Sapumal Ahangama

Widespread adoption of information systems has created a variety of digital data traces of users such as profile attributes, relational data and activity traces. As users use multiple information systems for different tasks that are not directly linked with each other, the data will be fragmented. Interconnecting these data traces of a user from various information systems is an important research problem known as “User Entity Resolution” (UER). Much of the current UER methods depend on the… Show more

Widespread adoption of information systems has created a variety of digital data traces of users such as profile attributes, relational data and activity traces. As users use multiple information systems for different tasks that are not directly linked with each other, the data will be fragmented. Interconnecting these data traces of a user from various information systems is an important research problem known as “User Entity Resolution” (UER). Much of the current UER methods depend on the similarity of profile attributes and network data. With the reliability and availability of user profile attributes decreasing, it is important to develop methods that incorporate user activity data to improve UER. We propose a novel approach to incorporate user activity data in UER. A deep neural network model is used as a cross-domain transfer model. Our evaluation results indicate that the proposed model has a far superior ability to rank users. Show less

Widespread adoption of information systems has created variety of digital data traces concerning the users, their relationships and personal activities. Interconnecting the digital data traces of a user from various information systems is not an easy task and is an open problem, commonly known as user entity resolution. Since prior research has loosely focused on user shared content for user entity resolution, in this study our intention is to develop methods that would enable the incorporation… Show more

Widespread adoption of information systems has created variety of digital data traces concerning the users, their relationships and personal activities. Interconnecting the digital data traces of a user from various information systems is not an easy task and is an open problem, commonly known as user entity resolution. Since prior research has loosely focused on user shared content for user entity resolution, in this study our intention is to develop methods that would enable the incorporation of user shared content in the task as an approximate matching method. We proposed the idea of identifying latent characteristics from user shared content and matching based on the similarity of these characteristics while modeling the cross domain differences of the data. In order to do so, we used topic models to represent external data sources as a knowledge base and to represent users as profiles of topic distributions. Similar user profiles will be searched using locality sensitive hashing to identify similarities in different localities in the data and to model cross domain differences. Show less

In spite of the virtual nature and the system operating purely based on pre-formulated computer algorithms, cryptocurrency networks have reached greater popularity with a significant follower base with people placing trust on the system operation. As credibility is an important factor for systems facilitating financial transactions, in this study we will be presenting a simple model facilitating identification of relevant important factors to be considered by users and a methodology for… Show more

In spite of the virtual nature and the system operating purely based on pre-formulated computer algorithms, cryptocurrency networks have reached greater popularity with a significant follower base with people placing trust on the system operation. As credibility is an important factor for systems facilitating financial transactions, in this study we will be presenting a simple model facilitating identification of relevant important factors to be considered by users and a methodology for assessing the credibility of cryptocurrency networks. We identify two routes, systems and the psychological perspective in the credibility assessment process which varies with the user expertise. Show less

Expectation shortfall is a common occurrence in outsourcing. Prior literature suggests that strategies such as strict contract terms and proper evaluation of the vendor capabilities are adopted to avoid expectation shortfall. However, in the case of highly specialised technical products custom made to vendor requirements (i.e., B2B IT innovation), traditional strategies in managing outsourcing projects may not work as expected. This is mainly due to the complexity of the product requirements… Show more

Expectation shortfall is a common occurrence in outsourcing. Prior literature suggests that strategies such as strict contract terms and proper evaluation of the vendor capabilities are adopted to avoid expectation shortfall. However, in the case of highly specialised technical products custom made to vendor requirements (i.e., B2B IT innovation), traditional strategies in managing outsourcing projects may not work as expected. This is mainly due to the complexity of the product requirements and the inability to assess the scope of the project in depth at the beginning. In this research, we adopt the vendor’s perspective to better understand how organizations in the highly specialized B2B IT innovation handle outsourced projects to avoid expectation shortfall. We uncover a dynamic innovation process which the client and the vendor go through. In addition, we suggest strategies to achieve B2B IT innovation in a win-win scenario while elucidating reasons of failure. Show less

Motion Blur due to the relative motion between the camera and object can seriously degrade image quality. We present an FPGA based motion blur detection and correction algorithm which is implemented on top of a configurable soft-processor based architecture. The system consists of two main modules. The blur detection module identifies the blur length and angle, and the restoration module uses regularized inverse filtering to remove the motion blur. The Processing algorithms are implemented as… Show more

Motion Blur due to the relative motion between the camera and object can seriously degrade image quality. We present an FPGA based motion blur detection and correction algorithm which is implemented on top of a configurable soft-processor based architecture. The system consists of two main modules. The blur detection module identifies the blur length and angle, and the restoration module uses regularized inverse filtering to remove the motion blur. The Processing algorithms are implemented as separate cores on the FPGA fabric where the soft processor core is only used for managing system configuration. The system can achieve a frame rate of 15fps for a 720p HD video stream. Show less