What made us love our work?

Before becoming a Foundation, K4A was PASCAL,  a distributed institute – an institute without walls that spans many countries. It worked because, despite their differences in culture, experience and research, the members of PASCAL form a community. This community was made from individuals who visit, challenge, educate, brainstorm and support each other. It is also an open community, excited to welcome new members. In this section, we hear the personal perspectives of some of the members of PASCAL, in their own words.

 

Nicolò Cesa-Bianchi – I am Professor of Computer Science, Università degli Studi di Milano. My research interests are mainly in learning theory, applications of machine learning to text analysis and relationships between learning theory and game theory. PASCAL, as a network of excellence helps you to keep in touch with co-workers in Europe, it gives you an opportunity to participate in events. It also gives you money. In this, PASCAL was different from other networks of excellence. Other networks were mainly about mobility; PASCAL is also about giving real money to fund research – a new dimension. This provides a lot more incentive to participate in PASCAL. I was contacted by John when the network was initially conceived and so was one of the founding members of PASCAL. I became a member of the steering committee and responsible for the Pump-Priming Programme and the Conference and Workshop Attendance Programme. The Pump-Priming was totally new experience for me– like running a small funding agency and so you’re on the other side of the fence – evaluating proposals, assigning funding, monitoring progress. We’re used to judging scientific quality of proposals, but this was a bit different because we also had to monitor progress. You were a peer not a project officer, but you had to politely insist that certain tasks had to be carried out. But all the participants, in the end, were a nice crowd, so there were no serious problems. I think the programme was very valuable for the network, especially at the beginning when it forced sites to team up to write proposals. In the end we calculated that half of the PASCAL sites were involved with at least one of these projects. My students have also benefited a lot from PASCAL. They could choose from a huge list of events being proposed all the time, in my field and related fields, which I pointed out to them. So it was very easy to recruit people because of the breadth. PASCAL strengthened links within a central core, but also encouraged participation by providing funding to a wider community following a special measurment system – which really worked, it was amazing. The management was really flexible – we weren’t afraid of changing roles and we were tolerant of small mistakes. The whole management group was really nice. In particular, John is a great coordinator because of his light touch. Problems get solved without you almost even noticing they were there.

Nicola Cancedda – I am area manager of the Cross-Language Technologies group, at the Xerox Research Centre Europe. My research interest is in the use of machine learning for language problems in general, in particular for problems involving crossing language barriers: machine translation, cross-language information retrieval, and related problems. My role in PASCAL was on the one hand to bring Xerox into the network, and on the other to be in charge of the hardware and software infrastructure. Since we wanted PASCAL to become a virtual institute we needed the researchers in different places to be able to collaborate as tightly as possible – possibly as if they were in the same building. This involved identifying the software and hardware to best support the situation, which resulted in the adoption of a videoconferencing tool and document management support tools, buying equipment for videolectures.net, and eventually purchasing a cluster of PCs that was made available to all the researchers. PASCAL has been an extremely valuable network. I’m now coordinating SMART – a European project that would not have existed without PASCAL, for that’s where I met the partners who came into SMART. Also, I’m much more aware of the research going on in machine learning and optimisation now than I could ever be if I was not part of PASCAL. We participated in several pump-priming projects and actually ended up hiring some of the people for a longer term or even in permanent positions. We have a couple of PhD students (one is Anastasia, also in the brochure) – they greatly benefited from PASCAL, attending conferences and visiting other groups. I think PASCAL helped a lot to form a community. I have some experience in European projects. This is one of the very few cases where I really see enthusiasm and people putting more into the network than they would get in terms of financial benefit. I think it worked very well. My role for PASCAL 2 is now to be in charge of the Harvest programme, which is the main mechanism for outreach to new companies. I’m optimistic that there is a high potential to be explored there.

