Computer Laboratory

Aliaksei Mikhailiuk

I am a PhD student at the Computer Laboratory, working under the supervision of Dr. Rafal Mantiuk. I can be contacted at: am2442[at]cam[dot]ac[dot]uk.

Research

My current research is focused on applications of Machine Learning to a problem of modelling Human Visual System. I am particularly interested in Deep Learning and Bayesian Inference. However my general interest is not constrained by these.

Supervising

I have supervised or am supervising for:


In the past I have supervised Part IB/II students and have a few ideas for Part II/MPhil projects in machine learning, multimedia experience and computer vision. If you want to discuss possible supervision, please drop me an email.

Before

Before starting a PhD I did an MPhil in Scientific Computing at Cambridge. During this time I was working under the supervision of Dr. Anita Faul on the dissertation entitled “Deep Learning and Parallelisation Applied to Seismic Data”. Before joining Cambridge I did a BEng in Computer Science and Electronics at the University of Bristol.

Publications ("GoogleScholar" )

Perez-Ortiz M., Mikhailiuk A., Zerman E., Hulusic V., Valenzise G. and Mantiuk R., 2018. "From pairwise comparisons and rating to a unified quality scale". In: Transactions on Image Processing (TIP) 2018. IEEE (Under review)

Mikhailiuk A., Perez-Ortiz M. and Mantiuk R., 2018. "Psychometric scaling of TID2013 dataset" . In: International Conference on Quality of Multimedia Experience (QoMEX) 2018. IEEE
github dataset bibtex

Mikhailiuk A. and Faul A., 2018. "Deep Learning Applied to Seismic Data Interpolation". In: European Association of Geoscientists and Engineers (EAGE) 2018. IEEE
github bibtex

Mikhailiuk A.. and Dahnoun N., 2016. "Real-time pothole detection on TMS320C6678 DSP" . International Conference on Imaging Systems and Techniques (IST), 2016. pp 123-128. IEEE
github bibtex

Blog posts and media coverage

Unsupervised deep learning for data interpolation

Confidence intervals: parametric and non-parametric resampling

Usefull links

37 Reasons why your Neural Network is not working

28 Jupyter Notebook tips, tricks, and shortcuts

A Simple Guide to the Versions of the Inception Network

The hacker's guide to uncertainty estimates

Software to plot DNN's architectures

Address

SS04, William Gates Building, 15 J J Thomson Ave, Cambridge CB3 0FD

  • © 2017 Computer Laboratory, University of Cambridge