Department of Computer Science

The application of deep learning in robotics leads to very specific problems and research questions that are typically not addressed by the computer vision and machine learning communities. In this paper we discuss a number of robotics-specific learning, reasoning, and embodiment challenges for deep learning. We explain the need for better evaluation metrics, highlight the importance and unique challenges for deep robotic learning in simulation, and explore the spectrum between purely data-driven and model-driven approaches. We hope this paper provides a motivating overview of important research directions to overcome the current limitations, and helps to fulfill the promising potentials of deep learning in robotics.

Feature selection is of paramount concern in document classification process which improves the efficiency and accuracy of text classifier. Vector Space Model is used to represent the "Bag of Word" BOW of the documents with term weighting phenomena. Documents representing through this model has some limitations that is, ignoring term dependencies, structure and ordering of the terms in documents. To overcome this problem semantic base feature vector is proposed. That is used to extracts the concept of term, co-occurring and associated terms using ontology. The proposed method is applied on small documents dataset, which shows that this method outperforms then term frequency/ inverse document frequency (TF-IDF) with BOW feature selection method for text classification.

Opinion target identification is an important task of opinion mining problem. Several approaches have been employed for this task that can be broadly divided into two major categories: supervised and unsupervised. The supervised approaches require training data which need manual work and are mostly domain dependent. Unsupervised technique is most popularly used due to its two main advantages: domain independent and no need of training data. This paper presents optimization of unsupervised likelihood ratio test (LRT) technique for opinion targets identification from unstructured reviews. The LRT approach basically depends on frequently observed features and hence do not predict infrequent features (IFF). This paper exploits WordNet lexical resource to identify IFF based on semantic relatedness to the frequent features in review documents which improves the overall accuracy of features extraction algorithms. Empirical results show the validity of the proposed technique.

Feature selection is of paramount concern in document classification process which improves the efficiency and accuracy of text classifier. Vector Space Model is used to represent the "Bag of Word" BOW of the documents with term weighting phenomena. Documents representing through this model has some limitations that is, ignoring term dependencies, structure and ordering of the terms in documents. To overcome this problem semantic base feature vector is proposed. That is used to extracts the concept of term, co-occurring and associated terms using ontology. The proposed method is applied on small documents dataset, which shows that this method outperforms then term frequency/ inverse document frequency (TF-IDF) with BOW feature selection method for text classification.

In the computational linguistics the extraction of actual sense of words from text has a long history in the field. Due to its importance in the field of sentiment analysis it is considered the most important one. During sentiment analysis more challenging problems are faced due to the ambiguous senses of words. In this work we propose a new method of word sense disambiguation (WSD) using matrix map of the semantic scores extracted from SentiWordNet of WordNet glosses terms. The correct sense of the target word is extracted and determined for which the similarity between WordNet gloss and context matrix is greatest. Our empirical results have shown that the proposed method improves the result of sentence level sentiment classification as evaluated on different domain datasets. From the result it is clear that the propose method achieves an accuracy of 90.71% at sentence level sentiment classification of online reviews.

The exponential increase in the explosion of Web-based user generated reviews has resulted in the emergence of Opinion Mining (OM) applications for analyzing the users' opinions toward products, services, and policies. The polarity lexicons often play a pivotal role in the OM, indicating the positivity and negativity of a term along with the numeric score. However, the commonly available domain independent lexicons are not an optimal choice for all of the domains within the OM applications. The aforementioned is due to the fact that the polarity of a term changes from one domain to other and such lexicons do not contain the correct polarity of a term for every domain. In this work, we focus on the problem of adapting a domain dependent polarity lexicon from set of labeled user reviews and domain independent lexicon to propose a unified learning framework based on the information theory concepts that can assign the terms with correct polarity (+ive,-ive) scores. The benchmarking on three datasets (car, hotel, and drug reviews) shows that our approach improves the performance of the polarity classification by achieving higher accuracy. Moreover, using the derived domain dependent lexicon changed the polarity of terms, and the experimental results show that our approach is more effective than the base line methods.

With the increasing availability of electronic documents and the rapid growth of the World Wide Web, the task of automatic categorization of documents becomes the key method for organizing the information, knowledge and trend detection. With the growing availability of online resources, and popularity of fast and rich resources on web, classification of e-documents, news, personal blogs, and extraction of knowledge and trend from the documents has become an interesting area for research, as the World Wide Web is the fastest media for news and events collection from world. So the growing phenomenon of the textual data needs text mining, machine learning and natural language processing techniques and methodologies to organize and extract pattern and knowledge from the documents. This overview focused on the existing literature and explored the main techniques and methods for automatic documents classification i.e. documents representation, classifier construction and knowledge extract...

Recently, Cognitive-based Sentiment Analysis with emphasis on automatic detection of user behaviour, such as personality traits, based on online social media text has gained a lot of attention. However, most of the existing works are based on conventional techniques, which are not sufficient to get promising results. In this research work, we propose a hybrid Deep Learning-based model, namely Convolutional Neural Network concatenated with Long Short-Term Memory, to show the effectiveness of the proposed model for 8 important personality traits (Introversion-Extroversion, Intuition-Sensing, Thinking-Feeling, Judging-Perceiving). We implemented our experimental evaluations on the benchmark dataset to accomplish the personality trait classification task. Evaluations of the datasets have shown better results, which demonstrates that the proposed model can effectively classify the user's personality traits as compared to the state-of-the-art techniques. Finally, we evaluate the effectiveness of our approach through statistical analysis. With the knowledge obtained from this research, organizations are capable of making their decisions regarding the recruitment of personals in an efficient way. Moreover, they can implement the information obtained from this research as best practices for the selection, management, and optimization of their policies, services, and products.

