Back-propagation learning in expert networks

Expert networks are event-driven, acyclic networks of neural objects derived from expert systems. The neural objects process information through a nonlinear combining function that is different from, and more complex than, typical neural network node processors. The authors develop back-propagation learning for acyclic, event-driven networks in general and derive a specific algorithm for learning in EMYCIN-derived expert networks. The algorithm combines back-propagation learning with other features of expert networks, including calculation of gradients of the nonlinear combining functions and the hypercube nature of the knowledge space. It offers automation of the knowledge acquisition task for certainty factors, often the most difficult part of knowledge extraction. Results of testing the learning algorithm with a medium-scale (97-node) expert network are presented.     

A comparative analysis of methods for expert systems

Given the current widespread interest in expert systems, it is important to examine the relative advantages and disadvantages of the various methods used to build them. In this paper we compare three important approaches to building decision aids implemented as expert systems: Bayesian classification, rule-based deduction, and frame-based abduction. Our critical analysis is based on a survey of previous studies comparing different methods used to build expert systems as well as our own collective experience over the last five years. The relative strengths and weaknesses of the different approaches are analysed, and situations in which each method is easy or difficult to use are identified.     

Robust learning algorithm for multiplicative neuron model artificial neural networks

The two most commonly used types of artificial neural networks (ANNs) are the multilayer feed-forward and multiplicative neuron model ANNs. In the literature, although there is a robust learning algorithm for the former, there is no such algorithm for the latter. Because of its multiplicative structure, the performance of multiplicative neuron model ANNs is affected negatively when the dataset has outliers. On this issue, a robust learning algorithm for the multiplicative neuron model ANNs is proposed that uses Huber's loss function as fitness function. The training of the multiplicative neuron model is performed using particle swarm optimization. One principle advantage of this algorithm is that the parameter of the scale estimator, which is an important factor affecting the value of Huber's loss function, is also estimated with the proposed algorithm. To evaluate the performance of the proposed method, it is applied to two well-known real world time series datasets, and also a simulation study is performed. The algorithm has superior performance both when it is applied to real world time series datasets and the simulation study when compared with other ANNs reported in the literature. Another of its advantages is that, for datasets with outliers, the results are very close to the results obtained from the original datasets. In other words, we demonstrate that the algorithm is unaffected by outliers and has a robust structure.     

Model-free reinforcement learning from expert demonstrations: a survey

Reinforcement learning from expert demonstrations (RLED) is the intersection of imitation learning with reinforcement learning that seeks to take advantage of these two learning approaches. RLED uses demonstration trajectories to improve sample efficiency in high-dimensional spaces. RLED is a new promising approach to behavioral learning through demonstrations from an expert teacher. RLED considers two possible knowledge sources to guide the reinforcement learning process: prior knowledge and online knowledge. This survey focuses on novel methods for model-free reinforcement learning guided through demonstrations, commonly but not necessarily provided by humans. The methods are analyzed and classified according to the impact of the demonstrations. Challenges, applications, and promising approaches to improve the discussed methods are also discussed.

     

Combining expert neural networks using reinforcement feedback for learning primitive grasping behavior

This paper present an architecture for combining a mixture of experts. The architecture has two unique features: 1) it assumes no prior knowledge of the size or structure of the mixture and allows the number of experts to dynamically expand during training, and 2) reinforcement feedback is used to guide the combining/expansion operation. The architecture is particularly suitable for applications when there is a need to approximate a many-to-many mapping. An example of such a problem is the task of training a robot to grasp arbitrarily shaped objects. This task requires the approximation of a many-to-many mapping, since various configurations can be used to grasp an object, and several objects can share the same grasping configuration. Experiments in a simulated environment using a 28-object database showed how the algorithm dynamically combined and expanded a mixture of neural networks to achieve the learning task. The paper also presents a comparison with two other nonlearning approaches.     

Robust passivity analysis of a class of discrete-time stochastic neural networks

This paper addresses the passivity problem of a class of discrete-time stochastic neural networks with time-varying delays and norm-bounded parameter uncertainties. New delay-dependent passivity conditions are obtained by using a novel Lyapunov functional together with the linear matrix inequality approach. Numerical examples show the effectiveness of the proposed method.     

Alternative logics for approximate reasoning in expert systems: a comparative study

In this paper we report the results of an empirical study to compare eleven alternative logics for approximate reasoning in expert systems. The several “compositional inference” axiom systems (described below) were used in an expert knowledge-based system. The quality of the system outputs—fuzzy linguistic phrases—were compared in terms of correctness and precision (non-vagueness).In the first section of the paper we discuss fuzzy expert systems. The second section provides a brief review of logic systems and their relation to approximate reasoning. Section three contains the experimental design, and section four supplies the results of the experiment. Finally, a summary is given.     

