All interested researchers and students are invited to attend. No registration is required for the short course. Seating is limited and is available on a first come, first served basis.
Program
Talk 1 – “Foundation Models and Assured Autonomy for Self-Driving Vehicles” – Eric Wolff, G
- Large-scale foundation models are increasingly critical for the perception and decision-making capabilities of self-driving vehicles.
- At GM, our focus is on personal autonomous vehicles that can drive safely and efficiently anywhere a human can. To assure performance across such a varied and uncertain environment, we rely on foundation models to underpin our driving policies.
- I will discuss how these foundation models play key roles in improving perceptual robustness, anticipating interactions with other road users, and reasoning about longtail scenarios.
Talk 2 – “Assured Autonomy: Building AI Pilots We Can Trust” – Cristina Sorice, Shield AI
- Overview of Hivemind, Shield AI’s software product for developing and deploying AI pilots
- Examples of Hivemind in the real world (including in space)
- Shield AI’s approach to Assured Autonomy
Talk 3 – “Current Approach for Space Mission Verification and Validation (V&V)” – Issa Nesnas, JPL
- Defining autonomy
- Today’s approach. Mission assurance by design: the collaborative process of mission formulation where scientists and engineers work together to arrive at a feasible mission concept with minimal risks; hardware design for robustness, and testing of the hardware; difficulties in assuring more intelligent systems.
- Limitations of current V&V approaches
- Examples of mission that autonomous systems could enable
Talk 4 – “Forward Looking Space Exploration Missions” – James Keane, JPL
- Important scientific questions from the decadal survey, and potential future interesting mission concepts 20-30 years from now
- Use cases for autonomy to enhance scientific missions: long lunar traverse, data and power limited missions, multi-robot missions
- Is autonomy essential?
Travel Information
Parking (for Visitors and for JPL Personnel)
For Visitors: From the Arroyo Parkway, turn right (east) on Del Mar Avenue. Proceed approximately one and a quarter miles. The Caltech campus will be on your right. Turn right (south) onto Wilson Avenue. Turn right into the North Wilson Structure and park in an unmarked spot.
Visitor parking permits may be purchased through Caltech’s mobile parking payment system Honk Mobile. Visitors can purchase a parking session at payment signs in the campus lots including scan-to-pay (QR code) and text-to-pay options. Study participants: use the parking instructions sent via email to purchase a parking permit through Caltech’s mobile parking payment system Honk Mobile.
For JPL Personnel: JPLers may use their JPL hang tag for parking. Please park in an unmarked spot in the North Wilson Structure located on Wilson Avenue.
