Issue 19 spring 2010

PDF Downloads click to download PDF Email send via email Contact EIRMA send us a message Videos
Ian McKenna, meeting chair, on ICT in R&DLinks EIRMA meeting on ICT for R&D Microsoft Research Cambridge Graphics processors used in supercomputingAmazon Web Services Syngenta EMBL nucleotide sequence database Computational science Alexander Brändle Visualisation tool IBM Research Zurich Infineum I-Nova L’Oréal I-Nova software Natasa Milic-Frayling Research Desktop Research Desktop in action Mendeley Mendeley summary Mendeley video Microsoft’s OneNote Curio for Mac Microsoft Sharepoint Mindmap of collaboration tools IBM PfizerMatthias Kaiserswerth’s blog

Modern organisations are increasingly reliant upon information and communications technologies (ICT). The R&D function is no exception, with lab chiefs using it to manage the R&D process and its interface with partners, while researchers use it as an experimental tool, as a way to keep records and share results, and as a link to their communities. Is ICT a more important enabler of R&D than it is of other business functions? A group of EIRMA members met at Microsoft’s Cambridge, UK, research labs to find out.

In silico

One obvious advantage of ICT is its role as a virtual lab bench

One obvious advantage of ICT for R&D is its role as a virtual lab bench, on which experiments that used to be made in vivo or in vitro can now be made in silico, that is, in massive simulations spread out across almost limitless computing resources.

IBM Research, for example, has a pool of servers, mainframes and supercomputers, whose resources are available over the company’s network. At the other end of the scale, when emeritus professor Graham Richard, former head of chemistry at Oxford University, wanted to explore computational drug discovery he built a cluster of Linux servers using low-cost hardware and kept it under a desk. Even the graphics processors used to power computer games are now so powerful that they are being for calculations that used to need supercomputers. And if you can’t get your own hardware, you can rent computing power and storage from companies such as online retailer Amazon, which offers its spare capacity on the open market.

The use of computing as a research tool has its issues. According to Derek Scufell, senior information domain specialist at plant biotechnology company Syngenta, as much data was deposited in the European Molecular Biology Laboratory nucleotide sequence database in the first month of 2010 as was deposited in a year, five or six years ago. It cost $10 billion to sequence the first human gene at the turn of the millennium, but the more recent sequencing of DNA co-discover James Watson’s genome cost less than $1 million. The $1000 genome sequence, and the mass of data that will accompany it, can’t be far off.

“What previously was perceived as a wave of data is becoming a tsunami,” Scuffell said.

ICT-based research faces other challenges. One of these is finding the data: if it is held within the business, do you know where it is and can you actually access it?; if it is held outside the business, can you overcome the technical, corporate and legal barriers to using it? Another is the different formats in which data is held, which can be a problem when you want to bring large amounts of data together in new ways: “It's a lot more complex than merging two [sets of financial records],” said Scuffell.

Tools are being developed to ease the flow of scientific data. Microsoft Research in Cambridge is working on various aspects of what it calls computational science. For example, Alexander Brändle, head of technology in the computational science group, demonstrated a graphical tool designed to ease the process of gathering scientific data from where it is held on a network, conditioning it for use, applying it to a simulation and then visualising the results. The drag-and-drop metaphor used on screen hides some complex data management and computing from scientific researchers who want to focus on experimentation.

Thinking about scientists thinking

This desire to do what it takes to get results is just one aspect of the way in which scientists think and approach their work. It needs to be taken into account as managers try to make appropriate ICT provision.

According to Ian McKenna, technology business relationships manager for petroleum additives company Infineum, scientists think differently to the general population. They have deep knowledge, strong cognitive skills and can easily conceptualise complex information. But they don't like to change their minds, sometimes see patterns where there are none, and can be overconfident. Being human, they may also give greater weight than they should to information because it is easily accessible, or because it comes from colleagues whom they like.

“This has implications for us in the way we provide information to scientists,” said McKenna. Concepts of working memory, which suggest that humans can only hold seven or eight concepts in mind at once, mean that abstraction is also important. So how, McKenna asks, can R&D managers use ICT to augment cognition, improve the way we think, unblock mental bottlenecks and bring an organisation’s collective knowledge to the point of need?

It’s a big issue. And it becomes even bigger when you add in the cultural change necessary to make open innovation work in science, which already collaborates strongly yet still regards having knowledge as having power. Research scientists aren’t driven by a desire to be second to publish a breakthrough. And it will take time for lab managers’ assertions that ‘those who share knowledge most freely are the most highly valued’ to be accepted by their staff.

In the meantime, ICT tools can help companies know what they know, know what they need to know, and make that knowledge available to the people who need it.

Gilles Toulemonde, CEO of I-Nova , which makes software for the front end of the innovation process, gave an example of the power of such tools for L’Oréal, a customer. A researcher at the company was wondering why the flesh of an apple goes brown when exposed to air, and did some research to find out. She wrote up her conclusions and, since they weren’t relevant to her division, they were put on the shelf. A couple of years later she moved within the company and met a colleague in her new division who needed the results she had earlier shelved. The outcome was a new product line that could mean a stronger market position and up to an extra $1 billion of revenue for the company.

The I-Nova software is designed to help capture a more shareable view of the research that a company is doing, collecting ideas, fields of interest and results and then analysing them for their relationships. Various ways of presenting the data are available, so that users can, for example, graph the projected value of a research topic against when it is due to come to fruition, to get a sense of the payback on their research pipelines. Another presentation clusters related topics, exposing hidden commonalities as well as potential gaps in the research portfolio.

