SemanticZ at SemEval-2016 Task 3: Ranking Relevant Answers in Community Question Answering Using Semantic Similarity Based on Fine-tuned Word Embeddings

TL;DR

This paper presents a system that uses fine-tuned word embeddings and topic similarities to rank relevant answers in community question answering, achieving top-three results in SemEval-2016 tasks.

Contribution

The paper introduces a semantic similarity approach based on fine-tuned embeddings for answer ranking in community forums, demonstrating competitive performance.

Findings

Achieved third place in Subtask C with MAP of 51.68

Achieved third place in Subtask A with MAP of 77.58

System outperforms baseline methods in answer ranking

Abstract

We describe our system for finding good answers in a community forum, as defined in SemEval-2016, Task 3 on Community Question Answering. Our approach relies on several semantic similarity features based on fine-tuned word embeddings and topics similarities. In the main Subtask C, our primary submission was ranked third, with a MAP of 51.68 and accuracy of 69.94. In Subtask A, our primary submission was also third, with MAP of 77.58 and accuracy of 73.39.