Collaborating Authors

Learning concept graphs from text with stick-breaking priors

Neural Information Processing Systems

We present a generative probabilistic model for learning general graph structures, which we term concept graphs, from text. Concept graphs provide a visual summary of the thematic content of a collection of documents-a task that is difficult to accomplish using only keyword search. The proposed model can learn different types of concept graph structures and is capable of utilizing partial prior knowledge about graph structure as well as labeled documents. We describe a generative model that is based on a stick-breaking process for graphs, and a Markov Chain Monte Carlo inference procedure. Experiments on simulated data show that the model can recover known graph structure when learning in both unsupervised and semi-supervised modes.

Concept-Oriented Deep Learning Artificial Intelligence

Concepts are the foundation of human deep learning, understanding, and knowledge integration and transfer. We propose concept-oriented deep learning (CODL) which extends (machine) deep learning with concept representations and conceptual understanding capability. CODL addresses some of the major limitations of deep learning: interpretability, transferability, contextual adaptation, and requirement for lots of labeled training data. We discuss the major aspects of CODL including concept graph, concept representations, concept exemplars, and concept representation learning systems supporting incremental and continual learning.

KACC: A Multi-task Benchmark for Knowledge Abstraction, Concretization and Completion Artificial Intelligence

A comprehensive knowledge graph (KG) contains an instance-level entity graph and an ontology-level concept graph. The two-view KG provides a testbed for models to "simulate" human's abilities on knowledge abstraction, concretization, and completion (KACC), which are crucial for human to recognize the world and manage learned knowledge. Existing studies mainly focus on partial aspects of KACC. In order to promote thorough analyses for KACC abilities of models, we propose a unified KG benchmark by improving existing benchmarks in terms of dataset scale, task coverage, and difficulty. Specifically, we collect new datasets that contain larger concept graphs, abundant cross-view links as well as dense entity graphs. Based on the datasets, we propose novel tasks such as multi-hop knowledge abstraction (MKA), multi-hop knowledge concretization (MKC) and then design a comprehensive benchmark. For MKA and MKC tasks, we further annotate multi-hop hierarchical triples as harder samples. The experimental results of existing methods demonstrate the challenges of our benchmark. The resource is available at

Microsoft unveils Concept Graph: 'It's time AI learned some common sense'


Concept Graph contains more than 5.4 million concepts, which Microsoft says is far more than existing knowledge bases. Microsoft has released Concept Graph, a database of words linked to millions of concepts that it's using to help machines grasp meanings the way humans can when interpreting a sentence. As Microsoft notes, one thing that separates humans from machines when it comes to understanding a sentence is that humans have knowledge about concepts, such as dates, people, and animals. Humans also have the ability to conceptualize things, such as "cats are animals", or that a birthday is a significant date for a person. Machines don't have these qualities and that's why a computer could take the phrase "animals other than dogs such as cats" to mean that "cats are animals" or that "cats are dogs".

Note on Representing attribute reduction and concepts in concepts lattice using graphs Artificial Intelligence

Mao H. (2017, Representing attribute reduction and concepts in concept lattice using graphs. Soft Computing 21(24):7293--7311) claims to make contributions to the study of reduction of attributes in concept lattices by using graph theory. We show that her results are either trivial or already well-known and all three algorithms proposed in the paper are incorrect.