Faculty Members researchs

أنت هنا

Probabilistic Vector Quantization for Discrete HMM-Based Learning: Type-Written Arabic OCR; a Case Study Waleed Nazeeh Ahmed

Vector quantization (VQ) is a fundamental signal processing operation that attributes a given point in a multidimensional space (i.e. vector) to one of the centroids in a codebook, that in turn is inferred (via some offline codebook-making algorithm like LBG, K-means ... etc.) to optimally represent a population of points (e.g.. features) corresponding to some observable phenomenon [12, 17, 18, 24]. VQ is typically implemented via a "minimum distance" criterion that in turn is an instance of the hard deciding "winner-takes-all" policy.

Our intended project, on the other hand, introduces a novel probabilistic criterion to VQ (ProVQ) that is an instance of a fairer soft-deciding approach. Our probabilistic VQ builds a probability distribution for the belonging of some given point/vector to each centroid in the codebook that is inversely proportional to the distances (i.e. directly proportional to the closeness) between that point and all the codebook’s centroids. The actual runtime arbitration of the given point to a specific centroid is decided via a random election simulator following that probability distribution.

We speculate that our ProVQ - that results in smooth edges separating the different classes – will mitigate the negative effect of over-fitting that degrades the performance of machine learning/classification systems incorporating VQ [11], and may also make these systems more robust with the inevitable noise superimposed to their inputs.

To experimentally attest such a speculation, we will incorporate ProVQ in one of the state-of-the-art discrete HMM-based Arabic type-written OCR systems [2, 3, 12, 32], and hence compare its recognition performance with the...

Processing the Text of the Holy Quran: a Text Mining Study محمد عمر الحوارات

 The Holy Quran is the reference book for more than 1.6 billion of Muslims all around the world Extracting information and knowledge from the Holy Quran is of high benefit for both specialized people in Islamic studies as well as non-specialized people. This paper initiates a series of research studies that aim to serve the Holy Quran and provide helpful and accurate information and knowledge to the all human beings. Also, the planned research studies aim to lay out a framework that will be used by researchers in the field of Arabic natural language processing by providing a ”Golden Dataset” along with useful techniques and information that will advance this field further. The aim of this paper is to find an approach for analyzing Arabic text and then providing statistical information which might be helpful for the people in this research area. In this paper the holly Quran text is preprocessed and then different text mining operations are applied to it to reveal simple facts about the terms of the holy Quran. The results show a variety of characteristics of the Holy Quran such as its most important words, its wordcloud and chapters with high term frequencies. All these results are based on term frequencies that are calculated using both Term Frequency (TF) and Term Frequency-Inverse Document Frequency (TF-IDF) methods.

...
Aldaej, A., Krause, P. (2014). An Enhanced Approach to Semantic Markup of VLEs Content Based on Schema.org. In 4th Int. Workshop on Learning and Education with the Web of Data–13th Int. Semantic Web Conference. Riva del Garda, Italy. عبدالعزيز عبدالله الداعج

Coming soon..

...
Aldaej, A., Krause, P. (2014). SocialLearn Demo. Making it Matter: Supporting education in the developing world through open and linked data, LINKEDUP Project, London, UK, 16th May, 2014. عبدالعزيز عبدالله الداعج

Coming soon..

...
Aldaej, A., Krause, P. (2013). E-Learning evolution: Semantic Web and social networking. 3rd Postgraduate Research Conference, University of Surrey, Guildford, UK, 29th-30th January 2013. عبدالعزيز عبدالله الداعج

Coming soon..

...
Aldaej, A., Krause, P. & Strømmen-Bakhtiar, A. (2012). E-Learning evolution: Next Steps Semantic web and foundations of E-learning. In Proceedings of World Conference on E-Learning in Corporate, Government, Healthcare, and Higher Education 2012, Montreal, عبدالعزيز عبدالله الداعج

Coming soon..

...
Development of a Spatial Decision Support System for Farms Management الصادق عبدالله الجاك فضل الله

Mechanized agricultural operations are needed to increase productivity, efficiency and competitiveness of farms. Likewise, developing and promoting appropriate new technologies are needed in the decision making and management process. Farm decision-makers are faced with not only the need to manage traditional facilities, but, also, complex environmental management requirements. This research project aims to develop a conceptual model for a farm Spatial Decision Support System. The model uses Decision Support System (DSS) and Geographical Information System (GIS) to handle farm spatial and attribute information. The real power of GIS derives from the integration of database and graphics capabilities with engineering design, analytical, and cost models. Importance of research stems from its association with the community of Al-Kharj area. It is expected that the proposed system will be an effective tool for the advance of decision-making process on farms,  and will  contain many types of  descriptive, quantitative and spatial analysis models, that can be easily and friendly used.

...
End-to-End Route Reliability in Pervasive Multi-Channel Multi-Radio (PMCMR) Scheme for MANETS Usman Tariq

Multi-channel multi radio wireless communications technologies have changed the trend of traditional routing based Quality of Service (QoS) in wireless-mesh networks. This avalanche of growth in multimedia communication is trigged by Multi-Channel Multi-Radio Wireless Mesh Networks (MCMR-WMNs). 

...
Pattern recall in networks of chaotic neurons محمد عمر الحوارات

This research investigates the potential utility of chaotic dynamics in neural information processing. A novel chaotic spiking neural network model is presented which is composed of non-linear dynamic state (NDS) neurons. The activity of each NDS neuron is driven by a set of non-linear equations coupled with a threshold based spike output mechanism. If time-delayed self-connections are enabled then the network stabilises to a periodic pattern of activation. Previous publications of this work have demonstrated that the chaotic dynamics which drive the network activity ensure that an extremely large number of such periodic patterns can be generated by this network. This paper presents a major extension to this model which enables the network to recall a pattern of activity from a selection of previously stabilised patterns.

...
The Nonlinear Dynamic State neuron محمد عمر الحوارات

This research is concerned with using nonlinear dynamics to greatly enhance the 
range of possible behaviours of artificial neurons. A novel neuron model is presented which 
has a dynamic internal state defined by a set of nonlinear equations, together with a 
threshold driven spike output mechanism. With the aid of spike feedback control the model is 
able to stabilise one of a large number of Unstable Periodic Orbits in its internal dynamics. 
These orbits correspond to dynamic states of the neuron each of which generates a ...

...
QR Code for https://cces.psau.edu.sa/ar/sources/research/4