Abdul Rahim Ahmad

Online handwriting recognition using support vector machine

Discrete hidden Markov model (HMM) and hybrid of neural network (NN) and HMM are popular methods in handwritten word recognition system. The hybrid system gives better recognition result due to better discrimination capability of the NN [Y. Bengio et al., 1995]. Support vector machine (SVM) is an alternative to NN. In speech recognition (SR), SVM has been successfully used in the context of a hybrid SVM/HMM system. It gives a better recognition result compared to the system based on hybrid NN/HMM [A. Ganapathiraju, January 2002]. This paper describes the work in developing a hybrid SVM/HMM OHR system. Some preliminary experimental results of using SVM with RBF kernel on IRONOFF, UNIPEN and IRONOFF- UNIPEN character database are provided.

A Survey of Arabic language Support in Semantic web

Information availability is a key factor in the acquisition of knowledge. Access to information either in the general area or even in more specific ones like sciences, languages, and religion become wider since the use of semantics in World Wide Web. Semantic Web technologies assist in the acquiring of information by creating processes that link information to another. However, the technology supports mostly languages using Latin family scripts. Arabic is still not well supported. This paper, reports on the survey of the support for Arabic in some of the existing Semantic Web technologies, and give future scenario in applying Semantic Web for Arabic applications. Finally, multilingual support for these new technologies is also discussed.

An Arabic language framework for semantic web

This research study was basically aimed at addressing the current lack of Arabic support within majority of semantic web tools technologies and applications. It will also address the on-going problem of the lack of resources for Arabic research or semantic web technologies that support the Arabic text. While there may have already been various researches in the past that have been conducted in order to develop ontology-based and optimized semantic framework with the English content, there were very few researches that were developed for the Arabic language. This paper, reports on the survey of the support for the Arabic language in some of the existing semantic web technologies and gives future scenario in applying Semantic Web for Arabic applications. Finally, we propose a new framework intended to add a semantic web layer to the current web based application in order to improve the process of linking, integrating, and searching for different types of applications. The proposed framework will serve Arabic language processing, although it is designed to be adapted to any other languages.

Lexicon-Based Word Recognition Using Support Vector Machine and Hidden Markov Model

Hybrid of Neural Network (NN) and Hidden Markov Model (HMM) has been popular in word recognition, taking advantage of NN discriminative property and HMM representational capability. However, NN does not guarantee good generalization due to Empirical Risk minimization (ERM) principle that it uses. In our work, we focus on using the support vector machine (SVM) for character recognition. SVMs use of structural risk minimization (SRM) principle has allowed simultaneous optimization of representational and discriminative capability of the character recognizer. We first evaluated SVM in isolated character recognition environment using IRONOFF and UNIPEN character databases. We then demonstrate the practical issues in using SVM within a hybrid setting with HMM for word recognition. We tested the hybrid system on the IRONOFF word database and obtained commendable results.

MS-TDNN with global discriminant trainings

This article analyses the behavior of various hybrid architectures based on a multi-state neuro-Markovian scheme (MS-TDNN HMM) applied to online handwriting word recognition systems. We have considered different cost functions, including maximal mutual information criteria with discriminant training and maximum likelihood estimation, to train the systems globally at the word level and also we varied the number of states from one up to three to model the basic hidden Markov models at the letter level. We report experimental results for non-constrained, writer independent, word recognition obtained on the IRONOFF database.

Trust Formation Based on Subjective Logic and PGP Web-of-Trust for Information Sharing in Mobile Ad Hoc Networks

Trust play important roles in Mobile ad hoc networks since nodes may randomly joined and leave the network at their own pace. In catastrophe like massive accident or earthquake, groups of rescue personnel form a temporary network structure at the rescue site using their portable devices for coordinating the rescue relief. These groups of user from different agency need to use the temporary network for communication and also for sharing the information. Trust plays important roles in these groups since sharing of information must be among trusted and authorized nodes only. In this paper we form trust model adopting the PGP web-of-trust concept with subjective logic. The PGP web-of-trust is suitable method to adopt in mobile ad hoc networks since nodes are communicating based on peer to peer, hence peer recommendation can be use to create trust between peers. The use of subjective logic operators for calculating trust in the proposed trust model is appropriate since there is an element such as uncertainty as part of calculation. Uncertainty is important in this dynamic network since we do not have full information when decision is make. The trust value later is mapped with the access control policy to determine user access privileges.

Group based access control scheme (GBAC): Keeping information sharing secure in mobile Ad- hoc environment

Mobile ad-hoc network is a network that can dynamically setup on the fly by mobile nodes. Due to its unique characteristics, it is becoming an attractive choice for commercial and also military application and among the used is to support information sharing among mobile nodes. However due to its borderless infrastructure, this network is highly susceptible to adversaries nodes, that illegally able to access information shared in the network. Hence to overcome this problem we proposed a Group Based Access Control scheme, a step by step process or protocol that able to make an access to share information in the network secure. We use emergency rescue mission scenario to illustrate the use of the scheme. The protocol created meets the security properties such as data confidentiality, integrity and also non-repudiation.

Parallel execution of distributed SVM using MPI (CoDLib)

Support Vector Machine (SVM) is an efficient data mining approach for data classification. However, SVM algorithm requires very large memory requirement and computational time to deal with very large dataset. To reduce the computational time during the process of training the SVM, a combination of distributed and parallel computing method, CoDLib have been proposed. Instead of using a single machine for parallel computing, multiple machines in a cluster are used. Message Passing Interface (MPI) is used in the communication between machines in the cluster. The original dataset is split and distributed to the respective machines. Experiments results shows a great speed up on the training of the MNIST dataset where training time has been significantly reduced compared with standard LIBSVM without affecting the quality of the SVM.

Group based access control scheme: proof of method for secure access control architecture in mobile ad-hoc networks

In disaster area, where the infrastructures is partially or fully destroyed, a form of communication to allow information been shared among rescue team is needed. Since Mobile ad-hoc network is easy to setup and required less infrastructure therefore it is a suitable candidate to work in disaster area. Despite of its uniqueness, this network is highly vulnerable to malicious node and also to threats. In rescue mission scenario, information needs to be shared among trusted and legal nodes only hence a mechanism to restrict an access to information in this network is extremely important. In this paper, we outline the access control requirement for this network and proposed the secure access control architecture based on the requirements. Based on the proposed architecture, we derive a Group Based access control scheme, to show how an access to information in mobile ad-hoc environment at emergency rescue mission is working.

Kernel methods and support vector machines for handwriting recognition

This paper presents a review of kernel methods in machine learning. The support vector machine (SVM) as one of the methods in machine learning to make use of kernels is first discussed with the intention of applying it to handwriting recognition. SVM works by mapping training data for a classification task into a higher dimensional feature space using the kernel function and then finding a maximal margin hyperplane, which separates the mapped data. Finding the solution hyperplane involves using quadratic programming which is computationally intensive. Algorithms for practical implementation such as sequential minimization optimization (SMO) and its improvements are discussed. A few simpler methods similar to SVM but requiring simpler computation are also mentioned for comparison. Usage of SVM for handwriting recognition is then proposed.