08@London i Machine Learning

Special Session Interactive Machine Learning with the “human-in-the-loop”

2015 International Conference on Brain Informatics & Health (BIH) > London, 01-09-2015

TUESDAY, September 1, 2015
Venue: 1 Kensington Gore, London, SW7 2AR, Royal Geographical Society, Imperial College London
Underground: Piccadilly Line (dark blue): Knightsbridge or Circle Line (yellow) and District Line (green): South Kensington
Directions from Stansted Airport (approx 60 minutes): Railway to Tottenham Hale > Victoria Line towards Brixton to Green Park > Piccadilly Line
Directions from Heathrow Airport (approx 30 minutes): Take Piccadilly Line towards Cockfosters

Session Description and Motivation:

Machine learning is the most growing field in computer science  [Jordan, M. I. & Mitchell, T. M. 2015. Machine learning: Trends, perspectives, and prospects. Science, 349, (6245), 255-260], and it is well accepted that health informatics is amongst the greatest challenges [LeCun, Y., Bengio, Y. & Hinton, G. 2015. Deep learning. Nature, 521, (7553), 436-444 ].

Sometimes we do not have big data – but complex data. Applying fully automatic machine learning methods might – when lacking training samples  result in artefacts. In this case interactive machine learning may help by putting the human-into-the-loop of the algorithms and benefit from the power of human cognition. This can be particularly beneficial when dealing with NP-hard problems. Examples include subspace clustering, k-anonymization or protein folding – just to name some cases. The basis for interactive machine learning approaches can be found in three roots: reinforcement learning, preference learning and active learning. This session tries to promote some first steps in the fostering of interactive machine learning with particular emphasis on the doctor-in-the-loop to help to solve hard biomedical problems to support human health.

Keywords: Machine Learning, Knowledge Discovery, Data Mining, Health Informatics, Human-in-the-Loop, Doctor-in-the-Loop;