Briefing: Skilligent Drives Adaptive Robot Learning

November 25th, 2008  |  Published in Artificial Intelligence, Research Reports, Robotics

SciVestor provides open-source industry research reports available for free download.  These reports are licensed under a Creative Commons 2.0 license - Attribution.  Registration on SciVestor.com is required to download this report.  Once you register, SciVestor provides digital storage for all your requested reports.

You can register and download the presentation here.

Summary: Skilligent is developing a new breed of software with the prospect of teaching robots skills and behaviors without any conventional programming. The report highlights the SciVestor Model for Robotics Training Strategies research that points to three disparate strategies for implementing behaviors into robots.

1) Fixed training strategies rely on a sequenced set of preprogrammed skills.
2) Codified training relies on behaviors codified by the robot programmer.
3) Adaptive methods abstract the teaching of robot skills from the code-base, freeing the trainer from technical competency.

The report also discusses the work of Novamente LLC and The Electric Sheep Company who are training artificial intelligence robots in SecondLife.

Findings: This report recommends strategies for Investors, Robot Manufacturers, and Robot Technology Consumers when considering Skilligent solutions.

Author: Ray Renteria
Price: Free. Licensed under a Creative Commons (cc) Attribution license.
Length: 3 pages
Research ID: 7R-004
Concepts Discussed: Adaptive robot learning, Skilligent, SecondLife, Novamente, Sergey Popov, Microsoft Robotics Studio, SciVestor model for robotics training strategies, Roomba, EOD Robots, Human Robot Interaction.

Leave a Response


Categories