Training Mixed-Objective Pointing Decoders for Block-Level Optimization in Search Recommendation

TL;DR

This paper introduces a novel architecture for search suggestion ranking that considers all candidates simultaneously and employs mixed-objective training to optimize multiple metrics, improving block-level suggestion quality.

Contribution

The paper proposes a block-level suggestion architecture that processes all candidates together and enforces multiple metrics through mixed-objective training, advancing search recommendation methods.

Findings

Improved suggestion block quality over traditional pairwise models.

Effective enforcement of multiple ranking metrics during training.

Demonstrated benefits of joint candidate processing in search suggestions.

Abstract

Related or ideal follow-up suggestions to a web query in search engines are often optimized based on several different parameters -- relevance to the original query, diversity, click probability etc. One or many rankers may be trained to score each suggestion from a candidate pool based on these factors. These scorers are usually pairwise classification tasks where each training example consists of a user query and a single suggestion from the list of candidates. We propose an architecture that takes all candidate suggestions associated with a given query and outputs a suggestion block. We discuss the benefits of such an architecture over traditional approaches and experiment with further enforcing each individual metric through mixed-objective training.