Workshop on Cross-Platform Text Mining and Natural Language Processing Interoperability
At the 10th Language Resources and Evaluation Conference (LREC 2016)
Grand Hotel Bernardin Conference Center, Portorož (Slovenia) May 23rd 2016
The workshop is over, but the notes are still available to view and comment on. Thanks to all particpants for their constructive discussions and note taking.
Recent years have witnessed an upsurge in the quantity of available digital research data, offering new insights and opportunities for improved understanding. Following advances in Natural Language Processing (NLP), Text and data mining (TDM) is emerging as an invaluable tool for harnessing the power of structured and unstructured content and data. Hidden and new knowledge can be discovered by using TDM at multiple levels and in multiple dimensions. However, text mining and NLP solutions are not easy to discover and use, nor are they easy to combine for end users.
Multiple efforts are being undertaken world-wide to create TDM and NLP platforms. These platforms are targeted at specific research communities, typically researchers in a particular location, e.g. OpenMinTeD, CLARIN (Europe), ALVEO (Australia), or LAPPS (USA). All of these platforms face similar problems in the following areas: discovery of content and analytics capabilities, integration of knowledge resources, legal and licensing aspects, data representation, and analytics workflow specification and execution.
The goal of cross-platform interoperability raises many problems. At the level of content, metadata, language resources, and text annotations, we use different data representations and vocabularies. At the level of workflows, there is no uniform process model that allows platforms to smoothly interact. The licensing status of content, resources, analytics, and of the output created by a combination of such licenses is difficult to determine and there is currently no way to reliably exchange such information between platforms. User identity management is often tightly coupled to the licensing requirements and likewise an impediment for cross-platform interoperability.
Topics of interest
Workshop topics include but are not limited to:
- cross-repository discovery of content, language resources, and analytics
- uniform access to content repositories or heterogeneous data sources (content, knowledge)
- extraction of textual content from heterogeneous sources
- orchestration of analytics workflows composed from analytics from different sources
- orchestration of cross-platform analytics workflows
- linking knowledge sources and uniformly accessing them from analytics workflows
- annotation schema design best practices
- mapping and transformation between annotation schemata
- dynamic deployment of analytics to computing resources
- machine-interpretable representation of legal and licensing metadata
- policy making for TDM for an international open research environment and open access publishing