Professor Ata Kaban PhD

Professor Ata Kaban

School of Computer Science

Contact details

Address
School of Computer Science
University of Birmingham
Edgbaston
Birmingham
B15 2TT
UK

Ata Kaban is Professor in Computer Science working in statistical machine learning and data mining in high dimensional settings. The two major problems of her focus are the `curse of dimensionality’ and the gap between theory and practice.

 Ata Kaban is also an EPSRC Fellow (Jan 2017- Jan 2022) with the project “FORGING: Fortuitous Geometries and Compressed Learning”, and a part-time Turing Fellow.

 For more information please visit Ata's Computer Science page.

Qualifications

  • PhD in Computer Science 2002

  • PhD in Musicology 2000

  • BSc (Hons) in Computer Science 1999

  • MSc in Musicology 1994

  • BA in Musical Composition 1993

Biography

Ata Kaban obtained a BSc in Computer Science from the Babes-Bolyai University of Cluj-Napoca, Romania, alongside of finishing a PhD in Musicology. She went on to pursue a PhD in Computer Science in Scotland, at the University of Paisley, under the supervision of Professor Mark Girolami. Upon completing, she briefly took an assistant professor position at the Eotvos Lorand University of Budapest, Hungary, before joining the University of Birmingham as a lecturer in 2003. She has been working in Birmingham since then, from 2018 as a Professor.

Postgraduate supervision

  • Statistical machine learning - theory and practice

  • High-dimensional data spaces, distance concentration

  • Probabilistic modelling of data, Bayesian inference

  • Large scale black-box, optimisation

  • Dimensionality reduction, random projections

  • Compressive learning, compressive optimisation 

Research

Professor Kaban’s research contributed to the theory and practice of statistical machine learning, data mining, pattern recognition, as well as to evolutionary black-box optimisation. Her main focus has been to explain, test and resolve computational, statistical, inferential, geometric, and interpretational problems associated with the ‘curse of dimensionality’ in these areas. Her current work (supported by a 5 years EPSRC Fellowship) develops theory for high dimensional data analytics through compressive learning, to provide better risk guarantees and new algorithms that exploit naturally occurring structures in the high dimensional learning problems.

 In previous years, she had several fruitful inter-disciplinary collaborations where she developed novel machine learning algorithms to analyse data from Palaeontology (through a visit to the University of Helsinki), Astrophysics (as a Co-I on a PPARC-funded project), and Biology (as an MRC Discipline Hopping Award recipient).

Publications

Recent publications

Article

Reeve, H & Kaban, A 2019, 'Robust randomised optimisation with k nearest neighbours', Analysis and Applications, vol. 17, no. 5, pp. 819-836. https://doi.org/10.1142/S0219530519400086

Kabán, A, Bootkrajang, J & Durrant, RJ 2016, 'Toward Large-Scale Continuous EDA: A Random Matrix Theory Perspective', Evolutionary Computation, vol. 24, no. 2, pp. 255-291. https://doi.org/10.1162/EVCO_a_00150

Conference contribution

Reeve, HWJ & Kaban, A 2019, Fast Rates for a kNN Classifier Robust to Unknown Asymmetric Label Noise. in Proceedings of the Thirty-sixth International Conference on Machine Learning (ICML 2019). The Proceedings of Machine Learning Research , vol. 97, Thirty-sixth International Conference on Machine Learning (ICML 2019), Long Beach, CA, United States, 9/06/19.

Reeve, HWJ & Kaban, A 2019, Classification with unknown class-conditional label noise on non-compact feature spaces. in 32nd Annual Conference on Learning Theory (COLT 19). Proceedings of Machine Learning Research, vol. 99, Proceedings of Machine Learning Research, 32nd Annual Conference on Learning Theory (COLT 19), Phoenix, Arizona, United States, 25/06/19.

Kaban, A 2019, Compressive Learning of Multi-layer Perceptrons: An Error Analysis. in Proceedings of 2019 International Joint Conference on Neural Networks (IJCNN) . IEEE Computer Society Press, International Joint Conference on Neural Networks (IJCNN 2019), Budapest, Hungary, 14/07/19.

