BECCA (Brain-Emulating Cognition and Control Architecture), a general purpose learning agent developed by research group of Brandon R. Rohrer (the Intelligent Systems, Robotics, and Cybernetics Group (ISRC) of Sandia National Laboratories (SNL), Albuquerque, New Mexico, USA).
BECCA open-source resources: Github repository | BECCA wiki
Email: Этот адрес электронной почты защищён от спам-ботов. У вас должен быть включен JavaScript для просмотра.
Когнитивная архитектура BECCA (Brain-Emulating Cognition and Control Architecture) была разработана исследовательской Группой кибернетики, робототехники и интеллектуальных систем (the Intelligent Systems, Robotics, and Cybernetics Group) в Sandia National Laboratories в г. Альбукерке, Нью-Мексико, США) при участии Брэндона Рорера (Brandon Rohrer) (работал до 2013 года).
Общие сведения
BECCA is a Brain-Emulating Cognition and Control Architecture. It was developed in order to perform general reinforcement learning, that is, to enable unmodeled systems operating in unstructured environments to perform unfamiliar tasks.It accomplishes this through complementary feature creation and reinforcement learning (RL) algorithms.
Both of these algorithms are novel and have characteristics that have not been demonstrated by previously existing algorithms.
BECCA is a general learning program for use in any robot or embodied system. It consists of an automatic feature creator and a model-based reinforcement learner.
When using BECCA, a robot learns to do whatever it is rewarded to do, and continues learning throughout its lifetime.
BECCA was designed to solve the problem of natural-world interaction.
The research goal is to place BECCA into a system with unknown inputs and outputs and have it learn to successfully achieve its goals in an arbitrary environment. The current state-of-the-art in solving this problem is the human brain.
As a result, BECCA's design and development is based heavily on insights drawn from neuroscience and experimental psychology. The development strategy emphasizes physical embodiment in robots.
Knowledge and experiences are represented using: Perceptual experiences are represented features and combinations of features. Knowledge about the world is represented as transitions between experiences.
Main components: BECCA is a solution to the general reinforcement learning problem. It consists of two parts, an unsupervised feature creator and a model-based incremental learner. Both are incremental and on-line, designed for a physically embodied agent operating in an unstructured environment.
Implementation: MATLAB, Python
Funding program: Sandia National Labs, internal R&D funding
Main general paradigms: Goal-directed behavior
Internet Groups and Mailing Lists
becca_users This is the Google group for users of BECCA (Brain-Emulating Cognition and Control Architecture) to share questions, answers, observations, and experiences.
This is a forum where BECCA users, whether they be academics, industry professionals, or hobbyists, can share their expertise and contribute to the development process. While its long-term research goals are exciting, BECCA is intended above all to be useful.
Please visit Brandon Rohrer's Youtube channel for videos describing his research on general learning machines.
An Open-source Artificial General Intelligence Platform: http://openbecca.org/ (not available?)
Additional forums, tools, and Documentation Wiki are available at openbecca.org, the homepage for the open source BECCA project.
The BECCA project welcomes volunteers who want to contribute to the goal of creating Artificial General Intelligence. http://openbecca.org/participate
Publications
1. BECCA: Reintegrating AI for Natural World Interaction - Brandon Rohrer
2. A developmental agent for learning features, environment models, and general robotics tasks - Brandon Rohrer
