Scopus Index Publications
Permanent URI for this communityhttps://rda.sliit.lk/handle/123456789/2162
This collection consists of all Scopus-indexed publications produced by SLIIT researchers. Scopus is recognized worldwide as a leading and reputable academic indexing database.
Browse
2 results
Search Results
Publication Embargo Bi-directional long short-term memory based ensemble deep learning framework for non-linear steam turbine power forecasting: a biomass fuelled case study(Elsevier Ltd, 2026-04-10) Perera, H; Jayasekara, S; Wijesinghe, R.E; Silva, B. N; Cha, HIn palm oil manufacturing, steam turbines powered by biomass fuel are central to energy generation. However, fluctuating load demands and temporal variations lead to inefficiencies, while limited and variable supply of biomass waste constrains boiler feed flexibility. Current index-based boiler feeding methods overlook actual load demands and waste availability, resulting in significant energy wastage. This study presents a novel ensemble deep learning model combining Bidirectional Long Short-Term Memory (Bi-LSTM) and Gated Recurrent Units (GRU) with Attention Layers, trained on an eight-year operational dataset with structured preprocessing and feature selection, to forecast steam turbine power generation. The model captures complex non-linear temporal patterns more effectively than conventional and standalone ML models, achieving a Root Mean Square Error (RMSE) of 0.0684, Mean Absolute Error (MAE) of 0.0414, and an R-squared (R2) value of 0.9832, which outperformed eight benchmark models by approximately 25% in prediction accuracy. Additionally, the framework incorporates operational parameters such as kVA, total energy, and Fresh Fruit Bunch (FFB) production to dynamically optimise biomass feed rates, balancing energy output with resource availability. This approach minimises energy wastage, reduces grid reliance, and promotes both sustainability and profitability.Publication Embargo Data-centric single teacher guided knowledge distillation for alleviating sub-optimal supervision in image classification(Elsevier Ltd, 2026-02-23) Sharma, K; Silva, B. NIn recent years, larger, deeper, and more complex deep learning models have emerged as a result of advancements in deep learning techniques. Nevertheless, the computational costs have also increased with the growing model size. Thus, Knowledge Distillation has evolved into a cornerstone in contemporary machine learning, facilitating the transfer of knowledge from cumbersome teacher models to more compact student models. However, student learning is persistently challenged by sub-optimal supervision caused by erroneous and ambiguous teacher predictions. Moreover, the learning process is further deteriorated by the complications introduced through frequently encountered noisy labels in real-world datasets. Existing methods often resort to the ensemble of teachers, introducing additional complexity. We propose a novel, simple, and efficient learning method, Corrective Knowledge Distillation (CKD), to alleviate these drawbacks while relying solely on a single-teacher model. The proposed work employs a two-phase learning paradigm. In the initial phase, the teacher selectively teaches extremely confident knowledge to the student, and in the subsequent phase, the student leverages its own past learning experiences, conditioning its knowledge acquisition on the guidance of the teacher. The proposed method consistently exhibits superior performance in addressing sub-optimal supervision, as evidenced by comprehensive experiments on benchmark datasets such as CIFAR-100, CIFAR-100N-Fine, and ImageNet-1K. Notably, CKD surpasses established baselines, achieving substantial accuracy gains of up to 3.53% in real-world scenarios. Furthermore, CKD exhibits exceptional robustness in highly noisy environments, outperforming ensemble techniques by a significant margin of up to 5.18%. Our code is available at https://github.com/Karthick47v2/ckd.