Dr. Andrees Chattha
Overview of research interest & accomplishments
My research lies at the intersection of applied mathematics, statistical signal processing, and machine learning. The overarching goal of my work is to develop mathematical models, statistical inference methods, and scalable algorithms for extracting information from stochastic signals and large-scale distributed data.
A central theme of my research is the integration of optimization, Bayesian inference, and distributed computation to design algorithms capable of operating in modern data environments such as sensor networks, cyber-physical systems, and large-scale machine learning platforms. These systems often involve uncertainty, noisy observations, and decentralized data, requiring new theoretical and computational tools for reliable estimation and learning.
Research interests and expertise
1. Optimization and Machine Learning for Large-Scale Data
Recent work focuses on optimization methods for machine learning and large-scale data analysis. Many modern learning algorithms require solving high-dimensional optimization problems under stochastic and distributed settings.
My research investigates stochastic and distributed optimization algorithms that combine statistical modeling with efficient numerical methods. These approaches enable learning from massive datasets while accounting for uncertainty and limited communication between computing nodes. This line of work connects closely with problems in distributed signal processing, where multiple agents cooperate to solve global inference or optimization tasks.
2. Distributed Estimation and Data Fusion
Another major direction of my research concerns distributed estimation and data fusion in networked systems, particularly for sensor networks observing spatially distributed stochastic processes. In these systems, individual nodes collect noisy measurements and must collaborate to estimate signals or parameters of interest without relying on centralized processing.
I develop distributed filtering and learning algorithms based on adaptive and cooperative strategies such as:
- Least Mean Square (LMS) adaptive methods
- Diffusion learning algorithms
- Consensus-based estimation techniques
These algorithms allow networks of agents to perform cooperative inference using only local communication, making them suitable for applications such as environmental monitoring, risk assessment, and large-scale sensing systems.
3. Statistical Signal Processing and Stochastic Filtering
My work also explores statistical methods for parameter and state estimation in stochastic signal models. These problems arise widely in time series analysis, dynamic systems, and predictive modeling.
I develop both batch and recursive estimation methods for signal and time series models with Gaussian and non-Gaussian noise. The methodologies include techniques such as:
- Least squares and maximum likelihood estimation
- Expectation–Maximization (EM) algorithms
- Kalman and nonlinear filtering methods
- Particle filtering and Monte Carlo approaches
These methods provide tools for learning and inference in complex stochastic systems and form the mathematical foundation for many modern approaches in machine learning and data science.
Selected publications and outputs
Journal Articles, Conference Proceedings, Technical Reports, and Theses
- A. C. Andrees. Performance evaluation of single and multi-line competitive resource sharing systems. M.Phil dissertation, University of London, 2003.
- A. C. Andrees and M. Gani. Probabilistic algorithms for generating centralized Voronoi tessellations and their application in the deployment of sensor networks. Technical Report, King’s College London, 2005.
- M. Gani and A. C. Andrees. Optimal deployment control for heterogeneous mobile sensor networks. 9th International Conference on Control, Automation, Robotics and Vision, Singapore, 2006.
- A. C. Andrees, M. Gani, and F. Yang. Decentralized Kalman filtering algorithm with uncertain signal models for heterogeneous sensor networks. European Control Conference, 2007.
- A. C. Andrees and M. Gani. Distributed particle filtering for nonlinear stochastic processes over sensor networks. IEEE Control Systems Society Colloquium, London, 2007.
- A. C. Andrees, M. Gani, and F. Yang. Decentralized robust Kalman filtering for uncertain stochastic systems over heterogeneous sensor networks. EURASIP Journal on Signal Processing, 88(2), 1919–1928, 2008.
- A. C. Andrees and M. Gani. Distributed estimation for nonlinear stochastic signals over sensor networks. IEEE ICASSP, Las Vegas, 2008.
- A. C. Andrees. A novel approach to distributed filtering for nonlinear stochastic dynamical systems over sensor networks. International Conference on Applied Mathematics and Computing, Bulgaria, 2008.
- A. C. Andrees. A survey of sequential Monte Carlo simulations and particle filters: theory, methods, and algorithms. Technical Report, King’s College London, 2009.
- A. C. Andrees. Parameter estimation of hidden Markov models using particle filtering algorithms. Technical Report, University of Essex, 2009.