With the rapid increase in social networks and blogs, the social media services are increasingly being used by online communities to share their views and experiences about a particular product, policy and event. Due to economic importance of these reviews, there is growing trend of writing user reviews to promote a product. Nowadays, users prefer online blogs and review sites to purchase products. Therefore, user reviews are considered as an important source of information in Sentiment Analysis (SA) applications for decision making. In this work, we exploit the wealth of user reviews, available through the online forums, to analyze the semantic orientation of words by categorizing them into +ive and -ive classes to identify and classify emoticons, modifiers, general-purpose and domain-specific words expressed in the public's feedback about the products. However, the un-supervised learning approach employed in previous studies is becoming less efficient due to data sparseness, l...

Automatic extraction of features from unstructured text is one of the challenging problems of Opinion Mining. The trend of getting products and services reputation from online resources such as web blogs and customer feedback is increasing day by day. Therefore efficient system is required to automatically extract products features and the opinion of consumers about all aspects of the products. In this study our focus is on extraction of product features from customer reviews. We have proposed a concordance based technique for automatic extraction of features of product from customer reviews. In our proposed technique we extract patterns of lexical terms using concordance for candidate features extraction and identify features by grouping. The proposed grouping algorithm is used to remove irrelevant features. We conducted experiments on different products reviews and compared our results with existing methods. From empirical results we proved the validity of the proposed method.

Opinion targets identification is an important task of the opinion mining problem. Several approaches have been employed for this task, which can be broadly divided into two major categories: supervised and unsupervised. The supervised approaches require training data, which need manual work and are mostly domain dependent. The unsupervised technique is most popularly used due to its two main advantages: domain independent and no need for training data. This paper presents a review of the state of the art unsupervised approaches for opinion target identification due to its potential applications in opinion mining from user discourse. This study compares the existing approaches that might be helpful in the future research work of opinion mining and features extraction.

First of all, I would like to thank the Almighty Allah, for His divine guidance and providence. His support, blessings, goodness, and kindness were always with me. He blessed me with motivation, passion, and hard work. It was his blessings which made me able to plan, visualize, and execute my dreams into the reality. I would like to dedicate this achievement to Him and to my mother and wife, whose immeasurable sacrifices have led to what I am today. Every PhD student dreams for a visionary supervisor. I am extremely thankful to Allah, for connecting me up with the best possible supervisor Associate Professor Dr. Baharum Baharudin. Prof. Dr. Baharum"s guidance, inspirations and motivations enabled me to bring out the best of myself. I managed to learn a lot from him. The list would become quite long if I mention each and everything, to be short, it is Dr. Baharum"s dedication and perseverance that made me learn scientific writing, guided me to the art of critical thinking, taught me to work smart, gave me freedom to explore different ideas, linked me up with the best researchers, and provided me the best possible working environment. Without his continuous support and appreciation this research work would not have been accomplished. I pay my deep regards to him and his greatness. I am also very grateful to other lecturers and the entire staff of my department for their support during the course of my study. I am very obliged to the Universiti Teknologi PETRONAS (UTP) Malaysia, for awarding me a fully funded scholarship. Without it, I would not be able to pursue my studies. With my deep heart, I acknowledge my wife, who took care of my children while I was busy in my research activities. Her encouragement, motivations, support and understanding made me able to work on my research activities and to write this thesis.

The proposed System is an automatic front-end Code Generator Expert System (CGES) for ensuring wider business application for the generation of GUI with a source code for Databases. Safe keeping of data for smooth transaction in business has always been a matter of concern. With the help of the proposed CGES, with economy of effort and time, a customizable database application may be produced with a simple wizard. The CGES requires a database as a prerequisite input. Once the normalized database is featured with a diagram, the CGES shall apply techniques according to the pre-defined algorithm; the complete application with source code in various modules shall automatically produce. By selecting the CGES solutions, an N-tier application shall give rise to a product, comprising of SQL server queries, Object Oriented features and modules. The results prove on a test working principles of the system are written in MS SQL Server and on the Visual Basic.NET source code generated by CGES.

Internet is the most significant source of getting up thoughts, surveys for a product, and reviews for any type of service or activity. A Bulky amount of reviews are produced on daily basis on the cyberspace about online products and objects. For example, many individuals share their remarks, reviews and feelings in their own language utilizing social media networks such as twitter and so on. Considering their colossal Quantity and size, it is exceedingly knotty to look at with and interpret specified surveys. Sentiment Analysis (SA) aims at extracting people's opinion, felling and thought from their reviews in social websites. SA has recently gained significant consideration, however the vast majority of the resources and frameworks constructed so far are tailored to English as well as English like Western languages. The requirement for designing frameworks for different dialects is expanding, particularly as blogging and micro-blogging sites are becoming popular. This paper presents a comprehensive review of approaches of Urdu sentiment analysis and outlines of relevant gaps in the literature.