Learning from the latest advances in artificial intelligence

Manuel’s experience at the 11th International Conference on Hybrid Artificial Intelligence Systems

Published on 28th April 2016 by Manuel Martin Salvador

You know you have arrived in Seville, Spain when the taxi is playing “sevillanas” – a traditional music style from the region. This year, the 11th International Conference on Hybrid Artificial Intelligence Systems was held at the Universidad Pablo de Olavide in Seville. The work I developed as part of my PhD, together with Marcin Budka and Professor Bogdan Gabrys of Bournemouth University, “Towards Automatic Composition of Multicomponent Predictive Systems” was accepted to the conference, including an oral presentation at the Data Mining and Knowledge Discovery session.

Our work aims to speed up the process of creating predictive models, from several days or even weeks, to a few hours. To this end, we have contributed towards the theory behind, and development of, a tool called Auto-WEKA that can be used as a blackbox for people without any data mining knowledge. We have already used this tool in our recent experiments with gaze detection and we plan to use it for fraud detection in the near future.

More than 60 academic papers from different universities and institutions were presented. Although there were excellent works in pure research, the talks in applied research attracted more attention. The identification of skin cancer using computer vision techniques – analysing pictures of melanomas – is just one example of how artificial intelligence can help to save lives. This is something we’re very interested in at Base, with an emphasis on prevention rather than cure.

IMG-20160420-WA0012

Electric vehicles has been the hot topic recently. A couple of talks focused on fault detection in batteries for vehicles, meaning improvements in reliability. This is hugely important for this technology to become widely accepted among consumers.

In the decades since the first DNA sequencing in the 1970s, bioinformatics has grown exponentially. A talk I would like to highlight was one by Sánchez Medina about using genetic algorithms to discover genes that are highly related with cancer. Nowadays there are thousands of researchers working in this field and it is one of the most exciting areas as genome sequencing has become very cheap.

Robotics is another field that is benefiting from the latest advances in Deep Learning. For example, Google is working on an autonomous learning project in which robots are learning by themselves how to pick up things through trial and error. Rodrigo Salgado presented a similar project where a robot is able to achieve a given goal through motivation, getting rewarded for success! (see video below).

Here at Base we have been continuing to experiment with indoor positioning using Beacons, identifying where people are inside a building using Bluetooth. At the conference, Marcin Luckner presented an algorithm for detecting which floor people are on, using only the network signal of their mobile phones. This approach could be useful to find people in the event of a fire since there is no additional installation needed that may depend on electric power.

Many more topics showing the latest advances in artificial intelligence made the conference really interesting. Health, bioinformatics, robotics, indoor positioning and weather forecasting are just some of the application fields where artificial intelligence techniques make a great impact. At Base, transport and smart cities are areas in which we are applying machine learning to improve a user’s experience.
Attending this conference and meeting new colleagues makes me realise that even my small contribution is part of a larger scientific community that helps all knowledge progress further.

Published on 28th April 2016

Research and Development Engineer

Manuel Martin Salvador