Reeve, HWJ & Kaban, A 2019, Exploiting geometric structure in mixture proportion estimation with generalised Blanchard-Lee-Scott estimators. in 30th International Conference on Algorithmic Learning Theory (ALT'19). Proceedings of Machine Learning Research, vol. 98, Proceedings of Machine Learning Research, pp. 682-699, 30th International Conference on Algorithmic Learning Theory (ALT'19), Chicago, United States, 22/03/19.

Kaban, A & Thummanusarn, Y 2018, Tighter guarantees for the compressive multi-layer perceptron. in D Fagan, C Martín-Vide, M O’Neill & M A. Vega-Rodríguez (eds), Theory and Practice of Natural Computing: 7th International Conference, TPNC 2018 Dublin, Ireland, December 12–14, 2018 Proceedings. Lecture Notes in Computer Science, Springer, pp. 388-400, 7th International Conference on the Theory and Practice of Natural Computing (TPNC 2018), Dublin, Ireland, 12/12/18. https://doi.org/10.1007/978-3-030-04070-3_30

Kaban, A 2018, Dimension-free error bounds from random projections. in Thirty Third AAAI Conference on Artificial Intelligence (AAAI-19). AAAI Press, Thirty Third AAAI Conference on Artificial Intelligence (AAAI-19), Honolulu, Hawaii, United States, 27/01/19.

Turl, A & Kaban, A 2018, Joint blind source separation and declipping: a geometric approach for time disjoint sources. in 2017 IEEE International Symposium on Signal Processing and Information Technology (ISSPIT) (2017). Institute of Electrical and Electronics Engineers (IEEE), pp. 220-225, 17th IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017, Bilbao, Spain, 18/12/17. https://doi.org/10.1109/ISSPIT.2017.8388645

Kaban, A 2017, On Compressive Ensemble Induced Regularisation: How Close is the Finite Ensemble Precision Matrix to the Infinite Ensemble? in Proceedings of 28th International Conference on Algorithmic Learning Theory (ALT 2017). Proceedings of Machine Learning Research, vol. 76, JMLR , 28th International Conference on Algorithmic Learning Theory (ALT 2017), Kyoto, Japan, 15/10/17.

Sanyang, M, Durrant, R & Kaban, A 2016, How effective is Cauchy-EDA in high dimensions? in Proceedings of the IEEE Congress on Evolutionary Computation 2016., 16557, Institute of Electrical and Electronics Engineers (IEEE), IEEE Congress on Evolutionary Computation 2016, Canada, 25/07/16. https://doi.org/10.1109/CEC.2016.7744221

Xu, Q, Sanyang, M & Kaban, A 2016, Large scale continuous EDA using mutual information. in Proceedings of the IEEE Congress on Evolutionary Computation., 16598, Institute of Electrical and Electronics Engineers (IEEE), IEEE Congress on Evolutionary Computation 2016, Canada, 25/07/16. https://doi.org/10.1109/CEC.2016.7744260

Schleif, FM, Kaban, A & Tino, P 2016, Finding small sets of random fourier features for shift-invariant kernel approximation. in Artificial Neural Networks in Pattern Recognition - 7th IAPR TC3 Workshop, ANNPR 2016, Proceedings. vol. 9896 LNAI, Lecture Notes in Computer Science, vol. 9896 LNAI, Springer Verlag, pp. 42-54, 7th IAPR TC3 Workshop on Artificial Neural Networks in Pattern Recognition, ANNPR 2016, Ulm, Germany, 28/09/16. https://doi.org/10.1007/978-3-319-46182-3_4

Kaban, A 2016, A New Look at Nearest Neighbours: Identifying Benign Input Geometries via Random Projections. in ACML 2015 Proceedings. vol. 45, Proceedings of Machine Learning Research, vol. 45, JMLR , pp. 65-80, 7th Asian Conference on Machine Learning, Hong Kong, China, 20/11/15.

Paper

Kaban, A 2016, 'Non-asymptotic analysis of compressive Fisher discriminants in terms of the effective dimension' Paper presented at 7th Asian Conference on Machine Learning, Hong Kong, China, 20/11/15 - 22/11/15, pp. 17-32.

View all publications in research portal