CogPrime, a comprehensive architecture for embodied Artificial General Intelligence, is reviewed, covering the core architecture and algorithms, the underlying conceptual motivations, and the emergent structures, dynamics and functionalities expected to arise in a completely implemented CogPrime system once it has undergone appropriate experience and education. A qualitative argument is sketched, in favor of the assertion that a completed CogPrime system, given a modest amount of experience in an embodiment enabling it to experience a reasonably rich human-like world, will give rise to human-level general intelligence (with significant difference from humans, and with potential for progress beyond this level).

Most object recognition systems are based on supervised learning, and dealing with 2D image which gives little infor?mation about the background and foreground. Even state of art supervised object recognition can’t promise good result, let alone unsupervised based system. With the advent of Kinect, it’s more easy to get 2.5D data. This paper proposes an innovative unsu?pervised based object recognition system. The first step is to fit plane and use space based clustering to extract objects from the scene. The second step is segmenting objects based on geometry primitive fitting, constructing graph model for each object based on the segmentation result. Finally, embed the graph model to normal vector space and calculate the similarity between objects with extended Earth Mover’s Distance. The experiment gives a good result.

A new approach to translating between natural language expressions and hypergraph-based semantic knowledge representations is proposed. Language comprehension is formulated in terms of homomorphisms mapping syntactic parse trees into semantic hypergraphs, and language generation as constraint satisfaction based on constraints derived via applying the inverse relations of these homomorphisms. This provides an elegant approach to implementing semantically savvy NLP systems, and also to thinking about the feedbacks between syntactic and semantic processing that are the crux of generally intelligent NLP. A prototype of the approach created using the link parser and the OpenCog Atom semantic representation is described, and initial results presented. Routes to extending this prototype into something useful for aiding generally intelligent dialogue systems are discussed.

Fishgram, a novel algorithm for recognizing frequent or otherwise interesting sub-hypergraphs in large, heterogeneous hypergraphs, is presented. The algorithm's implementation the OpenCog integrative AGI framework is described, and concrete examples are given showing the patterns it recognizes in OpenCog's hypergraph knowledge store when the OpenCog system is used to control a virtual agent in a game world. It is argued that Fishgram is well suited to fill a critical niche in OpenCog and potentially other integrative AGI architectures: scalable recognition of relatively simple patterns in heterogeneous, potentially rapidly-changing data.