BERT-Residual Quantum Language Model Inspired by ODE Multi-Step Method

Quantum-inspired language models model finer-grained semantic interactions in higher-order Hilbert spaces. However, previous methods usually capture semantic features based on context-free word vectors such as Word2Vec and GloVe. Building on natural language encoding, incorporating quantum-inspired...

Full description

Bibliographic Details

Published in:	IEEE Access
Main Authors:	Shaohui Liang, Yingkui Wang, Shuxin Chen
Format:	Article
Language:	English
Published:	IEEE 2025-01-01
Subjects:	Pre-trained models quantum language models residual connection
Online Access:	https://ieeexplore.ieee.org/document/10852213/