Joint Modeling of Search and Recommendations Via an Unified Contextual Recommender (UniCoRn)

TL;DR

This paper introduces UniCoRn, a unified deep learning model that simultaneously handles search and recommendation tasks, aiming to reduce complexity and technical debt in system development.

Contribution

The paper proposes a novel unified model that integrates search and recommendation functionalities using deep learning, streamlining system architecture.

Findings

Demonstrates improved efficiency in handling both tasks

Reduces maintenance complexity and technical debt

Shows competitive performance with specialized models

Abstract

Search and recommendation systems are essential in many services, and they are often developed separately, leading to complex maintenance and technical debt. In this paper, we present a unified deep learning model that efficiently handles key aspects of both tasks.