Skip to content
Search events. View events.
 

All Categories

Welcome to the CU-Boulder Events Calendar.

Click for help in using calendar displays. Print the contents of the current screen.

Advanced Search

(New Search)
  From:
  To:



Submit
Event Details
Notify me if this event changes.Add this event to my personal calendar.Email this event to a friend.
Go Back
CompSci Colloquium: Nisar Ahmed (CU Boulder)
Start Date: 4/27/2017Start Time: 3:30 PM
End Date: 4/27/2017End Time: 4:30 PM
Event Description:
Blended Bayesian Intelligence for Collaborative Autonomous Systems (Big and Small) 

ABSTRACT: The age of autonomy has arrived. However, as autonomous systems are products of imperfect human engineering designed to make decisions in an uncertain world, the promise of ‘set-it-and-forget-it’ autonomy is still quite far off: autonomous systems will never operate out of the box “exactly right.” For sufficiently rich tasks that constantly push the technological cutting edge, they will encounter unexpected situations that require reasoning beyond their designed/immediate capabilities. As such, intelligent autonomy must not only be able to independently gather, process and act on information – it should also be cognizant of what it can and cannot accomplish, and know when and how to seek help.

Human-machine and machine-machine interaction are thus key components of autonomous system design, and these must naturally integrate with existing perception, planning, learning and reasoning algorithms for single agent autonomy. An autonomous robot, for instance, should ideally enable human stakeholders and users to fluidly (re-)delegate tasks to the system, assess information, and contribute  meaningful operational improvements -- without requiring users to “babysit,” act as a “Band-Aid” when all else fails, or think too hard about what the system is trying to do. When combined into large scale networked teams, intelligent autonomous robots should also be able to efficiently exchange and extract relevant information from one another to improve both local and global reasoning under uncertainty – even if they have heterogeneous/asymmetrical capabilities and tasks to complete. 

In this talk, Ahmed will discuss novel Bayesian approaches to collaborative human-robot and robot-robot reasoning that can be exploited from the outset in autonomous system design. The first part of the talk will focus on probabilistic modeling and inference techniques for augmenting autonomous sensing and state estimation algorithms with plug-in “human sensor” inputs, provided via user-friendly semantic interfaces (e.g. natural language, sketching). The second part of the talk will highlight the important issue of scaling autonomous Bayesian reasoning to large distributed networks. In particular, Ahmed will show how factorized decentralized data fusion (FDDF) methods can be used to create a “GitHub for uncertainty” that allows agents to efficiently exchange conditional chunks of statistical information on demand across disparate local world models. Results from collaborative human-robot target search and multi-platform target tracking applications show that these techniques allow autonomous teams to gracefully “fill in the gaps and cut the knots” for challenging uncertain reasoning problems – without undermining individual agent roles or ignoring the limitations of individual autonomous agents. 

BIO: Nisar Ahmed is an assistant professor of aerospace engineering sciences at the University of Colorado Boulder. His research interests are in the modeling and estimation for intelligent control of dynamical systems, especially for applications involving human-robot interaction, distributed sensor networks, and information fusion. He received his BS in engineering from Cooper Union, and PhD in mechanical engineering from Cornell University in 2012. He was a postdoctoral research associate in the Cornell Autonomous Systems Lab from 2012 to 2014. He was awarded the 2011 AIAA Guidance, Navigation, and Control Conference Best Paper Award; and an ASEE Air Force Summer Faculty Fellowship in 2014. His work has been funded by the Army, Air Force and industry, and he has also organized several workshops and symposia on autonomous robotics, sensor fusion and human-autonomy interaction. He is a member of the IEEE and an associate member of the AIAA Intelligent Systems Technical Committee.
 
Location Information:
Main Campus - Engineering Classroom Wing  (View Map)
1111 Engineering DR
Boulder, CO
Room: 265
Contact Information:
Name: Matt Hammer
Email: matthew.hammer@colorado.edu
This event is open to
  • Everyone
  • Of note:
    Light refreshments will be served.

    Calendar Software powered by Active Calendar and Active Data Calendar   
    Select item(s) to Search
    Select item(s) to Search
    Select item(s) to Search
    Select item(s) to Search

    Featured Events

    Today's Events