[Adapt] [Seminar] Knowledge Base Question Answering via Encoding of Complex Query Structures
jessieluo1991 at gmail.com
Wed Sep 12 11:47:29 CST 2018
I will give the second half of the seminar, talking about our work "ExtRA:
Extracting Prominent Review Aspects from Customer Feedback" which is
accepted by EMNLP 2018 as a long paper. In this paper, we propose a
framework, ExtRA, for extracting the most prominent aspects of a given
product type from textual reviews. ExtRA extracts the aspect candidates
from text reviews based on a data-driven approach, builds the aspect
taxonomy by leveraging WordNet, then ranks the aspect candidates using the
taxonomy, finally generates the expected prominent aspects for the product.
Extensive experiments show that ExtRA is effective and achieves the
state-of-the-art performance on a dataset consisting of different product
Luo Kangqi <luo.kangqi at qq.com> 于2018年9月12日周三 上午11:17写道：
> Hi Adapters,
> In our weekly seminar today, I'll give a talk on my paper "Knowledge Base
> Question Answering via Encoding of Complex Query Structures" accepted in
> EMNLP 2018.
> The goal of KBQA is to answer natural language questions which ask
> existing facts of some specific entities in the knowledge base. We attempt
> to solve the KBQA problem in a more complex scenario, where multiple
> entities and relations are involved in one question. In this work, we
> encode the complex query structure of a question into a uniform vector
> representation, and thus successfully capture the interactions between
> individual semantic components within a complex question. Experimental
> results on multiple KBQA datasets proved the effectiveness of our approach.
> Time: Sept 12nd, 5pm (Wed, *today*)
> Venue: Room 3-517A
> See you there!
> Kangqi Luo, PhD Candidate
> ADAPT Lab, Department of Computer Science
> SEIEE 3-341, Shanghai Jiao Tong University,
> No. 800 Dongchuan Road, Shanghai, China
> Adapt mailing list
> Adapt at cs.sjtu.edu.cn
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Adapt