[AI Seminar Series] Seminar Dr. Rajeev Yasarla, Friday November 20th, 12-1pm, MRB Seminar Room

Vassilis Tsotras vassilis.tsotras at ucr.edu
Sat Nov 14 10:48:57 PST 2026


The next AI Seminar will be on Friday November 20th, 12-1pm, in the MRB
Seminar Room (1st floor).

*** Pizza and refreshments will be provided ****

To keep track of the number of attendees, please *register* at:
https://www.eventbrite.com/e/ai-seminar-series-tickets-2004845088629

The talk will be given by* Dr. Rajeev Yasarla*, Qualcomm AI Research

TITLE: Learning Beyond Observed Data: Uncertainty, Latent World Modeling,
and Multi-Agent Play

ABSTRACT:
End-to-end autonomous driving systems typically learn from static driving
logs, where each scene captures only one realized future. Yet robust
driving requires reasoning under uncertainty, anticipating alternative
outcomes, and understanding how other road users may respond to the
vehicle’s actions. This talk presents three complementary approaches that
address these challenges through a progression from uncertainty-aware
learning to generative and interactive scenario modeling. RoCA (ICML 2026)
learns uncertainty-aware representations to improve robustness and
adaptation across driving environments. GeRo (NeurIPS 2026) generates
language-grounded future scenarios that connect ego planning,
surrounding-agent motion prediction, and descriptions of evolving
maneuvers, enabling learning beyond individual recorded trajectories. MAPLE
(NeurIPS 2026) introduces reactive multi-agent interactions within latent
scenario rollouts, allowing driving policies to learn from diverse
counterfactual experiences in which the ego vehicle and surrounding agents
respond to one another. Together, these approaches illustrate how
uncertainty modeling, future generation, and interactive experience can
enrich the training of autonomous driving systems. More broadly, they
explore a shift in learning-based decision-making from relying on observed
outcomes to learning from plausible alternatives and their consequences.

Bio:
Rajeev Yasarla is a Staff Researcher at Qualcomm AI Research, where he
develops predictive visual intelligence for robust embodied agents and
leads projects in end-to-end autonomous driving, world-action modeling,
multimodal learning, and dynamic 3D perception. He received his Ph.D. in
Electrical Engineering from Johns Hopkins University in 2022, advised by
Professor Vishal M. Patel, and his B.Tech. and M.Tech. in Electrical
Engineering from the Indian Institute of Technology Madras in 2016. His
research connects uncertainty-aware and data-efficient learning, temporal
and 3D/4D scene understanding, vision–language–action models, and latent
world models with prediction, planning, and control under distribution
shift. His recent work includes RoCA, GeRo, and MAPLE, which explore robust
adaptation, language-grounded future generation, and reactive multi-agent
learning for autonomous driving. His research has appeared in NeurIPS,
CVPR, ICCV, ECCV, ICML, WACV, IEEE TIP, and IEEE TPAMI, and received the
Best Paper Award at IEEE ICIP 2021.

------------------------------------
Sponsored by the RAISE at UCR Institute, the AI Seminar Series presents
speakers working on cutting edge Foundational AI or applying AI in their
research. The goal of these seminars is to inform the UCR community about
current trends in AI research and promote collaborations between faculty in
this emerging field. These seminars are open to interested faculty and
graduate/undergraduate students. Please forward this email to other
colleagues or students in your lab that may be interested. After the seminar a
discussion will follow for questions, open problems, ideas for possible
collaborations etc.

Sincerely,
Vassilis Tsotras
Professor, CSE Department
co-Director, RAISE at UCR Institute

Amit Roy-Chowdhury
Professor, ECE Department
co-Director, RAISE at UCR Institute
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ucr.edu/pipermail/raise-seminar/attachments/20261114/f5dd5878/attachment.htm>


More information about the raise-seminar mailing list