ai-lect2 - Last time: A driving example: Beobots Goal:...

Info iconThis preview shows pages 1–10. Sign up to view the full content.

View Full Document Right Arrow Icon
1 Last time: A driving example: Beobots Goal:  build robots that can operate in unconstrained environments  and that can solve a wide variety of tasks. We have: Lots of CPU power Prototype robotics platform Visual system to find interesting objects in the world Visual system to recognize/identify some of these objects Visual system to know the type of scenery the robot is in We need to: Build an internal representation of the world Understand what the user wants Act upon user requests / solve user problems
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
2 Beowulf + Robot = “Beobot”
Background image of page 2
3 Prototype Stripped-down version of proposed general system, for simplified goal: drive around USC olympic track, avoiding obstacles Operates at 30fps on quad-CPU Beobot; Layout & saliency very robust; Object recognition often confused by background clutter.
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
4 Major issues How to represent knowledge about the world? How to react to new perceived events? How to integrate new percepts to past experience? How to understand the user? How to use reasoning to decide on the best course of action? How to communicate back with the user? How to plan ahead? How to learn from experience?
Background image of page 4
5 General architecture
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
6 Ontology
Background image of page 6
7 The task-relevance map Scalar topographic map, with higher values at more relevant locations Navalpakkam & Itti, BMCV’02
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
8 More formally: how do we do it? - Use ontology to describe categories, objects and relationships: Either with unary predicates, e.g., Human(John), Or with reified categories, e.g., John   Humans, And with rules that express relationships or properties, e.g.,  2200 x Human(x)   SinglePiece(x)   Mobile(x)   Deformable(x) - Use ontology to expand concepts to related concepts: E.g., parsing question yields “LookFor(catching)” Assume a category HandActions and a taxonomy defined by catching   HandActions, grasping   HandActions, etc. We can expand “LookFor(catching)” to looking for other actions in the  category where catching belongs through a simple expansion rule: 2200 a,b,c   a   c   b   c   LookFor(a)   LookFor(b)
Background image of page 8
9 Last Time: Acting Humanly: The Full Turing Test Alan Turing's 1950 article  Computing Machinery and Intelligence  discussed  conditions for considering a machine to be intelligent “Can machines think?”  ←→  “Can machines behave intelligently?” The Turing test (The Imitation Game): Operational definition of intelligence. Computer needs to posses:
Background image of page 9

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 10
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 43

ai-lect2 - Last time: A driving example: Beobots Goal:...

This preview shows document pages 1 - 10. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online