Browsing by Author "Ahraz Asif, Mohammad"

Now showing 1 - 2 of 2

APIs as Digital Innovation Objects - Implications on Requirements Elicitation
(2017-06-22) Törnlund, Mikaela; Ahraz Asif, Mohammad; Göteborgs universitet/Institutionen för data- och informationsteknik; University of Gothenburg/Department of Computer Science and Engineering
APIs as a way to expose business assets have become common across industry. Consequently insights into strategic API development practices are becoming more and more relevant. However domain specific knowledge for making strategic design decisions for APIs is often lacking. In this study we present and analyse a method for requirement elicitation for APIs created through use of the design science methodology alongside careful consideration of standard requirements elicitation practices and API strategies - specifically insights from the ’API as Digital Innovation Objects’ model on API strategy. We consider this due to a lack of awareness about what information is significant for such strategic design decisions, especially during requirements elicitation. Our findings show that there are several aspects of strategic API design that can be considered in order to improve requirements elicitation. The results imply that strategic API considerations change requirements elicitation by changing how it is carried out - not necessarily the tasks and activities conducted.
Deep Neural Network Compression for Object Detection and Uncertainty Quantification
(2019-11-21) Ahraz Asif, Mohammad; Tzelepis, Georgios; Göteborgs universitet/Institutionen för data- och informationsteknik; University of Gothenburg/Department of Computer Science and Engineering
Neural networks have been notorious for being computational expensive. Their demand for hardware resources prohibits their extensive use in embedded devices and puts restrictions on tasks like real time tracking. On top of that, neural networks are usually deterministic and provide no uncertainty which is crucial on safety decision tasks and physical sciences. In this work techniques were developed to reduce the computational cost of neural networks, such as Model Compression infused with a novel dynamical clustering and Knowledge Distillation while estimating the impact of such techniques on the uncertainty of the model by using Bayesian neural networks. A brief introduction is made on deep learning and the tools used. Furhtermore the ideas of Model Compression, Knowledge Distillation and Bayesian Deep Learning were extended analytically. All three approaches were tied together followed by the final discussion.