The OpenCog Foundation

License: Affero GNU Public License

Web Page: http://wiki.opencog.org/w/Ideas

Mailing List: http://groups.google.com/group/opencog

We're creating an open source framework for Artificial General Intelligence, intended to one day express general intelligence at the human level and beyond. That is, we're undertaking a serious effort to build a thinking machine. We've developed a detailed plan, possess the ability to execute it, and we're proceeding with the hard work step-by-step. As compared to most academic and industry AI research, the OpenCog project has a very different focus. We are not centrally concerned with building more accurate classification algorithms, or more efficient computer vision systems, or better language processing or information retrieval algorithms, etc. Neither are we centrally concerned with building a program that performs only one specific task like play chess, diagnose diseases, answer trivia questions, or drive a car. We are concerned with generic intelligence and the inter-related cognitive processes it entails. OpenCog is a moderate-sized and active project with: a team in Hong Kong working on intelligent game characters a new OpenCog lab in Addis Ababa, Ethiopia, with a staff of 5 developers in the Americas, Europe, Asia, Africa, the Middle East, and Australasia, working for various employers and applying OpenCog on commercial and government projects many fantastic volunteer developers Our vision is a huge one and we are eagerly recruiting new developers! The OpenCog Project consists of a number of software projects. The core functionality is in a mostly C++ codebase hosted on Github. The OpenCog Foundation is a nonprofit organization, incorporated in the US state of Delaware. OCF is a democratically run organization, with a Board elected by its membership, where the latter is defined as individuals who have substantially contributed to the OpenCog project, and have asked to become formal members and been approved by the Board.

Projects

• AGI Language Learning of Tagalog Morphology One aspect of AGI that the OpenCog Foundation is currently addressing is language learning. My project will be to develop and implement steps towards AGI language learning of Tagalog morphology. This will be done within the context of OpenCog’s current processes and outlined steps for learning the syntax of a language. The major hurdle of my project will be devising a method for precisely splitting Tagalog words at morpheme boundary locations, prior to the application of links between morphemes.
• Improve and Validate Accuracy of RelEx2Logic Semantic Relation Extraction The aim of this project is to improve Relex2logic semantic relation extractor. Currently the extractor is not handling the basic type of sentence in English grammar. In this project Adjective clause rule, conjunction rule and wh-question rule will be added and there helper functions will be written.
• Improved reference resolution with reasoning Combine PLN, Relex2logic and ConceptNet to perform reasoning and select the best antecedent for each anaphore.
• Incremental real-time sub-graphs miner in C++ Implementing a real time subgraphs miner mining patterns in the Atomspace Incrementally.
• Natural Language Generation using RelEx and the Link Parser The task of this project is to generate natural English sentence(s) using RelEx [1] and the Link Parser [2] and some graph similarity computing algorithm. Ideally, when we input the RelEx relationships produced by applying RelEx to a sentence, the NL generation system will generate one or more natural, well-formed sentence(s), which have the same meaning as the original sentence. It focuses on improving the work that I have done for the OpenCog GsoC 2009 [5]
• PLN Inference on extracted semantic relationships RelEx parses a sentence using Link-Grammar which produces sub-graphs. By matching patterns on these, RelEx can extract information which – as of now – pertains to gender, number and part of speech. On the way to an AGI, RelEx – as part of the OpenCog system – would need to learn how to make inferences and deductions from its input. This can be achieved by mapping it with RelEx2Logic rules to and adapting the existing implementation of probabilistic logic networks.
• PLN Inference on extracted semantic relationships Combining of two or more RelEx feature graphs, can be used to extract relationshionships/ conclusions of two or more sentences, This need a service of a logical reasoner to achieve the task. The Probabilistic Logic Network(PLN) is a systematic framework designed to carrying out reasoning under this kind of situation is a tool which can be used for above task.
• Using DeSTIN and Neural Nets to Effectively Classify Images The task I would be working on is basically using neural networks to classify images using DeSTIN.I would be concentrating mainly on several aspects of implementation of uniform DeSTIN such as scale invariance. I would be exploring the working of the online clustering algorithm and its role in extraction of spatial features and positioning of centroids and extraction of spatial patterns into belief states.
• Using DeSTIN and SVM or Neural Nets to Effectively Classify Images Features extracted from DeSTIN for natural objects are not enough to provide a good basis for further AI applications. By doing appropriate image preprocessing and other parameter tuning, DeSTIN is capable to extract high quality features for observations.