Dunja Mladenic – I’m a research and project manager at the Jozef Stefan Institute, Ljubljana, Slovenia. My research is in the area of machine learning, data/text mining and semantic Web. About ten years ago I started working on text – using machine learning techniques on textual data, web data. My PhD thesis was on adjusting certain machine learning approaches to handle large amounts of textual data. We first got involved in PASCAL as a collaborating site (project partner). In the first year I was in charge of the thematic programme on text mining, learning and natural language processing. Being female I also became involved with the Gender Action Plan – something each network has to write about (at the same time I was running side projects independent of PASCAL on Women in Science).  I was also having a personal issue, for right before PASCAL started, we got the baby! So it was very appropriate: women in science, with a baby, and I still had to go to meetings, which worked very well with PASCAL. Researchers in PASCAL are very flexible, so it was very helpful being able to bring a preschool child with me to events. He attended a scientific event for the first time when he was five months old; before he was one he had already attended many project meetings! He listens to music, plays games on the computer, watches cartoons, and being here has helped him learn some English and interact with new people. PASCAL has given me a great chance to do very interesting research and collaborate with new people. The network of excellence is also very special. I was very surprised how well the network was managed – the open structure of the network, and the impression that if you want, you could be even more active than originally listed in the proposal. For example, in the last internal call there was a call for “Brokerage of Expertise” and four people from our institute benefited, which was a very good experience for them. PASCAL was also exceptional for the events. PASCAL had a number of workshops and challenges, of very good quality, so attending these was very valuable, especially for meeting many enthusiastic people. It’s always good to be surrounded by enthusiastic people, and when their interests overlap with yours it’s even better. What was really fun in PASCAL was the dynamic that some of the partners who initially played a lesser role had a chance to show what they knew, to establish their credibility and travel – something our site benefited from. Through PASCAL we were able to host several established professors who came to visit us in Slovenia for several weeks, giving talks, performing joint work. In addition our students visited other sites, and students from other sites visited us, which was good. This was only possible because the funding had not been distributed to all the separate sites in advance, unlike in other networks. We’ve also learned lessons, particularly about the financial management of things like visits – in PASCAL 2 we will make sure each site has a financial secretary to record the details, rather than a researcher like myself doing the job. I think PASCAL is a great example of a network of excellence – I don’t know of any other network of excellence that set such an example. Actually some of the other networks applying now for a new round of funding are following the example of PASCAL in many respects.

David R. Hardoon – In my perspective, the PASCAL Network of Excellence manifests the idea of ‘connection’ in many ways. PASCAL is a system of interconnected people whose exchange goes beyond the field of research alone and has managed to create dynamic working relationships that span across the globe.  My relationship with PASCAL began when I was a second year PhD student at the University of Southampton. At my initial introduction to PASCAL, I remember being perplexed as to what PASCAL was about. My first, and also my most memorable, interaction with PASCAL was quick to answer this curiosity. The PASCAL funded Machine Learning Summer School, in 2004, held at the Island of St. Berder in France was all about bringing different levels of expertise within the field together, creating interaction and discussion for the purpose of current and future opportunities and techniques in Machine Learning, and no doubt, any other relevant fields where applicable. The long and intense sessions of studying during the day were followed by equally long and intense sessions of partying in the evening, and sometimes, throughout the night. Concurrent to learning from experts in the field we had formed new networks, some of which would be lasting friendships within and outside of the working relationships. My journey with PASCAL has been intellectually and personally fascinating.  I am now a Research Fellow at University College London and consider myself privileged for having been able to partake of the interaction PASCAL has to offer, without which my experience and development would not have been as fruitful or enjoyable. PASCAL has introduced me to some of the most fascinating people in the field and continues to provide a setting for me to work on truly amazing projects. In my mind, the objectives that make PASCAL, like a Phoenix, can never truly end.

My name is Anastasia Krithara and I am a third year PhD student in Xerox Research Center Europe, in Grenoble, France. I am working in collaboration with the computer science laboratory (LIP6) of Pierre et Marie Curie University (Paris VI) in Paris. My research interests fall into the area of Machine Learning. In particular, I am interested in semi-supervised categorisation and active learning. I have been a PASCAL member, as a postgraduate student, since 2005. Being a PASCAL member, and thanks to the “short visiting program” it offers, I was given the possibility to work closely with my professors Massih-Reza Amini and Patrick Gallinari in Paris (University Paris VI). I visited the university 8-10 times during the last 2 years. The latter has strengthened our collaboration and has really contributed in my PhD research. PASCAL gave me also the possibility last summer, to attend the 9th Machine Learning Summer School (MLSS07), in Tübigen, Germany. It has been a great opportunity to learn about many different aspects of machine learning and broaden my knowledge in the diverse fields. In addition, I believe that the summer school has been a great environment to meet people with similar interests, and international experts of the domain. I have the opportunity to discuss with them, exchange ideas and expand my network. I found the whole summer school really inspiring and motivating for the continuation of my research. Also, thanks to PASCAL, I was given the opportunity to attend 3 workshops: In 2005,  the ICML workshops in “ROC analysis”, and “Learning from partially labeled data” in Bonn, Germany, and in 2006 the “International Workshop in Intelligent Information Access” in Helsinki.  All three of them gave me the chance to learn about new contributions in the field and also present my own research.

Nello Cristianini – I am Professor of Artificial Intelligence at University of Bristol. I’m interested in all aspects of learning and computers: statistical, algorithmic, implementational, and their application to problems mostly in the fields of genomics and the Web. I am also becoming more interested in cognitive systems, intelligent systems and producing intelligent behaviour in general. I was not part of Pascal from its start as at the time I was working in the US. My first involvement with PASCAL was organising a workshop in Italy with John Shawe-Taylor. In 2006 I came back to the UK and joined the consortium, taking over the role of dissemination and public relations (the previous person left). I benefited through PASCAL funding workshops, meetings, hosting visitors and summer schools. I was already very closely involved with many of the researchers, but my students certainly benefited. I believe PASCAL is almost like a trademark – being involved in PASCAL is like an indicator of quality.  Perhaps the main benefit is that if you have a good idea, you don’t have to worry about funding, you can concentrate on the work. PASCAL provided a good, non-bureaucratic way of getting things done. In our community I don’t feel we should have any envy of America or Japan. We are leading, we contribute a lot of ideas, and this is the result of organisation, as well as talent. PASCAL (and its predecessor EU Networks) have played a central role in fostering our research community. PASCAL members are constantly in touch, not only by email but also by regular videoconferencing. We also attend the same lectures, watch the same videolectures, So we have a common, shared base of knowledge which makes it very easy to interact when we do meet. We are constantly thinking, working and acting as if we are part of the same group. I think it is a model for many activities, other than research – we really work together! This is the recipe: you take some clever people, you give them just enough money to work together. These are people who like to work. You give them difficult problems. And because of this, a research community forms, naturally self-organising. One person is in a company, another is a student, another is a professor, and they all work together and get things done. In many ways the next generation of machine learning has come out of PASCAL.

Ron Begleiter – The importance of being part of a research community, its diversity, and openness are important to a yet to be matured researcher as myself. Thus, I was extremely lucky to participate in the Machine Learning Summer School (MLSS) that was held in Tübigen in the summer of 2007. MLSS 2007 gathered approximately 100 researchers and research students from North and South America, Europe, the Middle East, South Africa, and Far East Asia. I was overwhelmed by the diversity of interests, we had a (research) psychologist who applies Machine Learning methodologies in his lab, vision and text classification practitioners, Neuroscience researchers, a guy who applies machine learning to Geodesic problems, Machine Learning foundations theorists, industry peers, and more. The corresponding teaching classes did not disappoint. The summer school covered theory and empirical issues, research paths rooted in subjective probability view of point (Bayesian) and corresponding paths rooted in objective probability (Frequencies); very fundamental material as well as state of the art ideas. Each lecture revealed a new connection between the different paradigms, and gave us a peek of the big picture. I came to  MLSS with some confidence that my research theme is itself the definition of Machine Learning, and the paradigms I use (my research toolbox) are the natural ones. However, I left with an understanding that there is not such a single Machine Learning definition, I enriched my (research) toolbox, and gained many new multidisciplinary ideas. Aside from lectures, the summer school gave me several precious moments to cherish: discussing a mixture of politics, philosophy, and research ideas over a Doner kebab with a Turkish peer (and now a friend); exchanging ideas over a bottle of cold beer with famous machine learning researchers; and sharing my research confusions with a new friend on a sunny Tübigenian afternoon. All these little moments, lectures, and deep research materials are gestalted together. I know it has made me a better research student, enhanced my research palette, connected me with the (ML) community, and made me more confident in the path I walk. I deeply thank PASCAL, MPI Tubingen, and others who organised and supported the summer school, giving me this unique opportunity.

