The final figures show that LangTech was attended by some 330 representatives from over 30 countries and across five continents. The actual programme featured presentations from over 70 companies from 20 nations. Most importantly, nearly two-thirds of LangTech attendees came from industry or commercial concerns.
We are naturally delighted with this result, particularly the level of geographic diversity and support from industry. These demographics naturally led to a balanced and comprehensive account of issues, business models and future opportunities for the speech and language technologies sector across the globe.
An overview of some of the plenary sessions is here on this page. For those who attended LangTech, speaker's presentations are available online . Participants should have received a separate email by now, informing of the login and password to access this information.
At the end of the conference, Joseph Mariani, Director of the ICT Department - French Ministry of Research and New Technologies, announced that LangTech 2003 will be held in Paris, France next autumn. Send all comments, suggestions and requests regarding next year's conference to .
Many thanks for all your support in helping LangTech become 'The New European Language Technology Forum'. We hope to see you in Paris for LangTech 2003!
Professor Hans Uszkoreit, LangTech Programme Chair, opened the conference by pointing out that the key current challenge to the speech and language technology sector was not so much bringing research concepts to market but dealing with the depressed business climate. There is a fairly advanced capacity to absorb innovation on the demand side, and despite current pessimism, the market is set to rebound strongly.
Other key strategic points to emerge from the LangTech programme:
A 'user centric' drive toward 'natural' communication and interfaces is widely regarded as the way forward. The EC FP6 programme appears to be addressing this explicitly in its 'multimodal' roadmap, and many of the SME 'pitches' at the event had this concept at the centre of their business model.
Many voice and multilinguality-based technologies are now mature. As more and more applications are reaching the market, this process is set to gather greater momentum. However, several groups called for more EC support for translation technology efforts.
A key catalyst for market penetration is visibility at 'board level'. Marketing of language technologies must incorporate a greater effort to reach corporate decision-makers. Business consultants may emerge as an important 'champion' for this cause.
There are usually almost no successful generic solutions in language technology - solutions have to be customised to a specific company, sector, task etc.
Language technology currently represents around 2% of the value added to software products.
Bill Dolan , Head of Natural Language Processing at Microsoft, reminded the audience that deployable language technologies have been expected 'in 5 years time' right from the beginning of machine translation (MT) in the 1950s. Yet we still, have not developed a feasible commercial model for rolling out the technologies to the mass market.
Whilst Natural Language Processing smarts are gradually being integrated into consumer software, Dolan stressed that current user interfaces are far too clumsy: going forward, computers must now adapt to users rather than the opposite model that has driven the market. Microsoft is deploying NLP in the form of behind-the-scenes grammar checkers, smart tags and other morphological analysers in consumer software products.
He also showed how high quality MT tools can learn 'automatically' from available bilingual texts in a specific domain, claiming that a single general purpose MT solution is probably not feasible. We are more likely to see thousands of specialised MT engines distributed over the web.
Professor Wolfgang Wahlster from the German research centre DFKI, focused on the use of language technologies in the mobile internet environment, maintaining that the natural interface will indeed be multimodal. Mobile based UTMS and 3G devices will eventually provide access to all communication messages, information, entertainment and web based content, creating significant opportunities for the language technology sector. After introducing the revolutionary transportable interface concept, Smartkom, Professor Wahlster concluded by stressing that multimodal interfaces increase the robustness of user interaction and lead to more intuitive and efficient dialogues.
This theme was further supported by Giovanni Varile, from the IST Intelligent Interfaces & Surfaces Unit. Through the IST programme, the EC has a vision of building a knowledge society for all with user-focused interfaces in the foreground. This is evidenced in a research budget of over 3,600 million euro for Knowledge and Interface Technologies within the IST Framework. Mr Varile identified the development of semantic-based and context-aware knowledge systems together with natural and adaptive multimodal interfaces as key EC objectives.
Guests at the LangTech evening reception on Thursday 26th September were addressed by Mr Paul Hector, representing the Information Society Division of UNESCO. Mr Hector stressed UNESCO's dedication to support measures that help preserve the right of individuals to participate in the information age through their native language.
UNESCO is particularly concerned about the pace at which minority languages are disappearing. With this in mind, Mr Hector outlined , which seeks through policy, awareness raising and the development of software applications and tools to foster the development of information content and promote equitable access within a multilingual cyberspace.
Funding Language Technology Innovation
The LangTech programme featured a dedicated venture capital session, with a panel of four venture capitalists discussing some of their recent deals and their different approach to evaluating and selecting ventures for funding. Again it was stressed that the development of a natural user interface was a key interest area.
Marcus Jochim from Deutsche Telekom VC unit, T-Venture, pointed to the comparatively low level of investment intensity and investor confidence in the current market, but suggested strong future potential for voice based services. Jochim indicated some of the key success criteria for technology VC propositions as: quality of management, status of marketplace, 'uniqueness' of technology, a flexible and open architecture, valuable business model, attractive expected ROI and potential synergies with the VC firm.
The SME 'Elevator Pitch' Competition
During the two days of LangTech, 23 SME companies from across the globe gave five minute 'elevator pitch' presentations of their corporate project with a view to attracting venture capital interest, and of course, competing for the LangTech prize!
Voted by an international jury, prizes worth a total of 3,000 euro were awarded to the three best presentations. The jury - comprising technology and investment know-how - paid particular attention to the overall impact, degree of innovation / R&D capabilities of the organisation, relevance of market scenario (size, development, competitors), company development potential (human resources), and appropriateness of investment required. With a large number of high-quality submissions, judging these entries proved to be challenging. But we are pleased to announce the following prize winners.
1st Prize (1,500 euro): Language and Computing, Belgium
2nd Prize (1,000 euro): Natural Speech Communication, Israel
3rd Prize (500 euro): The Language Technology Centre, UK
Our congratulations go to these parties, and our thanks to all the SMEs for their presentations.
A Big Thank You
The success of LangTech 2002 is by no means an accident and we would like to thank all those responsible for contributing to the event. Special thanks go to the Investitions Bank Berlin (IBB) for their Platinum Sponsorship and the local organisation team at VDI VDE-IT. Other supporting partners include EUROMAP Language Technologies, the German National Competence Center for Language Technology, DFKI - the German Research Centre for Artificial Intelligence, the European Language and Speech Network (ELSNET) and ELRA - the European Language Resources Association.
We would also like to thank all the Aculab, BaBel Technologies, Connexor, DIME Technologies, DFKI, ESTeam, Euromap Language Technologies, GALA, Global Words, Morphologic, Natlanco, Navigo Systems, Ontotext, Scansoft, SENECA, Sympalog, TC-STAR, University of Sheffield - GATE, VoiceObjects and of course the various speakers, presenters and panellists who contributed to the programme.