Modeling ecological expert: comparative analysis of human skill development with fuzzy controller

 The main difficulties of the skill transfer from human to machine exist in the fact that human skills are based upon two types of knowledge: one is the fundamental content knowledge needed to perform complex tasks, and the other one is knowledge of the process, by which those tasks should be executed. Distinguishing between those two types of knowledge, we present a comparative analysis between a fuzzy controller and a human expert. Regarding a human proficient expert as an ecological expert after Kirlik, we demonstrate that skillful control lies not only inside of the skill-performer's brain, but in the actor-environment system. In order to investigate into the relations between the human judgments and the environmental information, we adopt Brunswik's Lens Model to quantify both types of knowledge from the performance data. By analyzing how the ways of an operator's interacting with the task environment change and how the cues in the environment utilized by him/her alter, we formalize his/her control-skill improving process. We investigate these in comparison with the conventional fuzzy controller. We conclude in the aspects in which the human expert is superior to the fuzzy controller.     

Robust hebbian learning and noisy principal component analysis

The classical analysis of a stochastic signal into principal components compresses the signal using an optimal selection of linear features. Noisy Principal Component Analysis (NPCA) is an extension of PCA under the assumption that the extracted features are unreliable, and the unreliability is modeled by additive noise. The applications of this assumption appear for instance, in communications problems with noisy channels. The level of noise in the NPCA features affects the reconstruction error in a way resembling the water-filling analogy in information theory. Robust neural network models for Noisy PCA can be defined with respect to certain synaptic weight constraints. In this paper we present the NPCA theory related to a particularly simple and tractable constraint which allows us to evaluate the robustness of old PCA Hebbian learning rules. It turns out that those algorithms are not optimally robust in the sense that they produce a zero solution when the noise power level reaches half the limit set by NPCA. In fact, they are not NPCA-optimal for any other noise levels except zero. Finally, we propose new NPCA-optimal robust Hebbian learning algorithms for multiple adaptive noisy principal component extraction.     

Misinformation-oriented expert finding in social networks

Due to the distributed and decentralized nature of social media, respective content that contains misinformation is usually propagated without any type of moderation, which may mislead the public and have a profound real-world impact. In addition, it is quite challenging to distinguish misinformation with high precision, since the content is often short and lacks of semantics. A promising solution is to utilize the crowdsourcing wisdom that pushes the suspected misinformation to relevant users based on the expertise and collects the assessments to judge the credibility. Even though a lot of expert finding models have been employed, however, these methods cannot effectively deal with the misinformation-oriented expert matching tasks since the data collected from social network is different form traditional text collection. To this end, we focus on how to obtain an appropriate matching between the suspect misinformation and corresponding experts, and propose a multi-topic expert finding method, called LTM (List based Topic Model), to sufficiently utilize crowdsourcing wisdom. Moreover, we optimize the query results with the help of supervised information that extracted from Twitter Lists. Finally, we demonstrate the effectiveness of our work with experiments on real-world data and verify the superiority of our proposed model in accuracy.

     

A comparative study of the scalability of a sensitivity-based learning algorithm for artificial neural networks

Until recently, the most common criterion in machine learning for evaluating the performance of algorithms was accuracy. However, the unrestrainable growth of the volume of data in recent years in fields such as bioinformatics, intrusion detection or engineering, has raised new challenges in machine learning not simply regarding accuracy but also scalability. In this research, we are concerned with the scalability of one of the most well-known paradigms in machine learning, artificial neural networks (ANNs), particularly with the training algorithm Sensitivity-Based Linear Learning Method (SBLLM). SBLLM is a learning method for two-layer feedforward ANNs based on sensitivity analysis, that calculates the weights by solving a linear system of equations. The results show that the training algorithm SBLLM performs better in terms of scalability than five of the most popular and efficient training algorithms for ANNs.     

Robust reinforcement learning

This letter proposes a new reinforcement learning (RL) paradigm that explicitly takes into account input disturbance as well as modeling errors. The use of environmental models in RL is quite popular for both offline learning using simulations and for online action planning. However, the difference between the model and the real environment can lead to unpredictable, and often unwanted, results. Based on the theory of H(infinity) control, we consider a differential game in which a "disturbing" agent tries to make the worst possible disturbance while a "control" agent tries to make the best control input. The problem is formulated as finding a min-max solution of a value function that takes into account the amount of the reward and the norm of the disturbance. We derive online learning algorithms for estimating the value function and for calculating the worst disturbance and the best control in reference to the value function. We tested the paradigm, which we call robust reinforcement learning (RRL), on the control task of an inverted pendulum. In the linear domain, the policy and the value function learned by online algorithms coincided with those derived analytically by the linear H(infinity) control theory. For a fully nonlinear swing-up task, RRL achieved robust performance with changes in the pendulum weight and friction, while a standard reinforcement learning algorithm could not deal with these changes. We also applied RRL to the cart-pole swing-up task, and a robust swing-up policy was acquired.     

