We consider continuous state, continuous action batch reinforcement learning where the goal is to learn a good policy from a sufficiently rich trajectory generated by some policy. We study a variant of fitted Q-iteration, where the greedy action selection is replaced by searching for a policy in a restricted set of candidate policies by maximizing the average action values. We provide a rigorou...

Random Noise and RandomWalk algorithms are local search strategies that have been used for the problem of satisfiability testing (SAT). We present a Markov-chain based analysis of the performance of these algorithms. The performance measures we consider are the probability of finding a satisfying assignment and the distribution of the best solution observed on a given SAT instance. The analysis...

In this paper, we leverage both deep learning and conditional random fields (CRFs) for sequential labeling. More specifically, we propose a mixture objective function to predict labels either independent or correlated in the sequential patterns. We learn model parameters in a simple but effective way. In particular, we pretrain the deep structure with greedy layer-wise restricted Boltzmann mach...

Because general and extended logic programs behave nonmonotonically, it is in general diicult to predict how even minor changes to such programs will aaect their meanings. This paper shows that for a restricted class of extended logic programs | those with signings | it is possible to state a fairly general theorem comparing the entailments of programs. To this end, we generalize (to the class ...

Confidential algorithm for the approximate graph vertex covering problem is presented in this article. It can preserve privacy of data at every stage computation, which very important context cloud computing. Security of~our solution based on fully homomorphic encryption scheme. The time complexity and security aspects considered are described.

This paper looks at the problem of multi-join query optimization for symmetric multiproceasore. Optimizrtlion algorithms based on dynamic programming and greedy heuristics are described that, unlike traditional methods, include memory resources and pipelining in their cost model. An analytical model is presented and used to compare the quality of plans produced by each optimization algorithm. E...

Given a Gaussian Markov random field, we consider the problem of selecting a subset of variables to observe which minimizes the total expected squared prediction error of the unobserved variables. We first show that finding an exact solution is NP-hard even for a restricted class of Gaussian Markov random fields, called Gaussian free fields, which arise in semi-supervised learning and computer ...