[Lisa_teatalk] Talk by Alessandro Sordoni on Friday (27 March) at 14.00 in AA3195

Kyung Hyun Cho cho.k.hyun at gmail.com
Wed Mar 25 09:49:29 EDT 2015


Dear all,

This week, Alessandro Sordoni from RALI will tell us about how a
hierarchical recurrent neural net can be used to suggest a
information-retrieval query based on a history of previous queries/returns.

Hope to see many of you there!
- K

===
Speaker: Alessandro Sordoni
Date/Time: 27 March @14.00
Place: AA3195
Title: Generating Context-Aware Query Suggestions with Hierarchical
Recurrent Neural Networks
Abstract:
Query suggestion consists in presenting related queries to users after
they issue one or several queries to a web search engine. Typically,
suggestions are computed on the basis of the most recent query only.
Taking into account the full history of previous queries as important
context for suggestion is crucial but challenging. We propose a
hierarchical generative model of query sequences that allows to
generate synthetic suggestions given an unbounded history of previous
queries as context. Our approach is based on a novel neural network
approach that generates two levels of sequences, each with its own
recurrent state: sequences of words, i.e. user queries, and sequences
of queries, i.e. user sessions. Our model outperforms several
context-aware approaches in next query prediction setting.
Additionally, the model is able to generate fluent recommendations
even for long-tail queries. Our findings are corroborated by a user
study.

Short bio:
Alessandro Sordoni is a PhD student at RALI, under the direction of
Jian-Yun Nie.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://webmail.iro.umontreal.ca/pipermail/lisa_teatalk/attachments/20150325/399b9e92/attachment.html 


More information about the Lisa_teatalk mailing list