BriefGPT.xyz
Jan, 2021
为黑匣子上色：通过共感觉告诉我们关于角色嵌入的事情
Coloring the Black Box: What Synesthesia Tells Us about Character Embeddings
HTML
PDF
Katharina Kann, Mauro M. Monsalve-Mercado
TL;DR
通过研究语言学和神经科学，对比不同模型的英文字符嵌入模型，发现LSTMs模型比transformers模型更符合人的认知。同时，ELMo模型和其他模型存在差异。
Abstract
In contrast to their word- or sentence-level counterparts,
character embeddings
are still poorly understood. We aim at closing this gap with an in-depth study of English
character embeddings
. For this, we use res
→