   ChatGPT, Google Bard, and now Microsoft 365 Pilot—AI and machine learning tools are at the top of today’s news feeds. The rapid pace of development in this field and the almost-weekly appearance of powerful new AI tools have universities and other institutions scrambling to develop policies and applications for their use. Academic libraries are no exception. Panels on AI/ML have started popping up at national library conferences to standing-room-only crowds. The Auburn University Libraries (AUL) have been experimenting with AI/ML since 2019, when librarians and library IT staff began using IBM Watson in an exploratory project with the Military REACH Program in the College of Human Sciences. Although that project has been eclipsed by internal developments at IBM and the advent of more-powerful and easier-to-use solutions, it introduced Auburn librarians to a pioneering suite of AI tools and provided them with hands-on experience in using AI for a real-world application. AUL Research Data Management Librarian Ali Krzton and AUL AD for Technology and Research Support Aaron Trehub will describe their experience of working with the IBM Watson team, their attempt to use Watson with the Military REACH library of research publications, and current discussions around the professional and ethical implications of AI and machine learning in the academic library world.  

Click here for the Presentation

Quantitative Techniques in Cardiovascular MRI

Dr. Tom Denney, Jr.,   Mr. & Mrs. Bruce Donnellan & Family Endowed Professor

Abstract:

   Magnetic resonance imaging (MRI) is increasingly used in clinical settings to image the anatomy and physiological function of the cardiovascular system. While the image acquisition can be done in a reasonable amount of time, the analysis of these images to obtain quantitative measures of cardiac function is time consuming and requires considerable manual intervention. Consequently, only left ventricular volumes are measured at two points in the cardiac cycle, end-diastole and end-systole, to obtain stroke volume and ejection fraction. In this talk, techniques are presented to leverage the manual analysis that is typically done at two time points in the cardiac to compute parameters of cardiac function and shape throughout the cardiac cycle. Techniques for analyzing advanced cardiac MRI such as tagged MRI and phase contrast MRI are presented along with future directions including machine learning based techniques for automating the image analysis process for cardiac MRI.

Click here for the Presentation

Trustworthy and Explainable Artificial Intelligence

Dr. Anh Nguyen

Abstract:

   Magnetic resonance imaging (MRI) is increasingly used in clinical settings to image the anatomy and physiological function of the cardiovascular system. While the image acquisition can be done in a reasonable amount of time, the analysis of these images to obtain quantitative measures of cardiac function is time consuming and requires considerable manual intervention. Consequently, only left ventricular volumes are measured at two points in the cardiac cycle, end-diastole and end-systole, to obtain stroke volume and ejection fraction. In this talk, techniques are presented to leverage the manual analysis that is typically done at two time points in the cardiac to compute parameters of cardiac function and shape throughout the cardiac cycle. Techniques for analyzing advanced cardiac MRI such as tagged MRI and phase contrast MRI are presented along with future directions including machine learning based techniques for automating the image analysis process for cardiac MRI.

Click here for the Presentation