Accions del document

Intelligent Environment

The research group KEMLG applies Artificial Intelligence Techniques to solve real problems highly complex. Environmental processes and water quality management at the river basin are some of the pillars of their research. "Environmental Modelling and Software 2008" will be the fourth International Conference hold by the Environmental Modelling and software Society. Miquel Sànchez and Joaquim Comas (UdG) are organising such a big event this year. The conference preparation has given me the opportunity to talk to Miquel Sànchez about the intersection between IA and Environment.

Intelligence for life

Today we are interviewing Miquel Sànchez, head of the research group KEMLG. The main goal of Knowledge Engineering and Machine Learning Group (KEMLG) is the analysis, design, implementation and application of several Artificial Intelligence techniques, to support the operation or behaviour analysis of real-world complex systems or domains. The research is focussed on the analysis, design, management or supervision of these domains, such as in the health and medical field, in environmental processes and systems, and in the industrial and enterprise sector. Professor Miquel Sànchez has spent half of his life researching on environmental processes and software.

The research

"Environmental Modelling and software"

When did you start your research in environmental processes?

A long time ago... In fact, my PhD Thesis Dai-Depur: an integrated supervisory multi-level architecture for wasterwater treatment plans was related to water treatments in water purifying factories. I have been contributing to solve environmental issues for fifteen years. The research group has probably worked on it for about twenty years!.

Why environmental processes?

In my opinion, environmental processes are very interesting subjects because they are encouraging and a challenge at the same time. On one hand, as a planet Earth citizen, I am interested in social and environmental issues. On the other hand, as a researcher, I love finding theoretical problems in daily situations. I feel absolutely satisfied when this theory brings you to find new solutions and algorisms.

Everything has a beginning, what was the spark that ignites this research engine?

We have always been open to other engineers and scientist needs. Some chemist engineers from the Autonoma University contacted us many years ago. They were working on water purifying management problems and they encourage us to participate on his difficulties. At that point they were finding limitations on mathematical models application. When the system worked perfectly, mathematical models were enough. However, when they were facing an unusual situation, mathematics couldn't cope with the solution. Paradoxical, non normal situations happen regulary, therefore we thought that Artificial Intelligence techniques would be the best on this situations.

Could you explain to us the science that seems to have captivated you?

We work mainly on water purifying factories supervision and management issues. Our main goal is to be able to support those management decisions. Although it is true that in fact all these decisions could be automatized; we believed that is good to have a human being saying the last word. However, in a few years time, who knows...

We try to build integrated systems that use Artificial Intelligence. We work with:

Experts systems. Thinking Systems based on the experts knowledge, previous bibliography or earlier data. They are learning systems based on previous information.
Thinking based on cases. They use the experience as the head of the water purifying factory would do, in order to solve any problem. This technique can be useful in non typical situations where expert systems are not able to get a solution.
Extended Logic. Classical logic works in a dual system, either it is true or it is false; however we all know that reality is more complex than that and grey is the colour. Thanks to extended group theory we can fix the value of our parameters with different levels of possibility. Working this way our calculations are closer to reality and they can solve, not just ideal models, but real situations.

There is something else we have to take in consideration, too. Working on paper, we may think that every time something is not working is due to one specific mistake with a unique solution. Again, reality overcomes idealization. In the majority of cases we have to face a multiple breaking down situation where mistakes have been stored up. This represents a higher complex and challenging situation.

Rellotge

Scientists are hunters of new realities and challenges. That is why I am very pleased with my job. Getting closer to reality represents one of my main satisfactions although it is not the only one. Working with real problems spring up theoretical questions related to space and time. Working with the data temporary dependence is something very enriching and satisfying, too.

Therefore you are scientist able to work with many perspectives, theoretical and practical ones. Have you though about commercializing these techniques?

Yes, some researcher from LEQUIA research group at University of Girona and I created an intelligent drainage company called Sanejament Intel·ligent (SISLTech) that commercializes the atl_EDAR system introduced in several water purifying factories in Catalunya a while ago.

However, we could go even further with our computing architecture because we have designed it as general as possible so it could be used in other environmental issues: Intelligent Environmental decision support systems.

Environmental issues are not just water purifying factories, nor just water treatments... there are also other issues like land management or atmospheric pollution.

It is peculiar to analyze the research done all over the world because you can spot different tendencies depending on the part of the globe. On one hand Europe seems worried about water resources and management. On the other hand, countries like United States of America seem more focus on land management, atmospheric pollution or risk preventing. We all live in the same planet with same needs; however I guess each country lacks of a specific natural resource.

We are dealing with a very global subject. Are there any international meetings to talk about environmental issues and software?

Yes, sure! Inside the framework of Artificial Intelligence they organize European Conferences every two years (ECAI). Our group with Professor Ulises Cortés collaboration has been organizing workshops (BESAI) related to Artificial Intelligence applied to environmental issues for the last eight years.

However, this year we are going to miss our date because we have a bigger project in hands.

What are you preparing?

With Joaquim Comas support, I am preparing the fourth International Conference on Environmental Modelling and Software.

In 2000 the Environmental Modelling & Software Society was born in Canada. This initiative brought many scientists working on environmental issues together. From 2002 and every year from then, the environmental Modelling & Software Society has been organizing conferences every two years. In 2002 Swiss was the host country. In 2004 was Germany's turn. Two years ago was hold in United States of America and this year we are proud of being the host country for such a unique event.

When and where we will be able to take part on the conference?

The Conference will start the 7th of July and it will last until the 10th of the same month. We will be able to assist to many speeches by visiting professors, as well as to participate in several workshops. All this will be set here, in the Campus Nord of UPC.

Why are these events special?

Obviously, one of the main characterises of the Conference is that it is a multidisciplinary and international meting and this is incredible interesting. Everybody participating in it, it is someone with similar interests with very different perspective. therefore it is a unique opportunity to see what are other scientists doing, to realize about new problems and new applications, to make contacts...etc. It is an integration occasion to make science progress forward.

Could you tell us a concrete example?

Look, we were talking about commercializing our water purifying technology before. When we decided to step forward and jump into the business world we found out the difficulties of dealing with costumers. On one hand, we were really impressed by the theoretical part of the problem. On the other hand, the costumers were looking for an easy interface with easy manipulation. Through dialogue we were able to reach a compromise. That situation taught us a lot. It looks like a triviality but we discover that it is important to adapt yourself to the costumer needs.

Well... These issues are going to be some of the subjects for the conferences that will be hold here next July.

In conclusion, I would say that these high level conferences are intimately related to our research and they are unique opportunities. We hope that these three days of July inspire us enough to go on working many Julies more.

Press Contact:

ilapuente@lsi.upc.edu

Personal tools

Accions del document

Intelligent Environment