Towards Incremental Fuzzy Classifiers

Fuzzy classification systems (FCS) are traditionally built from observations (data points) in an off-line one shot-experiment. Once the learning phase is exhausted, the classifier is no more capable to learn further knowledge from new observations nor is it able to update itself in the future. This paper investigates the problem of incremental learning in the context of FCS. It shows how, in contrast to off-line or batch learning, incremental learning infers knowledge in the form of fuzzy rules from data that evolves over time. To accommodate incremental learning, appropriate mechanisms are applied in all steps of the FCS construction: (1) Incremental supervised clustering to generate granules in a progressive manner, (2) Systematic and automatic update of fuzzy partitions, (3) Incremental feature selection using an incremental version of Fisher’s interclass separability criterion. The effect of incrementality on various aspects is demonstrated via a numerical evaluation.     

Decentralized Networked Control System Design Using Takagi-Sugeno (TS) Fuzzy Approach

This paper proposes a new method for control of continuous large-scale systems where the measures and control functions are distributed on calculating members which can be shared with other applications and connected to digital network communications.At first, the nonlinear large-scale system is described by a Takagi-Sugeno(TS) fuzzy model. After that, by using a fuzzy LyapunovKrasovskii functional, sufficient conditions of asymptotic stability of the behavior of the decentralized networked control system(DNCS),are developed in terms of linear matrix inequalities(LMIs). Finally, to illustrate the proposed approach, a numerical example and simulation results are presented.     

Learning with partly labeled data

Learning with partly labeled data aims at combining labeled and unlabeled data in order to boost the accuracy of a classifier. This paper outlines the two main classes of learning methods to deal with partly labeled data: pre-labeling-based learning and semi-supervised learning. Concretely, we introduce and discuss three methods from each class. The first three ones are two-stage methods consisting of selecting the data to be labeled and then training the classifier using the pre-labeled and the originally labeled data. The last three ones show how labeled and unlabeled data can be combined in a symbiotic way during training. The empirical evaluation of these methods shows: (1) pre-labeling methods tend be better than semi-supervised learning methods, (2) both labeled and unlabeled have positive effect on the classification accuracy of each of the proposed methods, (3) the combination of all the methods improve the accuracy, and (4) the proposed methods compare very well with the state-of-art methods.     

Scaling up stochastic gradient descent for non-convex optimisation

Stochastic gradient descent (SGD) is a widely adopted iterative method for optimizing differentiable objective functions. In this paper, we propose and discuss a novel approach to scale up SGD in applications involving non-convex functions and large datasets. We address the bottleneck problem arising when using both shared and distributed memory. Typically, the former is bounded by limited computation resources and bandwidth whereas the latter suffers from communication overheads. We propose a unified distributed and parallel implementation of SGD (named DPSGD) that relies on both asynchronous distribution and lock-free parallelism. By combining two strategies into a unified framework, DPSGD is able to strike a better trade-off between local computation and communication. The convergence properties of DPSGD are studied for non-convex problems such as those arising in statistical modelling and machine learning. Our theoretical analysis shows that DPSGD leads to speed-up with respect to the number of cores and number of workers while guaranteeing an asymptotic convergence rate of \(O(1/\sqrt{T})\) given that the number of cores is bounded by \(T^{1/4}\) and the number of workers is bounded by \(T^{1/2}\) where T is the number of iterations. The potential gains that can be achieved by DPSGD are demonstrated empirically on a stochastic variational inference problem (Latent Dirichlet Allocation) and on a deep reinforcement learning (DRL) problem (advantage actor critic - A2C) resulting in two algorithms: DPSVI and HSA2C. Empirical results validate our theoretical findings. Comparative studies are conducted to show the performance of the proposed DPSGD against the state-of-the-art DRL algorithms.

     

A new route for the preparation of cyano‐containing poly(N‐isopropylacrylamide) microgel latex for specific immobilization of antibodies

A prepolymerization process was used to prepare functionalized poly(N‐isopropylacrylamide) latexes with surface cyano groups. The functionalized latexes prepared were characterized by FTIR, ¹H NMR, scanning electron microscopy and quasi‐elastic light scattering. In addition, the polymerization conversion and the water‐soluble polymer amounts were quantified. The polymerization conversions were found to be above 80% with 5–14 wt% of water soluble polymer formation. The immobilization of antibody (immunoglobulin) onto such cyano‐containing thermally sensitive particles, suggests the feasibility of specific dipole–dipole interactions between the cyano and hydroxyl functional groups from particle and antibody, respectively. Copyright © 2004 Society of Chemical Industry     

Sulfated Zirconia-Based Strong Solid-Acid Catalysts: Recent Progress

This review article deals with recent progress in the preparation of sulfated zirconia (SZ)-bassed, strong solid-acid catalysts, the characterization of their physicochemical properties and the evaluation of their catalytic performance in various promising applications. Strong emphasis was put on discussion of controversial issues such as the strength of acid sites, the nature of active sites, the reaction mechanism, and the role and state of supported platinum. An important part of this work was devoted to recent catalytic applications.     

The Role of Crosslink Density in Influencing the Electrical Resistance Behaviour of Polymeric Coatings

Paint films used to protect metalic surfaces are commonly polymeric in nature. The extent of protection offered by film depends on many factors including the characteristic electrical resistance behaviou and its effect on impeding local electrochemical processes. In the present work a range of polymeric coatings have been produced with systematically varied crosslinked density using an ultra-violet light curing technique. Their electrical resistance behaviour in an environment of varying concentrations of KC1 electrolyte has been examined. It is demonstrated that there are signs of the beginnings of a mechanism changeover from “D-type” to “I-type” behaviour at higher levels of crosslink density thus giving some tenuous support to previously unproven hypotheses in this area.     

Closed loop torque SVM-DTC based on robust super twisting speed controller for induction motor drive with efficiency optimization

This paper presents a space vector modulation (SVM) based Direct Torque Control strategy (DTC) for induction motor (IM) in order to overcome the drawbacks of the classical DTC. SVM can reduce the high torque and flux ripples by preserving a fixed switching frequency. This technique is known by the closed loop torque SVM-DTC. Moreover, the control scheme performance is improved by inserting a second order sliding mode super twisting controller in the outer loop for speed regulation. This nonlinear technique ensures a good dynamic and high robustness against external disturbance. Furthermore, the IM energy optimization is treated in the second objective of this paper. A proposed model based loss minimization strategy is presented for efficiency optimization. This strategy chooses an optimal flux magnitude for each applied load torque. The proposed optimized SVM-DTC algorithm will be investigated by simulation and real time implementation using Matlab/Simulink with real time interface based on dSpace 1104 signal card.