A Unified Encoder-Decoder Framework with Entity Memory

TL;DR

This paper introduces EDMem, a unified encoder-decoder framework with an entity memory that efficiently incorporates entity knowledge into text generation tasks, outperforming previous models.

Contribution

The paper proposes EDMem, a novel entity memory-augmented encoder-decoder model pre-trained on Wikipedia for improved entity-aware text generation.

Findings

EDMem outperforms memory-based auto-encoder models.

EDMem surpasses non-memory encoder-decoder models.

The framework effectively generates entity names with constrained decoding.

Abstract

Entities, as important carriers of real-world knowledge, play a key role in many NLP tasks. We focus on incorporating entity knowledge into an encoder-decoder framework for informative text generation. Existing approaches tried to index, retrieve, and read external documents as evidence, but they suffered from a large computational overhead. In this work, we propose an encoder-decoder framework with an entity memory, namely EDMem. The entity knowledge is stored in the memory as latent representations, and the memory is pre-trained on Wikipedia along with encoder-decoder parameters. To precisely generate entity names, we design three decoding methods to constrain entity generation by linking entities in the memory. EDMem is a unified framework that can be used on various entity-intensive question answering and generation tasks. Extensive experimental results show that EDMem outperforms both memory-based auto-encoder models and non-memory encoder-decoder models.