Accession Number : ADA508405


Title :   Learning On-Line from a Few Examples


Descriptive Note : Final rept.


Corporate Author : MASSACHUSETTS INST OF TECH CAMBRIDGE CENTER FOR BIOLOGICAL AND COMPUTATIONAL LEARNING


Personal Author(s) : Poggio, Tomaso


Full Text : https://apps.dtic.mil/dtic/tr/fulltext/u2/a508405.pdf


Report Date : 31 Aug 2009


Pagination or Media Count : 34


Abstract : Learning from the Brain with Applications to Scene Understanding and to Speech Synthesis and Understanding -- EXECUTIVE SUMMARY: The problem of learning represents a gateway to understanding intelligence in brains and machines and to making intelligent machines that learn from experience. Since our brain represents the best known example of a learning machine, we have developed algorithms directly based on the architecture of the cortex. In the process we have demonstrated systems in applications of interest to IPTO. In particular, we have demonstrated * a system for scene understanding in street environments * video realistic synthesis of a speaking agent.


Descriptors :   *SPEECH , *SCENE GENERATION , *LEARNING , *BRAIN , *SYNTHESIS , ONLINE SYSTEMS , HIERARCHIES , VIDEO IMAGES , MULTIAGENT SYSTEMS , KERNEL FUNCTIONS , LEARNING MACHINES , INTELLIGENCE , ALGORITHMS , RECOGNITION


Subject Categories : Linguistics
      Numerical Mathematics
      Bionics


Distribution Statement : APPROVED FOR PUBLIC RELEASE