READERS Project

[+] | [-]
READERS Project logo CHIST-ERA logo

Readers: Evaluation And DEvelopment of Reading Systems

Demos

Intelligent FAQ

This is a demonstrator for a possible application of the technology developed by READERS project. This demonstrator takes a text as input and generates a set of questions that can be answered withe the given text. This can be used, for example, to automatically create a site of Questions and Answers.

Link: Intelligent FAQ (Beta version)

This video explains how Intelligent FAQ works and can be used:



Automatic Reading of Texts and Graphical Representation

Here is a demo page for the automatic procesing of texts up to a graphical representation: http://syn-web01.synapse-fr.com/DemoReaders/index/text.aspx

Automatic text processing using improved semantic role labelling

Here is a web service to demonstrate the text processing pipeline using our Semantic Role Labelling model. The output is a xml file in NAF machine-readable format.

Enter some text to be processed, and please allow around 30 seconds or so for the system to return output:


Query Proposition Store

Temporarily out of service

Write a couple of words and retrieve the most frequent propositions that contain them directly or in a phrase. The query language is very simple:

  • Just write the words that bind any of the arguments or the predicate. For example quarterback pass
  • Use prop or rel followed by question mark (i.e. prop? or rel?) if you want to explore just propositions (predicate structures) or relations (e.g. instance_of, has, is, etc.) respectively. You can also mark a proposition type (e.g. nvn, nvpn, nvnpn, etc.). For example: rel? quarterback, nvn? quarterback
  • Use question mark ('?') to force the retrieval of propositions with instances of a class. Classes can be the usual for Named Entities (e.g. PERSON, LOCATION, ORGANIZATION, DATE, TIME, ...), but also any common noun (e.g. quarterback, coach). For example: nvn? quarterback?
  • Use '!' if you want to force the retrieval of propositions with the common noun that might express a class. For example: nvn? quarterback!

Write a couple of words: