Skip to yearly menu bar Skip to main content


Spotlight Poster

By Tying Embeddings You Are Assuming the Distributional Hypothesis

Bertolotti Francesco · Walter Cazzola

Hall C 4-9 #805
[ ] [ Project Page ] [ Paper PDF ]
[ Poster
Thu 25 Jul 4:30 a.m. PDT — 6 a.m. PDT

Abstract:

In this work, we analyze both theoretically and empirically the effect of tied input-output embeddings—a popular technique that reduces the model size while often improving training. Interestingly, we found that this technique is connected to Harris (1954)’s distributional hypothesis—often portrayed by the famous Firth (1957)’s quote “a word is characterized by the company it keeps”. Specifically, our findings indicate that words (or, more broadly, symbols) with similar semantics tend to be encoded in similar input embeddings, while words that appear in similar contexts are encoded in similar output embeddings (thus explaining the semantic space arising in input and output embedding of foundational language models). As a consequence of these findings, the tying of the input and output embeddings is encouraged only when the distributional hypothesis holds for the underlying data. These results also provide insight into the embeddings of foundation language models (which are known to be semantically organized). Further, we complement the theoretical findings with several experiments supporting the claims.

Chat is not available.