- Our Tools
- For Developers
Natural Language Processing (NLP) is a wide field and typically ranges from such remote areas as speech recognition to dialogue systems and to text mining. Within LTfLL, we have been focusing on a focused set of core technologies that serve positioning, the (automated or quality improved) provision of feedback, and social & informal learning.
In the application scenarios, along which we group the services, mostly combinations of techniques come into use. In several of the cases this has enabled us to use the domain of learning to validate improvements in natural language processing technology beyond the state of the art.
On this page we want to further document these achievements and mention a few interesting exploitation stories, where others are already working with these improved technologies -- to support learning or in other application areas.
The basic idea of latent semantic analysis (LSA) is, that text do have a higher order (=latent semantic) structure which, however, is obscured by word usage (e.g. through the use of synonyms or polysemy). By using conceptual indices that are derived statistically via a truncated singular value decomposition (a two-mode factor analysis) over a given document-term matrix, this variability problem can be overcome.
The lsa package for the statistical programming language and environment R can be downloaded here.
The latent semantic analysis based services and software packages are widely used through out the world by now. Here we document a selection of recent stories about where and how people are using this.
The first interview is by Benoit Borrel, technical lead at Espeo based in the area of Montreal, Canada. Benoit has implemented a semantic similarity recommender for Drupal. Fridolin Wild interviews him on how it came to that and how it works.
Here are the a selection of guest lectures, keynotes, and other talks given on analysis techniques based on LSA and its extensions.
The first on is from a guest lecture given at the Goethe University in Frankfurt, Germany. The chair of information systems engineering of the Goethe University Frankfurt Prof. Holten invited KMi's Fridolin Wild to tell the inside story of the latent semantic analysis research done in the EC-funded language technologies for lifelong learning (LTfLL) project. Background is research on competence and performance of consulting professionals done at Prof. Holten's institute -- which is very closely related to the research on monitoring of conceptual development done by KMi within the LTfLL project. The guest lecture focused on latent semantics and social interaction.
Social Network Analysis is a technology that uses graph algorithms for computing various metrics (for example, centrality, betweenes, etc.) on the networks of participants (nodes) and relations between them (arcs). These metrics may be individual, for each node-particpant or global. They indicate different facts, like the degree in which a participant is involved in collaboration, the persons through which all the communications pass (betweenes), etc.
Below you can download the related materials: