ANew Inference Framework for Dependency Networks
نویسندگان
چکیده
Dependency networks (DNs) have been receiving more attention recently because their structures and parameters can be easily learned from data. The full conditional distributions (FCDs) are known conditions of DNs. Gibbs sampling is currently the most popular inference method on DNs. However, sampling methods converge slowly and it can be hard to diagnose their convergence. In this article, we introduce a set of linear equations to describe the relations between joint probability distributions (JPDs) and FCDs. These equations provide a novel perspective to understand reasoning on DNs. Based on these linear equations, we develop both exact and approximate algorithms for inference on DNs. Experiments show that the proposed approximate algorithms can produce effective results by maintaining low computational complexity.
منابع مشابه
Building Subcomponents in the Cooperative Coevolution Framework for Training Recurrent Neural Networks: Technical Report
Cooperative coevolution decomposes a large problem into its subcomponents and uses evolutionary algorithms for solving them in order to gradually solve the large problem. This paper uses cooperative coevolution framework for training recurrent neural networks for grammatical inference problems. In the past, different encoding schemes were used to build subcomponents from the neural network for ...
متن کاملMean Field Inference in Dependency Networks: An Empirical Study
Dependency networks are a compelling alternative to Bayesian networks for learning joint probability distributions from data and using them to compute probabilities. A dependency network consists of a set of conditional probability distributions, each representing the probability of a single variable given its Markov blanket. Running Gibbs sampling with these conditional distributions produces ...
متن کاملAn Efficient Cluster Head Selection Algorithm for Wireless Sensor Networks Using Fuzzy Inference Systems
An efficient cluster head selection algorithm in wireless sensor networks is proposed in this paper. The implementation of the proposed algorithm can improve energy which allows the structured representation of a network topology. According to the residual energy, number of the neighbors, and the centrality of each node, the algorithm uses Fuzzy Inference Systems to select cluster head. The alg...
متن کاملThe Prediction of Forming Limit Diagram of Low Carbon Steel Sheets Using Adaptive Fuzzy Inference System Identifier
The paper deals with devising the combination of fuzzy inference systems (FIS) and neural networks called the adaptive network fuzzy inference system (ANFIS) to determine the forming limit diagram (FLD). In this paper, FLDs are determined experimentally for two grades of low carbon steel sheets using out-of-plane (dome) formability test. The effect of different parameters such as work hardening...
متن کاملRobust Classification using Mixtures of Dependency Networks
Dependency networks have previously been proposed as alternatives to e.g. Bayesian networks by supporting fast algorithms for automatic learning. Recently dependency networks have also been proposed as classification models, but as with e.g. general probabilistic inference, the reported speed-ups are often obtained at the expense of accuracy. In this paper we try to address this issue through t...
متن کامل