Pedrazzini, Sandro and Keibel, Holger (2022) Integrating ML-based Classifiers into an Enterprise Search System. In: SwissText 2022, June 8-10 2022, Lugano. (Submitted)
Full text not available from this repository.Abstract
HIBU is a proprietary solution platform we use to build customer solutions around enterprise search and multilingual text analysis. Its architecture is based on two analysis components’ pipelines: a first one embeds some NLP steps, based on the detected language, used to pre-elaborate the document content; a second one contains a sequence of high-level annotators, able to analyze the text and related elements, adding further contextual information to the final enriched document. Both pipelines are based on Apache UIMA, a framework used to combine analysis components that we apply to discover information in the document considered relevant for the target application. Some examples are discovering entities in a text, such as persons, places and organizations, identifying paragraphs in a document containing confidential information, etc. The single annotators can be adapted and can be switched on and off by configuration. Moreover, the framework allows us to add new annotators based on the individual customer’s needs. In this context, we recently integrated some new ML-based analysis annotators as part of the results of an Innosuisse project carried out in collaboration with SUPSI and DSwiss (EXTRA: presented in a separate SwissText presentation, leveraging our own fine-tuned version of the pre-trained BERT model and other ML-based technologies). The wrapped functionalities allow us to provide document classification, as well as relevant and tailored information extraction, to be used by customer applications for further workflow-based functionalities. In this demo session we will show how we could wrap the new analysis functionalities into the base platform, and how these are currently integrated to further enrich the final results.
| Item Type: | Article in conference proceedings or Presentation at a conference (Speech) |
|---|---|
| Subjects: | Computer sciences > Artificial intelligence > Speech & natural language processing Computer sciences > Artificial intelligence > Machine learning |
| Department/unit: | Dipartimento tecnologie innovative |
| Depositing User: | Sandro Pedrazzini |
| Date Deposited: | 17 Jul 2023 05:32 |
| Last Modified: | 17 Jul 2023 05:43 |
| URI: | http://repository.supsi.ch/id/eprint/14148 |
Actions (login required)
 |
View Item |