I first became interested in neural networks in about 1987. The idea that machines could learn was intriguing. But it seemed that there was a need for a more rigorous analysis. Aspects of that led me to computational learning theory (COLT) which was just getting off the ground at that point. I founded a kind of European COLT (EuroColt) in 1993 with Martin Anthony. Then in 1994 I coordinated NeuroColt 1 , followed by Neurocolt 2. Most recently, PASCAL began in 2004. Through these I tried to understand some of the more successful learning algorithms, particularly support vector machines. Over time I gradually moved back into applications, from just theoretical work; taking the ideas and seeing if they could be applied. I now have a bit of a mix, probably slightly more on the application side. I think coordinating PASCAL is enjoyable partly because I try to do it with a minimum of fuss, which I prefer. You can make these jobs as complicated as you want. The other thing is that for a network to succeed you can’t tell people what to do. That is the first step to failure. So what you really have to do is enable – give them the environment in which they can do what they want to do anyway, but you’re encouraging them. I also think it’s been enjoyable because we’ve had an incredibly good group of people in Europe. I look around and everywhere there are really strong people working in this area. It’s a very nice convergence of talent that you can see in the theory and practical sides that has meant that you feel it’s worthwhile – stuff is coming out that’s really good. And also they really are very decent people: they help, they play ball, they do what needs to be done. My previous experience made me more confident when helping to plan PASCAL; I was drawing on a lot of the experience of seeing how a network can run. You can’t be too laissez faire, you have to build in feedback mechanisms that ensure people want to get involved. They’re not going to do something unless there’s something in it for them, so you have to set it up in way that it is a win-win situation. I also think it’s important to keep a theme of well-founded rigorous research – a quality line that you have to try to keep well-defined, so people feel that this is worth being a part of. The key to the success we’ve had is setting up the right kind of cooperative way of working. People feel that there’s a basic sense of trust. You don’t have people continually saying, “what about that money, why wasn’t it spent here?” There’s almost no dissent – which is extraordinary. I think that comes down to trust that’s been built up through these earlier projects. Trust is something we have to be very careful to nurture and we must behave in a way that is worthy of it. That has been enormously helpful in making the network run smoothly. It’s also enabled us to implement some funding mechanisms that on the surface looked like they might create some conflict – like every year each site gets an allocation of money dependent on how much activity they generated. That looks like a recipe for disaster, but it worked amazingly well. We have a mechanism for taking on board comments, so people genuinely sense that if they put effort in and do something then they will be rewarded, and that there’s a fair system to assess it. That encourages people to get involved. And because activity is linked to promoting their work (for example in workshops, paper repository, videolectures) people see it as a win-win; once the momentum is going people are keen to be involved because they see it as something that is giving them exposure. PASCAL is a research environment in which you’re going to learn a lot of new things. You see new work and ideas emerge, which is exciting. In future I think we’re going to see more and more larger-scale learning systems coming out. We’re beginning to have the potential to impact all kinds of systems in the wider community. For me PASCAL has succeeded because it has enabled a shared creativity to develop and flourish. It’s a tribute to the quality and character of the scientists across Europe and the spirit which they have brought to the PASCAL network. This spirit also helped make Steering Committee meetings very constructive; decisions were made with little fuss and often good humour. Perhaps my biggest sense of satisfaction in helping to set up PASCAL came when new workshops and events were created that we had not even considered let alone proposed. PASCAL had flown the nest and was showing a life of its own! It’s been very fulfilling to see the network succeed. You can point to the articles in this brochure for concrete examples of the success of PASCAL, but I also think that the success of PASCAL is the fact that it has got people to work together in Europe. I think we’re giving the rest of the world a run for its money.