Robust and Unified VLC Decoding System for Square Wave Quadrature Amplitude Modulation Using Deep Learning Approach

Alfarozi, Syukron Abu Ishaq; Pasupa, Kitsuchart; Hashizume, Hidetoshi; Woraratpanya, Kuntpong; Sugimoto, Masanori

doi:10.1109/access.2019.2952465

Cited by 11 publications

(2 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The implementation of LA is generally based on traditional machine learning and statistical models [3]- [7]. However, the merging of deep learning model [8], which is adopted on various applications in medical imaging [9], telecommunication [10], economics [11], also become a new standard to be adopted in LA [12]- [14].…”

Section: Introductionmentioning

confidence: 99%

Transformer Encoder Model for Sequential Prediction of Student Performance Based on Their Log Activities

Kusumawardani

Alfarozi

2023

IEEE Access

Self Cite

View full text Add to dashboard Cite

Learning management systems (LMSs) have been used massively due to the growing utilization of distance learning. This advancement has led to increased educational data that can be analyzed to improve the quality of the learning process. Learning analytics (LA) is one of the most important methods that can be used to analyze student performance. In this paper, we proposed an LA method based on deep learning, i.e., transformer encoder, to sequentially predict the student's final performance based on log activities provided by an LMS. The objective is to predict at-risk students of failing so that they can be mitigated as soon as possible. The proposed model was evaluated on the Open University LA Dataset (OULAD) for daily or weekly prediction. The results show that the model could predict at the early stage with an accuracy of 83.17% on withdrawn versus pass-distinction classes. Meanwhile, for other tasks, i.e., withdrawn-fail versus pass-distinction and fail versus pass-distinction tasks, the accuracy was at least 76% at the early stage. The proposed model was compared to the LSTM model. We found that the transformer encoder performed better than the LSTM, with the average difference values from 1% to 3% in terms of accuracy and from 3% to 7% in terms of F1-score for all tasks, based on the statistical testing. Furthermore, the ablation study using positional encoding, different feature aggregation methods, and weighted loss function for the imbalanced class problem was done. In OULAD, we found that model without positional encoding was better in all cases. In contrast, weekly feature aggregation and the use of the weighted loss function performed better only for some cases. INDEX TERMS Learning analytics, transformer encoder, student at-risk prediction, massive open online courses, sequential model, imbalanced dataset.

show abstract

Section: Introductionmentioning

confidence: 99%

Transformer Encoder Model for Sequential Prediction of Student Performance Based on Their Log Activities

Kusumawardani

Alfarozi

2023

IEEE Access

Self Cite

View full text Add to dashboard Cite

show abstract

“…Work reported in [18] and [20] demonstrate the user positioning in the VLC systems with ANN estimation. References [21], [22] and [23] investigate the ANN-based physical layer techniques in the VLC systems, but to the best of our knowledge, there is no work applying an ANN to predict the transmitter patterns and data rates of a MIMO-VLC system.…”

Section: Introductionmentioning

confidence: 99%

High-Speed Adaptive MIMO-VLC System With Neural Network

Dong

O’Brien

2022

J. Lightwave Technol.

View full text Add to dashboard Cite

Rooms are typically lit with multiple luminaires, which open the possibility of creating Multiple-Input Multiple-Output (MIMO) Visible Light Communications (VLC) systems.At its most complex, the luminaires in the system might transmit different data at different power levels to a terminal containing multiple receivers, allowing a substantial increase in data rates. However, crosstalk between the transmitted channels, dependent on the location and orientation of the receiver, may cause the best strategy to be to group transmitters together and transmit the same data stream. In this paper, we report a transmitter coordination algorithm determining how to use the transmitters optimally as the receiver location varies. The data rate using this approach is on average 41 per cent higher than the conventional spatial multiplexing approach. Neural networks are then employed in the coordination algorithm. It increases the speed of operation by a factor of four compared to the initial coordination algorithm while achieving the same level of performance. The neural network also gives good performance for room geometries and receiver orientations outside the scope of the training data. Finally, the neural network is benchmarked against a water-filling algorithm approach with maximum system capacity. The neural network shows a factor of three speed-up in computing time with only a 12 per cent reduction in the average data rate. These results show the potential of the approach to achieve a near-maximal data rate with a straightforward and efficient channel selection technique.

show abstract