PRECISE Industry Day 2019

PRECISE’s Industry Day 2019 is an informal and engaging day-long symposium for leading executives and engineers involved in working towards the same goal of making the world a safer, better place through AI/ML.  The event allows industry the opportunity to see first-hand, the latest, ground-breaking research being undertaken by PRECISE. Given its success year-after-year, PRECISE's Industry Day has grown to become one of Penn Engineering's signature events.

In the past, the presentations, given by our Industry Day speakers, covered several areas, such as designing and developing Cyber-Physical Systems, Embedded Systems, Hybrid Control Systems, and IoT.  However, in 2019, PRECISE has shifted its focus to developing cutting-edge approaches, algorithms, tools and technologies for "Safe AI" - ensuring the safety of autonomous systems across various industries (e.g., automotive, medical, etc.)

AGENDA

  1. Session I:  How industries are using AI to make different aspects of life safer or more secure?
  2. Session II:  Featured Keynote
  3. Session III:  How PRECISE makes AI algorithms safer to ensure greater accuracy and fault-tolerance when AI is being used in these great AI-based systems that so many companies and organizations are creating to help the world?  
Some of the questions that our speakers may answers are listed below:
  • What is your company's overview of Safe Autonomy/AI?
  • What are your views on how verification technologies can help/hope to scale to highly sophisticated autonomous systems?
  • What are your views on how ever-increasing autonomy will change society in the future?
  • What do you prioritize, or deprioritize, when building an algorithm for safe autonomy?
  • How is autonomy be used in your product(s)?  What challenges/problems are you facing?
  • How do you process and learn from large volumes of customer-driven data?  How do you encourage data sharing to improve safety, and what type of data to be shared?
  • What system architecture enables smooth software updates and how do you evaluate software updates?
  • What is the most important open problem in safe autonomy that you are trying to solve? Why is it so hard to solve? What must be done to solve it?
  • What is your approach to assuring autonomous systems are safe?  What could be improved about this process?
  • Can the combination of simulation, real-world and controlled test environments provide robust autonomous systems for mass production?
  • What ways can Safe Autonomy/AI affect your business in the future?
  • How do you build social costs or human value into algorithms (e.g., when a system must decide to spare harm to drivers, passengers, or pedestrians)? 
  • What kind of road/AV test, if any, would be adequate for determining autonomous vehicle safety?  What contributes "sufficient confidence" in terms of passing the test versus gaming the test?
  • How "safe" is safe enough for deploying autonomous vehicles?  Do you rely primarily on testing?  How do you decide whether you have tested your system enough to call that "safe"?  How do you treat potential freaky corner cases; and how do you identify/resolve them?  
  • What behavior should be expected when an AV experiences abnormal/unexpected driving scenarios?
  • How can “miles driven” metric be used when the quality of the “miles” could be highly variable and not reflect the environment in which the AV is deployed?

We look forward to seeing you here!

Date & Time
Starts
Fri, October 25, 2019 @ 8:00am EDT
Ends
Fri, October 25, 2019 @ 7:00pm EDT
Venue
Glandt Forum
Singh Center for Nanotechnology
3205 Walnut Street
Philadelphia, PA 19104
Registration Fee
To be release shortly.
Event Contact
Agenda