Department of Computer Science and Software Engineering

Permanent URI for this collectionhttp://rda.sliit.lk:8081/handle/123456789/3214

Browse

Now showing 1 - 5 of 5

Open Access
A Graph Pointer Network-Based Multi-Objective Deep Reinforcement Learning Algorithm for Solving the Traveling Salesman Problem
(MDPI, 2023-01-13) Perera, J; Liu, S.H; Mernik, M; Črepinšek, M; Ravber, M
Traveling Salesman Problems (TSPs) have been a long-lasting interesting challenge to researchers in different areas. The difficulty of such problems scales up further when multiple objectives are considered concurrently. Plenty of work in evolutionary algorithms has been introduced to solve multi-objective TSPs with promising results, and the work in deep learning and reinforcement learning has been surging. This paper introduces a multi-objective deep graph pointer network-based reinforcement learning (MODGRL) algorithm for multi-objective TSPs. The MODGRL improves an earlier multi-objective deep reinforcement learning algorithm, called DRL-MOA, by utilizing a graph pointer network to learn the graphical structures of TSPs. Such improvements allow MODGRL to be trained on a small-scale TSP, but can find optimal solutions for large scale TSPs. NSGA-II, MOEA/D and SPEA2 are selected to compare with MODGRL and DRL-MOA. Hypervolume, spread and coverage over Pareto front (CPF) quality indicators were selected to assess the algorithms’ performance. In terms of the hypervolume indicator that represents the convergence and diversity of Pareto-frontiers, MODGRL outperformed all the competitors on the three well-known benchmark problems. Such findings proved that MODGRL, with the improved graph pointer network, indeed performed better, measured by the hypervolume indicator, than DRL-MOA and the three other evolutionary algorithms. MODGRL and DRL-MOA were comparable in the leading group, measured by the spread indicator. Although MODGRL performed better than DRL-MOA, both of them were just average regarding the evenness and diversity measured by the CPF indicator. Such findings remind that different performance indicators measure Pareto-frontiers from different perspectives. Choosing a well-accepted and suitable performance indicator to one’s experimental design is very critical, and may affect the conclusions. Three evolutionary algorithms were also experimented on with extra iterations, to validate whether extra iterations affected the performance. The results show that NSGA-II and SPEA2 were greatly improved measured by the Spread and CPF indicators. Such findings raise fairness concerns on algorithm comparisons using different fixed stopping criteria for different algorithms, which appeared in the DRL-MOA work and many others. Through these lessons, we concluded that MODGRL indeed performed better than DRL-MOA in terms of hypervolumne, and we also urge researchers on fair experimental designs and comparisons, in order to derive scientifically sound conclusions.
Embargo
"Talking Books" : A Sinhala Abstractive Text Summarization Approach for Sinhala Textbooks
(IEEE, 2023-05-23) Rathnayake, B.R.M.S.R.B.; Manathunga, K; Kasthurirathna, D
The ability for books to talk would be an exciting concept, and this research discussion paves the path for an identical approach. The research objectives discussed in this paper address several burning problems, solve them and adapt them to future technological enhancements from a Sri Lankan context. Burning problems include reducing printing costs for textbooks, addressing students’ health, promoting green technology, and identifying a suitable summarising approach to the native language, Sinhala resulting in students’ learning ease. Other symptoms for the betterment indicate paths taken to reduce the weight of school bags carried by students, reduce paper usage by the government on printing textbooks, and spread technological awareness to teenagers regarding e-Learning. Textbooks issued by the government will be digitized and centralized into a single system that the government officials themselves can administer. The paper discusses limited hindsight literature and proposes 2 new algorithms for abstractive and extractive summarization for Sinhala text. The 2 algorithms are compared against one another in terms of performance, efficiency, precision and accuracy. Experts in the education domain have verified the derived summary of both algorithms. The deliverable artefacts are the mobile application, a RESTful auto-summarization plugin service, and new data sets extracted to train the GPT-3 models.
Embargo
Static Code Analyser to Enhance Developer Productivity
(IEEE, 2023-05-23) Peiris, D. R. I.; Kodagoda, N
One of the main metrics important to software development is productivity. The measure of productivity helps a organization/ individual or team identify how well their employees/ software runs and how they could improve. In this research paper, a new software is proposed along with a background study of how improving code quality will improve developers productivity and efficiency. The software proposed to aid developer productivity is a custom built static code analyser using python and it’s rule set has been tailored through carefully selected research papers and with feedback from experts of the software industry
Open Access
COVID-19 symptom identification using Deep Learning and hardware emulated systems
(Elsevier, 2023-06-28) Liyanarachchi, R; Wijekoon, J; Premathilaka, M; Vidhanaarachchi, S
The COVID-19 pandemic disrupted regular global activities in every possible way. This pandemic, caused by the transmission of the infectious Coronavirus, is characterized by main symptoms such as fever, fatigue, cough, and loss of smell. A current key focus of the scientific community is to develop automated methods that can effectively identify COVID-19 patients and are also adaptable for foreseen future virus outbreaks. To classify COVID-19 suspects, it is required to use contactless automatic measurements of more than one symptom. This study explores the effectiveness of using Deep Learning combined with a hardware-emulated system to identify COVID-19 patients in Sri Lanka based on two main symptoms: cough and shortness of breath. To achieve this, a Convolutional Neural Network (CNN) based on Transfer Learning was employed to analyze and compare the features of a COVID-19 cough with other types of coughs. Real-time video footage was captured using a FLIR C2 thermal camera and a web camera and subsequently processed using OpenCV image processing algorithms. The objective was to detect the nasal cavities in the video frames and measure the breath cycles per minute, thereby identifying instances of shortness of breath. The proposed method was first tested on crowd-sourced datasets (Coswara, Coughvid, ESC-50, and a dataset from Kaggle) obtained online. It was then applied and verified using a dataset obtained from local hospitals in Sri Lanka. The accuracy of the developed methodologies in diagnosing cough resemblance and recognizing shortness of breath was found to be 94% and 95%, respectively.
Embargo
WONGA: The Future of Personal Finance Management – A Machine Learning-Driven Approach for Predictive Analysis and Efficient Expense Tracking
(IEEE, 2023-07-10) Uyanahewa, M.I.R; Jayawardana, G.V.H.D; Bandara, M.B.D.N; Hapugala, H.A.V.V; Attanayaka, B
The financial literacy of Sri Lankans is relatively low, leading to difficulties in managing personal finances. This research presents a smart solution to simplify the complexities associated with money management and assist individuals in managing their finances more efficiently to achieve better financial health without requiring a comprehensive knowledge of money management from the user. The proposed system automates personal finance management with minimal user effort, reducing manual data entry by tracking cash flow by utilizing SMS messages and expense bills to extract bank transaction data and cash expenditures. Each extracted expense will automatically be categorized into the correct expense category. The system also generates a custom budget plan for each user based on spending patterns to help them stay on the budget throughout the month and avoid irrational overspending. Furthermore, the system provides a mechanism to predict future expenses associated with upcoming events based on calendar events, allowing users to devise the most efficient budget plan and avoid facing financially unprepared events in the upcoming month. All these smart solutions are bundled up in the "Wonga" mobile application to help users make better financial decisions to achieve personal financial success.