International Conference on Advancements in Computing [ICAC]

Permanent URI for this communityhttps://rda.sliit.lk/handle/123456789/312

The International Conference on Advancements in Computing (ICAC) is organized by the Faculty of Computing of the Sri Lanka Institute of Information Technology (SLIIT) as an open forum for academics along with industry professionals to present the latest findings and research output and practical deployments in computing.

The primary objective of ICAC is to promote innovative research that addresses real-world challenges and contributes to the social well-being of communities. The conference provides a dynamic platform for researchers from around the world to present groundbreaking findings, exchange ideas, and establish meaningful collaborations.

https://icac.lk

Browse

Filters

2019 - 201912020 - 20252

Advanced Search

Filter by

Settings

Author: search.filters.author.Fernando, SHas files: Yes

Search Results

Now showing 1 - 3 of 3
  • Thumbnail Image
    ItemEmbargo
    HCLIP: Beyond CLIP for Cost-Effective Multimodal Retrieval in Education
    (Institute of Electrical and Electronics Engineers Inc., 2025-12-09) Weerasinghe, S; Gunatunga, O; Dewpura, W; Fernando, S; Kasthurirathna, D; Rathnayake, S
    Multimodal retrieval systems have gained significant attention due to their ability to process and cross-retrieve data containing images and text. However, the factors such as high cost of development, limitation on resources, and the proper addressing of the modality gap, the inherent representational differences between modalities pose a challenge to building effective and efficient retrieval models. In this work, we propose a low-resource, cost-efficient hybrid multimodal retrieval model that integrates Contrastive Language-Image Pre-training (CLIP) and All-MiniLM-L6-v2 to create a shared embedding space while storing raw images in an unstructured database. Our primary contributions include (1) the development of a hybrid model that outperforms CLIP-native retrieval, (2) a novel bidirectional neural network alignment technique that brings textual and visual modalities closer together, and (3) a comprehensive analysis of the modality gap's impact on downstream retrieval performance. Through proper evaluation using transparent techniques such as Mean Reciprocal Rank (MRR) and Cosine-Weighted MRR, our method demonstrates improved retrieval accuracy over baseline approaches. Experimental results exhibit that a lower modality gap does not always prove to be efficient on the downstream retrieval. Our findings pave the way for more efficient, adaptable, and cost-effective multimodal retrieval methodologies in low-resource environments, not limited to the education domain.
  • Thumbnail Image
    PublicationEmbargo
    Smart Caring System for Ornamental Fish
    (IEEE, 2022-12-09) Fernando, S; Jayaweera, N; Pitawala, S; Kaushalya, R; Ratnayake, P; Siriwardana, S
    Ornamental Fish Industry continues to be one of the fastest growing sectors worldwide. Healthy fish production at aquariums requires intensive care and ensures a stable and an optimum production environment inside the fish tanks, which is a challenging task. Unfortunately, due to the limitations in fish industry, productivity of well-developed, healthy fish has drastically depreciated. Limited skills and knowledge of aquarists have been a challenging task which has led to inaccurate predictions on certain factors such as quantification and length of estimation, amounts and types of fish food and servicing the filters at proper time intervals. Existing aquariums depend on the experience and availability of the aquarists, which can be a challenging process in real life. Developing a system to regulate these major concerns is a prominent solution. This research is done to propose an automated method, with the help of several fish aquariums and existing research papers, to encounter the mentioned major concerns which affects the aquarists and other stakeholders.
  • Thumbnail Image
    PublicationEmbargo
    VTutor: A Platform for Improving Searchability and Interactivity of Recorded Lectures
    (IEEE, 2019-12-05) Karunaratna, D; Hettiarachchi, I; Fernando, S; Epa, S; Kodagoda, N; Suriyawansa, K
    Recorded lectures have gained popularity as a method of delivering lecture content as they give learners a host of distinct advantages such as the ability to follow lectures without time or location constraints and to consume the lectures at their own pace. However, despite such benefits, they have a tendency to be lengthy and tedious to watch. They also prove cumbersome when precise information needs to be extracted from the content. Another drawback is that recorded lecture videos fail to show the connection between the lecture and its support material such as slides and questionnaires. Though many of the existing platforms allow editing lecture videos for more interactivity, the methods employed by these platforms have always been manual, and therefore time intensive. VTutor is a web platform that aims to address these drawbacks by introducing automation into the video enhancement process, eventually combining the lecture material to create an enhanced user experience. Specifically, VTutor allows users to navigate through a lecture video using subtopics, its corresponding slides and code samples. Furthermore, it is equipped with the ability to automatically generate questions by scraping the internet based on provided keywords thus improving the level of engagement that a learner has with the lecture.