USC AI Futures Symposium: January 12-13, 2021

Interacting with AI systems is now commonplace, whether conversing about a dinner reservation, directing them to drive a car, or requesting music. But it is very apparent that they do not communicate like we do. They know very little about the world, are ignorant of social conventions, cannot learn our quirks from numerous conversations, and seem clearly confounded when something unexpected happens.

When will we interact with AI systems as if they were one of us? The challenge is immense, and this symposium will present critical areas of future research to achieve this vision.


Invited Speakers

Elizabeth Churchill

Elizabeth Churchill is a director at Google, a research scientist, and a UX leader with a background in Psychology, Artificial Intelligence, Cognitive Science, Human Computer Interaction, and Cognitive Ergonomics.

Antonio Damasio

Antonio Damasio is University Professor, David Dornsife Professor of Neuroscience, Psychology and Philosophy, and Director of the Brain and Creativity Institute at the University of Southern California, Los Angeles.

Prem Natarajan

Prem Natarajan is a Vice President in Amazon’s Alexa and leads a multidisciplinary science, engineering, and product organization which improves customer experience worldwide through advances in dialog modeling, natural language understanding, entity linking and resolution, and related machine learning technologies.


Tuesday, January 12, 2021
8:45-9:00a PST
Welcome and opening remarks Yannis Yortsos
Dean, USC Viterbi School of Engineering
Zohrab A. Kaprielian Dean’s Chair in Engineering and Chester Dolley Professor of Chemical and Petroleum Engineering
Cyrus Shahabi
Chair, USC Department of Computer Science, Helen N. and Emmett H. Jones Professorship in Engineering and Professor of Computer Science, Electrical and Computer Engineering, and Spatial Sciences
Craig Knoblock
Keston Executive Director, Information Sciences Institute
Research Professor of Computer Science and Spatial Sciences
9:00-10:00a PST
Session I: Harnessing InformationSession Chair: Ralph Weischedel

"Controlling Misinformation"

Emilio FerraraAssociate Professor of Communication (USC Annenberg) and Computer Science (USC Viterbi)

"Threading Data"

Craig KnoblockKeston Executive Director, Information Sciences Institute;
Research Professor of Computer Science and Spatial Sciences

"Seeing is Not Believing"

Wael Abd-AlmageedResearch Associate Professor, Electrical and Computer Engineering Department

"Multi-Source Synthesis"

Ram NevatiaFletcher Jones Professor in Computer Science and Professor of Computer Science and Electrical and Computer Engineering

"Provenance Matters"

Liz BoscheeDirector, ISI Boston Office

10:00-10:15a PST
10:15-11:00a PST
Invited Talk I

"Deep Understanding"

Prem NatarajanAmazon

11:00-11:15a PST
11:15a-12:00p PST
Invited Talk II

"Intelligences, Brains, and Humanity"

Antonio DamasioUSC

12:00-12:15p PST
12:15-1:15p PST
Session II: Acting Responsibly and EthicallySession Chair: Sven Koenig

"Fair Behaviors"

Kristina LermanResearch Professor;
Research Team Leader;
Principal Scientist

"Cultural Differences"

Fred MorstatterResearch Assistant Professor of Computer Science;
Research Lead

"Inclusive Design"

Shri NarayananResearch Director, University Professor and Niki & C. L. Max Nikias Chair in Engineering

"Bias and Opinions"

Keith BurghardtComputer Scientist

"Making Good Decisions"

Bistra DilkinaAssociate Professor of Computer Science

1:15-1:30p PST
1:30-2:15p PST
Panel I: AI agency: legality, responsibility, ethics

Panel Chair: Kristina Lerman
Humans bestow trust carefully to other humans, but rarely to AI systems in matters of importance. What research needs to be done to improve trustworthiness of AI systems? What are the factors that influence how people decide to trust AI systems? What are the limits of purely algorithmic solutions? How should we approach the design of trustworthy AI systems?


Oren EtzioniAllen Institute for AI

Oren Etzioni is Chief Executive Officer at the Allen Institute for AI and Professor Emeritus at UW. He has founded or co-founded several companies, including Farecast (acquired by Microsoft) and write on AI for The New York Times, Wired, and Nature.

Patrick LinCal Poly SLO

Patrick Lin is the director of the Ethics + Emerging Sciences Group at Cal Poly, where he is a full philosophy professor, and is currently affiliated with Stanford Law School, the 100 Year Study on AI, Czech Academy of Sciences, and the World Economic Forum.

Cynthia RudinDuke University

Cynthia Rudin is a professor of computer science, electrical and computer engineering, and statistical science at Duke University. She is a three-time winner of the INFORMS Innovative Applications in Analytics Award,and was named as one of the "Top 40 Under 40" by Poets and Quants in 2015.