Unsupervised learning in second-order neural networks for motion analysis

This paper demonstrates how unsupervised learning based on Hebb-like mechanisms is sufficient for training second-order neural networks to perform different types of motion analysis. The paper studies the convergence properties of the network in several conditions, including different levels of noise and motion coherence and different network configurations. We demonstrate the effectiveness of a novel variability dependent learning mechanism, which allows the network to learn under conditions of large feature similarity thresholds, which is crucial for noise robustness. The paper demonstrates the particular relevance of second-order neural networks and therefore correlation based approaches as contributing mechanisms for directional selectivity in the retina.     

Robust recursive least squares learning algorithm for principalcomponent analysis

A learning algorithm for the principal component analysis (PCA) is developed based on the least-square minimization. The dual learning rate parameters are adjusted adaptively to make the proposed algorithm capable of fast convergence and high accuracy for extracting all principal components. The proposed algorithm is robust to the error accumulation existing in the sequential PCA algorithm. We show that all information needed for PCA can he completely represented by the unnormalized weight vector which is updated based only on the corresponding neuron input-output product. The updating of the normalized weight vector can be referred to as a leaky Hebb's rule. The convergence of the proposed algorithm is briefly analyzed. We also establish the relation between Oja's rule and the least squares learning rule. Finally, the simulation results are given to illustrate the effectiveness of this algorithm for PCA and tracking time-varying directions-of-arrival.     

Credit rating analysis with support vector machines and neural networks: a market comparative study

Corporate credit rating analysis has attracted lots of research interests in the literature. Recent studies have shown that Artificial Intelligence (AI) methods achieved better performance than traditional statistical methods. This article introduces a relatively new machine learning technique, support vector machines (SVM), to the problem in attempt to provide a model with better explanatory power. We used backpropagation neural network (BNN) as a benchmark and obtained prediction accuracy around 80% for both BNN and SVM methods for the United States and Taiwan markets. However, only slight improvement of SVM was observed. Another direction of the research is to improve the interpretability of the AI-based models. We applied recent research results in neural network model interpretation and obtained relative importance of the input financial variables from the neural network models. Based on these results, we conducted a market comparative analysis on the differences of determining factors in the United States and Taiwan markets.     

A comparative survey of artificial intelligence applications in finance: artificial neural networks, expert system and hybrid intelligent systems

Nowadays, many current real financial applications have nonlinear and uncertain behaviors which change across the time. Therefore, the need to solve highly nonlinear, time variant problems has been growing rapidly. These problems along with other problems of traditional models caused growing interest in artificial intelligent techniques. In this paper, comparative research review of three famous artificial intelligence techniques, i.e., artificial neural networks, expert systems and hybrid intelligence systems, in financial market has been done. A financial market also has been categorized on three domains: credit evaluation, portfolio management and financial prediction and planning. For each technique, most famous and especially recent researches have been discussed in comparative aspect. Results show that accuracy of these artificial intelligent methods is superior to that of traditional statistical methods in dealing with financial problems, especially regarding nonlinear patterns. However, this outperformance is not absolute.     

Robust reinforcement learning control using integral quadratic constraints for recurrent neural networks.

The applicability of machine learning techniques for feedback control systems is limited by a lack of stability guarantees. Robust control theory offers a framework for analyzing the stability of feedback control loops, but for the integral quadratic constraint (IQC) framework used here, all components are required to be represented as linear, time-invariant systems plus uncertainties with, for IQCs used here, bounded gain. In this paper, the stability of a control loop including a recurrent neural network (NN) is analyzed by replacing the nonlinear and time-varying components of the NN with IQCs on their gain. As a result, a range of the NN's weights is found within which stability is guaranteed. An algorithm is demonstrated for training the recurrent NN using reinforcement learning and guaranteeing stability while learning.     

Anti-Hebbian learning in topologically constrained linear networks:a tutorial

Using standard results from the adaptive signal processing literature, we review the learning behavior of various constrained linear neural networks made up of anti-Hebbian synapses, where learning is driven by the criterion of minimizing the node information energy. We point out how simple learning rules of Hebbian type can provide fast self-organization, under rather wide connectivity constraints. We verify the results of the theory in a set of simulations.     

Datacab: a geographical-information-system-based expert system for the design of cable networks

Abstract: Telecommunication networks have evolved over time as a result of technological advances, and network topologies and equipment have become increasingly complex. Expert systems are being successfully applied to the management of telecommunication networks. However, applying expert systems to network design is another especially beneficial yet still not very common approach. In this paper we propose a rule-based expert system called Datacab. Datacab was developed at Enditel Endesa in collaboration with the Electronic Technology Department of the University of Seville, for the automatic design of hybrid fibre coax (HFC) cable networks. Using data from a geographical information system as input, it automatically generates viable HFC network designs.