ECML PKDD 2013

Machine learning, data mining and statistical data analysis is clearly a popular area now, judging by the number of attendees of this year's European conference, ECMLPKDD 2013.

It's been a long time since I last attended (2001 for multi-label classification by a modification to C4.5). I think the field has grown and matured a lot. There are far fewer papers now showing the results of a new algorithm on 10 different UCI datasets. There is far more presence from people in industry. And industry is varied: search engines, internet shopping and finance. Yahoo, Amazon, Zalando, Deloitte and many others sponsored the conference and sent people to speak or attend. There was an "industry track", and that room was full.

Themes that I picked up on were: regression (still popular!), lots of tensors and matrices, numerical analysis methods for large data sets, network mining, sequence mining, and generally using ML/DM to influence people (buying, voting, doing good, giving your system feedback).

The organisers this year have really done a good job: working wifi, lots of food and coffee, sessions running on time, plenty of mingling time, and choosing a venue in a beautiful city, with accommodation in a wide range of hotels within easy walking distance booked as an easy part of the registration process. It is appreciated!

Diversity is something that the ECMLPKDD community have started to work on improving. It has the usual male/female imbalance of a technical conference. Perhaps slightly more women than I expected, or maybe I'm just getting used to this. I'd hazard a guess at about 20% or a bit less. But next year's organising committee are more gender-balanced, and there will also be a Diversity Chair to keep an eye on the issue.

Openness of code and data is something else the community are working on improving. This year for the first time they had an award for "Open Science", and encouraged paper submissions to include a link to code/data. In order to award this, the organisers had to download, compile, run and test lots of submitted code. I don't know which of the organisers did this onerous task, but I'm very pleased they did.

If I had to point out one thing that could still be improved, my number 1 would be that the proceedings are owned by Springer, and are not open. For reasons known only to Springer, I can't make an account with them or reactive an existing account. Maybe Springer will reply to my email eventually. But if the proceedings were open access (papers deposited at arXiv for example) then this would really benefit the ML/DM community and others, and more widely promote the work of everyone who presented.

Next year, 2014, the conference moves to Nancy, France, a city with Art Nouveau architecture, and with many fine wines. www.ecmlpkdd2014.org

Artificial Intelligence and Microscopes

Artificial Intelligence has always been a branch of Computer Science that really catches the imagination of both scientists and the public. Trying to understand and replicate intelligence in all its different forms (reasoning, creativity, decision making, planning, language, etc) is exciting because it helps us to understand ourselves. Computer scientists such as Alan Turing have been pondering the implications and possibilities of AI since the 1940s and 50s. In 1951, Marvin Minsky built the first randomly wired neural network learning machine. He had studied mathematics and biology and was trying to understand brains. He's now famous for his work in AI, but, back in the 1950s, he wasn't just a mathematician, or just a computer scientist, but also studied optics, psychology, neurophysiology, mechanics and other subjects. Perhaps we pigeonholed people less into disciplines back then? Or maybe he was just amazing. Armed with all this knowledge, and a desire to learn about the brain and to look at neurons, he invented a new type of microscope, the confocal microscope. This gets rid of unwanted scattered light so that he could really focus in detail on a very specific part of the item he was looking at. Now he could see things that had never been seen before. He built the first one, and then patented this microscope in 1961. It would be another 20 years before the idea caught on (what would the research impact monitoring committees of today make of that?). Confocal microscopes are now in every biological lab and are taken for granted.

C. elegans is a 1mm long worm which lives in the soil. It is a very simple creature, easy to grow in the lab and it has a brain. Sydney Brenner (who is 85 years old today, 13th Jan 2012) has a Nobel Prize for introducing C. elegans to biologists as a "model organism": an ideal organism for studying the principles of life. In 1986, John White and his colleagues Southgate, Thomson and Brenner published a paper on the structure of the brain of C. elegans. Each worm has just 302 neurons and this number is the same for any C. elegans worm. They worked out where all the neurons were and what their connections to other neurons were, using a confocal microscope. John White had to make substantial improvements to Minsky's microscope design in order to do this. They took 8000 pictures ("prints", because it wouldn't have been digital back then) with the microscope and annotated them all by hand.

So we now have a complete picture of a simple brain. Other scientists have taken the data from White et al.'s work and created models of the brain. We understand a lot about the behaviour of the worm and which of its 302 neurons are responsible for which behviours. We have the entire C. elegans genome, so we know how many genes it has (approx 20,000), how many cells it has (approximately 1000), and we have a technique (RNA interference) for surpressing the behaviour of any gene we want to investigate. Are we nearly there yet? Are we at that tipping point where we've inspected all there is to inspect and found nothing except complexity? Have we already understood intelligence?

Further reading/viewing:

• See the video web stories of Marvin Minsky and Sydney Brenner. During their stories, Minsky describes inventing the confocal microscope, and Brenner describes his love of computing, including how he wrote a compiler in order to better teach himself about programming.
• See the Royal Society hosted talk: "How new science is transforming the optical microscope" by Dr Brad Amos FRS (a scientist who worked with White) Monday 13th Feb 2012.
• See White and Amos's paper on the story of confocal microscopes (including a section on "Why aren't we rich?") DOI: 10.1016/S0248-4900(03)00078-9
• See Minsky's lovely memoir of how and why he came to invent this microscope.
• How does a confocal microscope work? A page with colourful hand-drawn diagrams, explaining the basics.