Belief Propagation google
Belief propagation, also known as sum-product message passing is a message passing algorithm for performing inference on graphical models, such as Bayesian networks and Markov random fields. It calculates the marginal distribution for each unobserved node, conditional on any observed nodes. Belief propagation is commonly used in artificial intelligence and information theory and has demonstrated empirical success in numerous applications including low-density parity-check codes, turbo codes, free energy approximation, and satisfiability. The algorithm was first proposed by Judea Pearl in 1982, who formulated this algorithm on trees, and was later extended to polytrees. It has since been shown to be a useful approximate algorithm on general graphs. …

DSS-MAS google
For some decision processes a significant added value is achieved when enterprises’ internal Data Warehouse (DW) can be integrated and combined with external data gained from web sites of competitors and other relevant Web sources. In this paper we discuss the agent-based integration approach using ontologies (DSS-MAS). In this approach data from internal DW and external sources are scanned by coordinated group of agents, while semantically integrated and relevant data is reported to business users according to business rules. After data from internal DW, Web sources and business rules are acquired, agents using these data and rules can infer new knowledge and therefore facilitate decision making process. Knowledge represented in enterprises’ ontologies is acquired from business users without extensive technical knowledge using user friendly user interface based on constraints and predefined templates. The approach presented in the paper was verified using the case study from the domain of mobile communications with the emphasis on supply and demand of mobile phones. …

Deep Sparse Subspace Clustering google
In this paper, we present a deep extension of Sparse Subspace Clustering, termed Deep Sparse Subspace Clustering (DSSC). Regularized by the unit sphere distribution assumption for the learned deep features, DSSC can infer a new data affinity matrix by simultaneously satisfying the sparsity principle of SSC and the nonlinearity given by neural networks. One of the appealing advantages brought by DSSC is: when original real-world data do not meet the class-specific linear subspace distribution assumption, DSSC can employ neural networks to make the assumption valid with its hierarchical nonlinear transformations. To the best of our knowledge, this is among the first deep learning based subspace clustering methods. Extensive experiments are conducted on four real-world datasets to show the proposed DSSC is significantly superior to 12 existing methods for subspace clustering. …

Advertisements