Restoring Hebrew Diacritics Without a Dictionary

TL;DR

This paper introduces NAKDIMON, a two-layer character-level LSTM that can restore Hebrew diacritics without relying on curated resources, achieving comparable performance to complex systems.

Contribution

The paper presents a simple, resource-free neural model for Hebrew diacritization that matches the accuracy of more complex, resource-dependent systems.

Findings

NAKDIMON performs on par with curated systems

Diacritization achieved without human-curated resources

Effective across diverse Hebrew sources

Abstract

We demonstrate that it is feasible to diacritize Hebrew script without any human-curated resources other than plain diacritized text. We present NAKDIMON, a two-layer character level LSTM, that performs on par with much more complicated curation-dependent systems, across a diverse array of modern Hebrew sources.