Question transversal (Question de corpus)

The co-ordinators of the TREC 2002 QA track obtained their 500 questions corpus from question logs of WWW search engines (like the MSN portal) They extracted a thousand queries that satisfied determined patterns from the millions of questions registered in the logs and then after correcting linguistic errors they

date there are no large-scale question-answer corpora available In this paper we present the 30M Factoid Question-Answer Corpus an enormous question-answer pair corpus produced by apply-ing a novel neural network architecture on the knowledge base Freebase to trans-duce facts into natural language ques-tions Theproducedquestion-answerpairs

2 2 Are corpora the answer to all research questions in linguistics? 27 2 3 Corpus annotation 29 2 4 Introducing concordances 35 2 5 A historical overview of corpus analysis tools 37 2 6 Statistics in corpus linguistics 48 2 7 Summary 53 Further reading 54 Practical activities 55 Questions for discussion 55 3 The web laws and ethics 57 3 1

3 we describe the corpus data and the linguistic indicators we made use of Section 4 provides a description and eval-uation of our multilingual system Section 5 briefly de-scribes the corpus we make available and in the final section we discuss our findings and our goals for future work 2 Background

We thus conducted a corpus study on the two particles ça and donc in French wh-questions using the contemporary part of the Frantext database (www frantext fr) This paper can be seen as a contribution to the empirical turn in the study of ellipsis (see Pranav Hardt and McCloskey (2020) for a corpus study of English sluices based