Separate and Attend in Personal Email Search

TL;DR

This paper introduces SepAttn, a neural model that separately learns from sparse and dense email features and uses attention to improve personal email search ranking performance.

Contribution

The paper proposes a novel neural ranking model, SepAttn, which separately processes sparse and dense features and combines them with attention, outperforming existing models.

Findings

SepAttn outperforms baseline models in email search quality.

Direct concatenation of features is suboptimal for neural ranking.

Separate modeling of features enhances search performance.

Abstract

In personal email search, user queries often impose different requirements on different aspects of the retrieved emails. For example, the query "my recent flight to the US" requires emails to be ranked based on both textual contents and recency of the email documents, while other queries such as "medical history" do not impose any constraints on the recency of the email. Recent deep learning-to-rank models for personal email search often directly concatenate dense numerical features (e.g., document age) with embedded sparse features (e.g., n-gram embeddings). In this paper, we first show with a set of experiments on synthetic datasets that direct concatenation of dense and sparse features does not lead to the optimal search performance of deep neural ranking models. To effectively incorporate both sparse and dense email features into personal email search ranking, we propose a novel neural model, SepAttn. SepAttn first builds two separate neural models to learn from sparse and dense features respectively, and then applies an attention mechanism at the prediction level to derive the final prediction from these two models. We conduct a comprehensive set of experiments on a large-scale email search dataset, and demonstrate that our SepAttn model consistently improves the search quality over the baseline models.