Research Publications Authored by SLIIT Staff
Permanent URI for this communityhttps://rda.sliit.lk/handle/123456789/4195
This collection includes all SLIIT staff publications presented at external conferences and published in external journals. The materials are organized by faculty to facilitate easy retrieval.
Browse
2 results
Search Results
Publication Open Access Topology and size optimization of trusses using graph neural networks: Towards efficient surrogate modeling(Elsevier Ltd, 2026-06) Ariyasinghe, N; Wickremasinghe, T; Weeratunge, H; Mallikarachchi, C; Herath, SReal-time structural optimization of trusses using machine learning techniques, incorporating both topology and size optimization, is more effective in discrete domains than in continuous ones, with Graph Neural Networks (GNNs) showing strong potential. However, the impact of convolutions in GNNs is not yet fully understood, limiting their full applicability. This paper presents a GNN-based surrogate model for real-time structural optimization and identifies the most suitable convolution type for this task. The study assesses the predictive performance of models trained on a dataset of optimized structures spanning a range of loads, boundary conditions, and design domain sizes. The resulting model effectively predicts both optimal topology and member sizes once the design parameters are provided. Eight graph convolution types are investigated to identify the most suitable method, alongside an evaluation of optimal network architectures. Among the tested approaches, Generalised Graph Convolution achieves the highest accuracy, followed by Topology Adaptive Graph Convolution, producing near-ideal topology predictions for most test cases and maintaining section size prediction errors within ±2% across all test data points. This framework demonstrates strong potential for broader applications.Publication Embargo Spatio-temporal graph neural network based child action recognition using data-efficient methods: A systematic analysis(Elsevier Inc, 2025-06-03) Mohottala, S; Gawesha, A; Kasthurirathna, D; Samarasinghe, P; Abhayaratne, CThis paper presents implementations on child activity recognition (CAR) using spatial–temporal graph neural network (ST-GNN)-based deep learning models with the skeleton modality. Prior implementations in this domain have predominantly utilized CNN, LSTM, and other methods, despite the superior performance potential of graph neural networks. To the best of our knowledge, this study is the first to use an ST-GNN model for child activity recognition employing both in-the-lab, in-the-wild, and in-the-deployment skeleton data. To overcome the challenges posed by small publicly available child action datasets, transfer learning methods such as feature extraction and fine-tuning were applied to enhance model performance. As a principal contribution, we developed an ST-GNN-based skeleton modality model that, despite using a relatively small child action dataset, achieved superior performance (94.81%) compared to implementations trained on a significantly larger (x10) adult action dataset (90.6%) for a similar subset of actions. With ST-GCN-based feature extraction and fine-tuning methods, accuracy improved by 10%–40% compared to vanilla implementations, achieving a maximum accuracy of 94.81%. Additionally, implementations with other ST-GNN models demonstrated further accuracy improvements of 15%–45% over the ST-GCN baseline. The results on activity datasets empirically demonstrate that class diversity, dataset size, and careful selection of pre-training datasets significantly enhance accuracy. In-the-wild and in-the-deployment implementations confirm the real-world applicability of above approaches, with the ST-GNN model achieving 11 FPS on streaming data. Finally, preliminary evidence on the impact of graph expressivity and graph rewiring on accuracy of small dataset-based models is provided, outlining potential directions for future research. The codes are available at https://github.com/sankamohotttala/ST_GNN_HAR_DEML.