Toulemonde says that this works well for internal research, but does not take into account the kinds of external inputs that are available through open innovation. I-Nova is working on links to idea-broking websites, although there is an issue of controlling how much of a company’s research is exposed through these services. Tying in problem challenges and customer-suggestion schemes can help strike a balance between control and openness.

The issue with tools is getting individuals to use them in order for the team to benefit

The issue with such tools is getting individuals to use them in order for the team to benefit. One approach is to tell researchers that ideas will only find backing if they are listed in such tools. Another is to make the tools very easy to use. The ‘social economics’ of sharing and collaboration, that is, the return that the individual gets from improving the prospects of the whole group, needs particular attention for research managers trying to move beyond using tools to simply catalogue research to using them to enable collaboration though greater sharing.

Collaboration tools

Tools to do this are emerging. Natasa Milic-Frayling, head of a research partnership program that helps researchers and Microsoft teams across Europe, Middle East and Africa, to collaborate, demonstrated one such tool – the Research Desktop . It’s a response to the changing nature of research, in which access to information has become less of a challenge than structuring and prioritising the very large amount of information that is freely available.

Research Desktop uses four main metaphors. An Activities metaphor enables researchers to gather all the documents and applications they use when performing one activity in one place. Summarisation, clustering and topic-map tools help show relationships between documents and topics.

A Library metaphor provides a unified view of research materials, wherever they are held, and can analyse references and author networks. A Notes metaphor offers an onscreen whiteboard where researchers can plan, sketch, and outline ideas. According to Milic-Frayling, this encourages people to think with the visual areas of the brain.

The Research Desktop also supports tagging, so researchers can benefit from their colleagues’ efforts to identify key aspects of a document.

There’s a video of the tool in action here.

The rise of social media and Web 2.0 means that there are ways of achieving similar things to Microsoft’s Research Desktop through free, low-cost or open-source tools.

Mendeley is one example. It tries to achieve some of what the Research Desktop does (there’s a summary here) either using a  downloadable application or through an online account. It’s free to use and backed by some of the people behind Skype, last.fm and Warner Music Group.

There’s a video about it here.

Some of the concepts in Research Desktop, like the free-form capture and structuring of information, are also available in tools such as Microsoft’s OneNote product, and Curio for the Mac. This Curio slide show gives a good impression of what such tools can do.

In some companies, the combination of the free-form journaling style of OneNote and Microsoft’s Sharepoint collaboration tool, is replacing Powerpoint slides as the main way of sharing knowledge at meetings. Users create OneNote documents that combine the relevant source material, and then use it to structure the meeting instead of a slide presentation. The OneNote document is updated during the meeting, and the result is saved on a Sharepoint server as a record.

Social networking

Public social networking is connecting hundreds of millions of people in webs of relationships that share information and evolve very rapidly. Many companies would like to use such tools to help researchers share information and connect with each other, to avoid knowledge being trapped in silos. Many such tools, free, open source or commercial, are available. The question is which to you use and how to get them used. Social networking is a bit like opening a bar or restaurant – no amount of effort on your part will make people want to use a facility they don’t like.

There’s a more detailed discussion of how IBM and Pfizer have implemented social networking, with details of the tools they have used, here.

To summarise, both companies have implemented a number of social networking and collaboration tools, and have found that some worked and some didn’t. Both have also developed strategies to tackle the social economics of sharing, trying to answer the researchers’ question “What’s in it for me?” IBM has a site that seeks solutions to problems and lists a Top Ten of contributors, ordered by the take-up of their ideas. Pfizer analyses the people who respond to its online challenges, so it can tell who are the biggest contributors in each area.

Behaviour

Senior managers need to become engaged

Companies that want their staff to take up social networking need to do more than just provide tools. Hoping for appropriate behaviour to emerge from the rank and file is not enough. Senior managers need to become engaged, to demonstrate that they take social networking seriously in practice as well as theory. Matthias Kaiserswerth, director and vice president of IBM’s Zurich Research Laboratory, for example, has a blog, in part to demonstrate that he values blogging.

Issues of freedom and control, openness and secrecy also need to be addressed if companies are going to allow a freer flow of information within and perhaps outside their organisations. Part of this is to do with educating staff about appropriate sharing: part of it is to do with educating policymakers about the reality of social networking tools. One company did both, by asking their legal department to write a policy covering contributions to wikis using an internal wiki platform. The lawyers’ initial concern about wikis was reversed once they had experienced how useful they could be.

Issues of freedom and control, openness and secrecy need to be addressed

Know what you know

Can ICT make a greater contribution to the R&D function than to other areas of a business? Perhaps yes.

The declining cost of computing power is putting a powerful experimental and analytic tool into the hands of a widening group of scientists. Meanwhile the computer’s ability to organise and analyse very large amounts of data is making it easier for scientists to keep up with the broadening frontiers of their research.

Social networking tools may also help companies ‘know what they know’, find the expertise they need from internal or external sources and, as McKenna put it, bring an organisation’s collective intelligence to the point of need. But managers need to remember that scientists think and are motivated in very particular ways that must be taken into account if such tools are to succeed. One size doesn’t fit all, and a social networking tool that isn’t used as intended is probably far better than no tool at all.

As with the shift to open innovation, the rising importance of ICT in R&D exposes a tension between freedom and control. Managers will have to adapt by demonstrating acceptable behaviours, setting sensible policies and ultimately deciding whether they trust their employees to do the right thing. And that’s got nothing to do with processing power, storage density or communications bandwidth.

[back to top]

eiq-2010-019-0002