This paper proves that validity and satisfiability of assertions in the Fuzzy Description Logic based on infinite-valued Product Logic with universal and existential quantifiers (which are non-interdefinable) is decidable when we only consider quasi-witnessed interpretations. We prove that this restriction is neither necessary for the validity problem (i.e., the validity of assertions in the Fuzzy Description Logic based on infinite-valued Product Logic is decidable) nor for the positive satisfiability problem, because quasi-witnessed interpretations are particularly adequate for the infinite-valued Product Logic. We give an algorithm that reduces the problem of validity (and satisfiability) of assertions in our Fuzzy Description Logic (restricted to quasi-witnessed interpretations) to a semantic consequence problem, with finite number of hypothesis, on infinite-valued propositional Product Logic.

The satisfiability problem for monadic infinite-valued Gödel logic is known to be undecidable. We identify a fragment of this logic extended with strong negation whose satisfiability is not only decidable but it is decidable within classical logic. We use this fragment to formalize the rules of CADIAG-2, a well performing fuzzy expert system assisting in the differential diagnosis in internal medicine. A (classical) satisfiability check of the resulting formulas allowed the detection of some errors in the rules of the system.

Terrorism has led to many problems in Thai societies, not only property damage but also civilian casualties. Predicting terrorism activities in advance can help prepare and manage risk from sabotage by these activities. This paper proposes a framework focusing on event classification in terrorism domain using fuzzy inference systems (FISs). Each FIS is a decision-making model combining fuzzy logic and approximate reasoning. It is generated in five main parts: the input interface, the fuzzification interface, knowledge base unit, decision making unit and output defuzzification interface. Adaptive neuro-fuzzy inference system (ANFIS) is a FIS model adapted by combining the fuzzy logic and neural network. The ANFIS utilizes automatic identification of fuzzy logic rules and adjustment of membership function (MF). Moreover, neural network can directly learn from data set to construct fuzzy logic rules and MF implemented in various applications. FIS settings are evaluated based on two comparisons. The first evaluation is the comparison between unstructured and structured events using the same FIS setting. The second comparison is the model settings between FIS and ANFIS for classifying structured events. The data set consists of news articles related to terrorism events in three southern provinces of Thailand. The experimental results show that the classification performance of the FIS resulting from structured events achieves satisfactory accuracy and is better than the unstructured events. In addition, the classification of structured events using ANFIS gives higher performance than the events using only FIS in the prediction of terrorism events.

Fuzzy Description Logics (DLs) are a family of logics which allow the representation of (and the reasoning with) structured knowledge affected by vagueness. Although most of the not very expressive crisp DLs, such as ALC, enjoy the Finite Model Property (FMP), this is not the case once we move into the fuzzy case. In this paper we show that if we allow arbitrary knowledge bases, then the fuzzy DLs ALC under Lukasiewicz and Product fuzzy logics do not verify the FMP even if we restrict to witnessed models; in other words, finite satisfiability and witnessed satisfiability are different for arbitrary knowledge bases. The aim of this paper is to point out the failure of FMP because it affects several algorithms published in the literature for reasoning under fuzzy ALC.

Recently, Wang et al. discussed the properties of fuzzy information systems under homomorphisms in the paper [C. Wang, D. Chen, L. Zhu, Homomorphisms between fuzzy information systems, Applied Mathematics Letters 22 (2009) 1045-1050], where homomorphisms are based upon the concepts of consistent functions and fuzzy relation mappings. In this paper, we classify consistent functions as predecessor-consistent and successor-consistent, and then proceed to present more properties of consistent functions. In addition, we improve some characterizations of fuzzy relation mappings provided by Wang et al.

Medical diagnosis process involves many levels and considerable amount of time and money are invariably spent for the first level of diagnosis usually made by the physician for all the patients every time. Hence there is a need for a computer based system which not only asks relevant questions to the patients but also aids the physician by giving a set of possible diseases from the symptoms obtained using logic at inference. In this work, an ICD10 based Medical Expert System that provides advice, information and recommendation to the physician using fuzzy temporal logic. The knowledge base used in this system consists of facts of symptoms and rules on diseases. It also provides fuzzy severity scale and weight factor for symptom and disease and can vary with respect to time. The system generates the possible disease conditions based on modified Euclidean metric using Elders algorithm for effective clustering. The minimum similarity value is used as the decision parameter to identify a disease.

We introduce the first temporal-difference learning algorithms that converge with smooth value function approximators, such as neural networks. Conventional temporal-difference (TD) methods, such as TD($\lambda$), Q-learning and Sarsa have been used successfully with function approximation in many applications. However, it is well known that off-policy sampling, as well as nonlinear function approximation, can cause these algorithms to become unstable (i.e., the parameters of the approximator may diverge). Sutton et al (2009a,b) solved the problem of off-policy learning with linear TD algorithms by introducing a new objective function, related to the Bellman-error, and algorithms that perform stochastic gradient-descent on this function. In this paper, we generalize their work to nonlinear function approximation. We present a Bellman error objective function and two gradient-descent TD algorithms that optimize it. We prove the asymptotic almost-sure convergence of both algorithms for any finite Markov decision process and any smooth value function approximator, under usual stochastic approximation conditions. The computational complexity per iteration scales linearly with the number of parameters of the approximator. The algorithms are incremental and are guaranteed to converge to locally optimal solutions.

Existing value function approximation methods have been successfully used in many applications, but they often lack useful a priori error bounds. We propose approximate bilinear programming, a new formulation of value function approximation thatprovides strong a priori guarantees. In particular, this approach provably finds an approximate value function that minimizes the Bellman residual. Solving a bilinear program optimally is NPhard, but this is unavoidable because the Bellman-residual minimization itself is NPhard. We therefore employ and analyze a common approximate algorithm for bilinear programs. The analysis shows that this algorithm offers a convergent generalization of approximate policy iteration.Finally, we demonstrate that the proposed approach can consistently minimize the Bellman residual on a simple benchmark problem.

We introduce the first temporal-difference learning algorithm that is stable with linear function approximation and off-policy training, for any finite Markov decision process, target policy, and exciting behavior policy, and whose complexity scales linearly in the number of parameters. We consider an i.i.d.\ policy-evaluation setting in which the data need not come from on-policy experience. The gradient temporal-difference (GTD) algorithm estimates the expected update vector of the TD(0) algorithm and performs stochastic gradient descent on its L_2 norm. Our analysis proves that its expected update is in the direction of the gradient, assuring convergence under the usual stochastic approximation conditions to the same least-squares solution as found by the LSTD, but without its quadratic computational complexity. GTD is online and incremental, and does not involve multiplying by products of likelihood ratios as in importance-sampling methods.

Relational representation of knowledge makes it possible to perform all the computations and decision making in a uniform relational way by means of special relational compositions called triangle and square products. In this paper some applications in manufacturing related to cost analysis are described. Testing fuzzy relational structures for various relational properties allows us to discover dependencies, hierarchies, similarities, and equivalences of the attributes characterizing technological processes and manufactured artifacts in their relationship to costs and performance. A brief overview of mathematical aspects of BK-relational products is given in Appendix 1 together with further references in the literature.