AU ECE - AI & Machine Learning Research Group (AIML)

AI & Machine Learning Research Group (AIML)

The AI & Machine Learning (AIML) Group's interests are in machine learning and deep learning. The research of the group falls within the following major areas:

Hyperspectral Imaging

Research interests include remote sensing, pattern recognition and astronomy for problems of classification and unmixing, particularly in the presence of small training samples.

- - B. Baassou, M. He, M.I. Farid, and S. Mei, "Hyperspectral image classification based on iterative Support Vector Machine by integrating spatial-spectral information", in Proc. IEEE IGARSS, July 2013, pp. 1023-1026.

Dimension Reduction

Research interests include exploiting low-rank structure for developing algorithm for matrix completion, inverse covariance estimation, and reduce-rank regression.

G. Nabi & S.A. Pasha, "Convex & Non‐convex Approaches for the Matrix Completion Problem," Proc. 19th IBCAST, Aug, 2022.
A. Safdar, M. Hassan, J.Y. Kim, M.I. Farid, et al., "FF-PCA-LDA: Intelligent Feature Fusion Based PCA-LDA Classification System for Plant Leaf Diseases," Applied Sciences, vol. 12(7: 3514), Mar 2022.
V. Solo & S.A. Pasha, "Point-Process Principal Components Analysis via Geometric Optimization", Neural Computation, vol.25(1), 2013.
S.A. Pasha & V. Solo, "Hawkes-Laguerre Dynamic Index Models for Point Processes," in Proc. 52nd IEEE CDC, Firenze, Italy, 2013, pp. 7028-7033.
S.A. Pasha & V. Solo, "Hawkes-Laguerre Reduced Rank Model for Point Process," in Proc. 38th IEEE ICASSP, Vancouver, BC, Canada, 2013, pp. 6098-6102.

Time Series Analysis

Research interests include analysis of high-frequency financial time series that are characterized by nonuniformly sampled observations.

A. Mihirana, R. Davis, S.A. Pasha & P. Leong, "Forecasting Financial Time-series with Grammar Guided Feature Generation", Computational Intelligence, vol.33(2), 2017.
S.A. Pasha & P. Leong, "Cluster Analysis of High-Dimensional High-Frequency Financial Time Series", in Proc. IEEE SSCI, 2013.
V. Solo & S.A. Pasha, "A Test for Independence between a Point Process and an Analog Signal", Journal of Time Series Analysis, vol.33(5), 2012.

AIML for Next-Generation Energy Systems:

This research stream focuses on operational intelligence that draws actionable insights from energy related data and delivers both offline and online decision making across power systems. This primarily involves investigation, development and applications of machine learning algorithms and optimization techniques for the future integrated energy systems, not limited to, but in the following potential areas.

Resilience in Integrated Energy Management Systems
Transformer Health Diagnostics
Vehicle-to-Grid (V2G) Integration and its Impact
Electricity Theft Detection
Short-Term, Mid-Term, and Long-Term Load Forecasting

Below is a list of relevant publications in this stream.

H.M. Hussain, A. Ahmad, A. Narayanan, P.H.J. Nardelli, and Y. Yang. "Benchmarking of Heuristic Algorithms for Energy Router-Based Packetized Energy Management in Smart Homes." IEEE Systems Journal, accepted, in-press 2022.
A. Ahmad, and J.Y. Khan. "Optimal sizing and management of distributed energy resources in smart buildings." Energy 244: 123110, 2022.
A. Ahmad, and J.Y. Khan. "Real-time load scheduling, energy storage control and comfort management for grid-connected solar integrated smart buildings." Applied Energy 259: 114208, 2022.
Z. Aslam, N.Javaid, A. Ahmad, A. Ahmed, and S.M. Gulfam. "A combined deep learning and ensemble learning methodology to avoid electricity theft in smart grids." Energies vol. 13(21): 5599, 2020.
A. Ahmad, and J.Y. Khan. "Real-time load scheduling and storage management for solar powered network connected EVs." IEEE Transactions on Sustainable Energy vol. 11(3), pp. 1220-1235, 2019.