- A. C. Andrees. Distributed and decentralized filtering algorithms for nonlinear stochastic systems over sensor networks. PhD Thesis, University of London, 2010.
- A. C. Andrees. Distributed filtering algorithm for nonlinear stochastic systems over homogeneous sensor networks. Accepted, Digital Signal Processing Journal.
- A. C. Andrees. Recursive least squares methods for parameter estimation of high-frequency time series models. Technical Report, University of Essex, 2014.
- A. C. Andrees. Noise reduction algorithms for discrete-time stochastic signals. Signal Processing Seminar, University of Lorraine, 2015.
- A. C. Andrees. Lecture notes on signal transformation theory and methods with applications. Shanghai Technical University, 2016.
- D. Zhang and A. C. Andrees. M-estimation of parameters for polynomial and exponential signal models. Bachelor’s Thesis, Shanghai Technical University, 2017.
- W. Xinyu and A. C. Andrees. Robust filtering for state estimation under uncertain models. Bachelor’s Thesis, Shanghai Technical University, 2017.
- A. C. Andrees. Lecture notes on stochastic difference equations with applications to finance. China Agricultural University, Beijing, 2018.
- H. Yongtai and A. C. Andrees. Parameter estimation for autoregressive models. Bachelor’s Thesis, China Agricultural University, 2019.
- Q. Feng and A. C. Andrees. Recursive estimation of time-varying parameters in autoregressive models. Bachelor’s Thesis, China Agricultural University, 2019.
- A. C. Andrees. Lecture notes on optimization methods and algorithms for machine learning. Wen–Kean University, 2019.
- L. King and A. C. Andrees. Exact computation of transfer functions for discrete-time dynamical systems. Bachelor’s Thesis, Wen–Kean University, 2020.
- T. Jinghuan and A. C. Andrees. Classical interpolation methods for signal reconstruction. Bachelor’s Thesis, Wen–Kean University, 2020.
- K. Victoria and A. C. Andrees. Analytical methods for polynomial coefficient determination. Bachelor’s Thesis, Wen–Kean University, 2020.
- G. Junzhe and A. C. Andrees. Evaluation of Riemann Zeta functions from a signal processing perspective. Bachelor’s Thesis, Wen–Kean University, 2020.
- Z. Adair and A. C. Andrees. Continued fractions and Padé approximation in signal processing. Bachelor’s Thesis, Wen–Kean University, 2021.
- A. C. Andrees. Regularized least squares with penalty functions leading to quadratic programming. Workshop on Statistical Learning, Wen–Kean University, 2021.
- A. C. Andrees. Distributed robust Kalman filtering via consensus averaging. Accepted, Digital Signal Processing Journal.
- A. C. Andrees. Reconstruction of time-evolving manifolds from noisy observations using Kalman filtering. Research Report, Forward College Paris, 2022.
- A. C. Andrees. Probabilistic reconstruction of time-evolving manifolds via known transformations. Research Report, Forward College Paris , 2023.
- A. C. Andrees. Integral methodologies for characterization of non-stationary stochastic signals in linear dynamical systems. Submitted, European Journal on Signal Processing,
- A. C. Andrees. Sequential and Bayesian Expectation–Maximization algorithms for parameter estimation of Weibull distributions. In preparation.
- A. C. Andrees. A cluster-based optimization framework for search engine resource allocation. In preparation, to be submitted to IEEE International Conference on Big Data, 2026.
Professional Memberships
- Member, Institute of Mathematics and its Applications (IMA)
- Member, Society for Industrial and Applied Mathematics (SIAM)
- Member, Royal Statistical Society
- Member, IEEE Signal Processing Society
- Member, British Data Science Society (Data Science Connect)
- Member, Academy of Data Scientists and Analysts
Research Grants Information
- University of London Scholarship (MSc)
- EPSRC Doctoral Grant
Research Honours and Awards
King’s College, University of London: Award for Excellence in Teaching Engineering Mathematics
Current Teaching
- ST2133: Advanced Statistics – Mathematical Statistics (Distribution Theory and Methods)
- ST2134: Advanced Statistics – Mathematical Statistics (Estimation Methods and Inference)
- ST2195: Python, R Programming; SQL and Machine Learning for Data Science
- EC2020: Elements of Econometrics – Panel, Cross-Section, and Time Series Models
- CS 210: Web and Mobile Applications (HTML, CSS, JavaScript, React, React Native; Android and iOS Platforms)
