BriefGPT.xyz
Sep, 2022
用于对话建模的状态性记忆增强变压器
Stateful Memory-Augmented Transformers for Dialogue Modeling
HTML
PDF
Qingyang Wu, Zhou Yu
TL;DR
本研究提出一种新的记忆增强型 Transformer 模型，该模型可在不影响对话历史信息的情况下适应长序列处理，并且在相对于其他预训练 Transformer 模型存在着更高的效率和性能。
Abstract
transformer
encoder-decoder models have shown impressive performance in
dialogue modeling
. However, as Transformers are inefficient in processing long sequences, dialogue history length often needs to be truncate
→