Project: Semantic Analysis for Unrestricted Generalized Employment

The market of job offer and job demand traversed in the last times a deep refactoring. On the one hand on-line facilities are absorbing a large share of the job market. On the other hand social media are becoming an excellent channel for recruiting and resume discovering. These two tendencies are crossed with the fact that the localization of work is becoming less crucial (distance working) and people is more and more willing to change his/her living place (European Mobility)._x000D_The potential of this phenomenon is however hampered by lack of technological support: whereas world level platforms exist for crossing job offer and job demand (e.g. Monster), the matching is still mostly monolingual and constrained by a set of structured data that the user (either a job seeker or a job provider) must fill. Moreover no business effort has been made so far to "project" the job demand/offer mechanism in the context of the semantic web / linked data paradigm._x000D_The need for innovation in the field of CV/offer matching is further stressed by the conclusions of the American scientists Peter A. Diamond and Dale T. Mortensen and the dual British-Cypriot citizen Christopher A. Pissarides (Nobel Prize 2010 in economics). Their model is furthering the understanding of how mismatch problems affect the functioning of markets – not least the labour market. A lack of symmetry between different search mechanisms and the resulting imbalance between supply and demand often thwart the closing of a business transaction. In other words, the classical assumption of “perfect markets”, consisting of supposedly fully informed participants and a balanced supply and demand, is proven wrong. Markets are rather determined by incompletely informed and hesitating suppliers and demanders (e.g. employers and employees) who watch for better alternatives. As a result, only a rough balance between supply and demand is achieved at high costs in terms of time and money. “Matching problem” is the name of the game._x000D_One might be tempted to think that the Internet, making more information available than ever, has solved that problem. However, the digital age was and still is primarily characterized by one theme: search. Thus Internet users looking up any arbitrary term in a search engine like Google are flooded with a massive amount of search results – most of them having little to no relevance to what the user was looking for. The amount of available information is both massive and unmanageable. Countless users are therefore surfing the Net for hours without finding what they are looking for. The problem of search frictions, thus, is currently more relevant than ever. That is where SAUGE/JANZZ with its principle of symmetric high-quality matchmaking, it's capability of structuring job information and projecting it into an interlinked ecosystem comes in._x000D_The SAUGE project aims at providing a technology of unrestricted CV (or resume) parsing able to capture all nuances contained in a manually written CV and to transform them into structured information. The transformation will be crucial step in that it will be based on the hybridization of most performing technologies of symbolic and statistical parsing. As a consequence, also the matching procedure will be enhanced and will allow the formulation of queries for candidate search in natural language, thus overriding the barrier of filling on line forms/questionnaire. Finally all the information will be structured according to the linked data paradigm in such a way that job oriented information (places, companies, persons, etc.) will be connected with information sources able to disclose additional information (e.g. GeoNames, Europeana, Wikipedia, etc.)_x000D__x000D_The choice of the consortium has been made in order to maximize the European dimension and collaboration of research performing SMES with academic Ps. The European dimension is necessary in order to cover multiple languages and countries, in order to increase the impact of the project. Collaboration with the academic P is necessary in order to develop certain technologies which are currently unavailable to the involved SMES. In particular CELI France (FR) will assume coordination and bring in its long lasting expertise in parsing and semantic matching, 4uGroup will innovate on ontology production, enrichment and matching in the job doCO, and UiO will provide valuable technologies in terms of semi-structured parsing, dependency parsing as well as distributional semantics._x000D__x000D_SAUGE will last 30 months and will deliver not only a technology able to assure increased competitiveness to the participating companies, but also a public prototype to be demonstrated as marketing materials._x000D__x000D_Last but not least we believe that, if successfully implemented, the technology delivered by this project could foster the whole European job market, thus improving European competitiveness and employment capacity globally.

Acronym SAUGE (Reference Number: 7846)
Duration 01/03/2013 - 31/08/2015
Project Topic The SAUGE project aims at providing a technology for unrestricted CV (or resume) parsing able to capture all nuances contained in a manually written CV and to transform them into structured information. The extracted information is then linked in the context of Linked Open Data initiative.
Project Results
(after finalisation)
The project delivered the first solutions on the market able not only to parse CVs written in a natural way by humans, but also to link several informational items within the CV to external ontologies, in such a way to improve classification, search and finally use.
Network Eurostars
Call Eurostars Cut-Off 9

Project partner

Number Name Role Country
3 4uGroup Ltd Partner Switzerland
3 Universitetet i Oslo Partner Norway