English. In this paper, we describe our experience on using current methods developed for Community Question Answering (cQA) for a commercial application focused on an Italian help desk. Our approach is based on (i) a search engine to retrieve previously answered question candidates and (ii) kernel methods applied to advanced linguistic structures to rerank the most promising candidates. We show that methods developed for cQA work well also when applied to data generated in customer service scenarios, where the user seeks for explanation about products and a database of previously answered questions is available. The experiments with our system demonstrate its suitability for an industrial scenario. Italiano. In questo articolo, descriviamo la nostra esperienza nell’usare i metodi attualmente disponibili per il Community Question Answering (cQA) in un’applicazione commerciale riguardante il servizio clienti in lingua italiana. Il nostro approccio si basa su (i) un motore di ricerca per recuperare le domande candidate precedentemente risposte e (ii) metodi kernel applicati a strutture linguistiche avanzate per riordinare i candidati più promettenti. Mostriamo che i metodi sviluppati per il cQA funzionano bene anche quando applicati ai dati generati nell’ambito dell’assistenza clienti, dove l’utente cerca informazioni riguardo a dei prodotti e una base di dati di domande precedentemente risposte è disponibile. Gli esperimenti sul nostro sistema dimostrano l’appropriatezza del suo utilizzo in uno scenario industriale.
[1]
Li Cai,et al.
Phrase-Based Translation Model for Question Retrieval in Community Question Answer Archives
,
2011,
ACL.
[2]
Christian S. Jensen,et al.
The use of categorization information in language models for question retrieval
,
2009,
CIKM.
[3]
Annalina Caputo,et al.
Overview of the EVALITA 2016 Question Answering for Frequently Asked Questions (QA4FAQ) Task
,
2016,
CLiC-it/EVALITA.
[4]
Emanuele Pianta,et al.
The TextPro Tool Suite
,
2008,
LREC.
[5]
Alessandro Moschitti,et al.
Efficient Convolution Kernels for Dependency and Constituent Syntactic Trees
,
2006,
ECML.
[6]
Kai Wang,et al.
A syntactic tree matching approach to finding similar questions in community-based qa services
,
2009,
SIGIR.
[7]
Preslav Nakov,et al.
SemEval-2017 Task 3: Community Question Answering
,
2017,
*SEMEVAL.
[8]
Alberto Barrón-Cedeño,et al.
Learning to Re-Rank Questions in Community Question Answering Using Advanced Features
,
2016,
CIKM.
[9]
Yong Yu,et al.
Searching Questions by Identifying Question Topic and Question Focus
,
2008,
ACL.