Wednesday, January 13, 2021
8:45-9:00a PST
9:00-10:00a PST
Session III: Learning IndependentlySession Chair: Yan Liu

"All the World's Knowledge"

Pedro SzekelyResearch Associate Professor;
Research Director;
Principal Scientist

"Thinking About How to Learn"

Greg Ver SteegResearch Associate Professor

"The Science of Forecasting"

Aram GalstyanDirector of Artificial Intelligence Division;
Research Associate Professor;
Research Director;
Principal Scientist

"The More You Know"

Xiang RenAssistant Professor of Computer Science;
Research Lead

"Compartmentalizing Data"

Jose-Luis AmbiteResearch Associate Professor;
Research Team Leader

10:00-10:15a PST
10:15-11:00a PST
Poster session
Presenter Title of poster
Avijit Thawani NLP with Numbers
Mingxuan Yue Deep Trajectory Clustering for Mobility Behavior Analysis
George Constantinou Spatial Keyframe Extraction of Mobile Videos for Efficient Object Detection at the Edge
Daniel Garijo MINT: AI for Integrated Environmental Modeling
Yolanda Gil AI Scientists: Automated Hypothesis-Driven Scientific Discovery
Jiao Sun CrimeForecaster: Crime Prediction by Exploiting the Geographical Neighborhoods’ Spatiotemporal Dependencies
Huy Nghiem Investigation of Inter-racial Prejudice on Twitter during the COVID-19 Pandemic
Taoan Huang Learning Node-Selection Strategies in Bounded Suboptimal Conflict-Based Search for Multi-Agent Path Finding
Zekun Li An Automatic Approach for Generating Rich, Linked Geo-Metadata from Historical Map Images
Weiwei Duan Weakly-Supervised Image Classification Using Target-Guided Generative Clustering Models
Yijun Lin Building Autocorrelation-Aware Representations for Fine-Scale Spatiotemporal Prediction
Aaron Ferber MIPaaL: Mixed Integer Program as a Layer
Ninareh Mehrabi Exacerbating Algorithmic Bias through Fairness Attacks
Justin Cho Grounding Conversations with Improvised Dialogues
Minh Pham Hunter: A Human-inspired Probabilistic Approach for Active Error Detection
Dimitris Stripelis Federated Learning Training Protocols in Data and Computationally Heterogeneous Environments
Basel Shbita Building Linked Spatio-Temporal Data from Vectorized Historical Maps
Haowen Lin SemiFed: Semi-supervised Federated Learning with Consistency and Pseudo-Labeling Analysis
Umang Gupta Controllable Guarantees for fair outcomes via contrastive information estimation
Stefanos Nikolaidis A Quality Diversity Approach to Automatically Generating HRI Scenarios in Shared Autonomy
Jiaoyang Li Lifelong Multi-Agent Path Finding in Large-Scale Warehouses
11:00-11:15a PST
11:15a-12:00p PST
Invited Talk III

"Inclusive Design"

Elizabeth ChurchillGoogle

12:00-12:15p PST
12:15-1:15p PST
Session IV: Taking on Critical TasksSession Chair: Pedro Szekely

"Responding to Unexpected Events"

Jon MayResearch Assistant Professor;
Research Lead

"Protecting Children"

Mayank KejriwalResearch Assistant Professor of Industrial and Systems Engineering;
Research Lead

"Financial Accountability"

Jay PujaraResearch Assistant Professor;
Research Lead

"AI Scientists"

Yolanda GilResearch Professor of Computer Science and of Spatial Sciences;
Director for Major Strategic AI and Data Science Initiatives
Director of Knowledge Technologies; Principal Scientist

"Trust and Security"

Marjorie FreedmanSupervising Computer Scientist;
Research Team Leader

1:15-1:30p PST
1:30-2:15p PST
Panel II: AI in high-stakes situations

Panel Chair: Shri Narayanan
AI systems hold the promise of supporting our ability to navigate high-stakes scenarios such as crises (e.g., natural disasters, pandemics, and social unrest) that are placing increasing demands on people, especially when capacities are challenged by physical and psychological toll, cognitive load, and other situational and resource constraints. Can AI systems be collaborators and work together with people in high-stakes situations with complex social dynamics? What are some possibilities that can be imagined and enabled? What are the challenges and potential pitfalls in trying to realize this promise of AI? What do we need to do to overcome these?


Wil CorveyDARPA

William Corvey is a Program Manager in DARPA’s Information Innovation Office (I2O). He is the current program manager for several efforts including the Low Resource Languages for Emergent Incidents (LORELEI) program.

Ibrahim DemirUniversity of Iowa

Ibrahim Demir is Assistant Professor in Civil and Environmental Engineering and Electrical and Computer Engineering, and Director of UI Hydroinformatics with. research interests in disaster communication, intelligent systems, AI and machine learning, and scientific visualization.

Marta GonzalezUC Berkeley

Marta Gonzalez is Associate Professor of Civil and Environmental Engineering, UC Berkeley and directs the HumNet Lab that develops "numerical models and computational tools to better characterize and understand human interactions in the built and natural environments.""

2:15-2:30p PST
Closing Remarks

Maja MataricUSC Interim Vice President of Research;
Chan Soon-Shiong Chair and Distinguished Professor of Computer Science, Neuroscience, and Pediatrics

Craig KnoblockKeston Executive Director, Information Sciences Institute;
Research Professor of Computer Science and Spatial Sciences