Semantic Alignment across Ancient Egyptian Language Stages via Normalization-Aware Multitask Learning

TL;DR

This paper presents a multitask learning approach for semantic alignment across different stages of Ancient Egyptian, leveraging normalization, auxiliary views, and a shared encoder-decoder model to improve cross-stage language understanding.

Contribution

It introduces a normalization-aware multitask learning framework combining multiple tasks and auxiliary views for semantic alignment in historical languages with scarce data.

Findings

Translation tasks significantly improve alignment quality.

IPA reconstruction with KL consistency enhances cross-branch alignment.

Early fusion methods show limited benefits.

Abstract

We study word-level semantic alignment across four historical stages of Ancient Egyptian. These stages differ in script and orthography, and parallel data are scarce. We jointly train a compact encoder-decoder model with a shared byte-level tokenizer on all four stages, combining masked language modeling (MLM), translation language modeling (TLM), sequence-to-sequence translation, and part-of-speech tagging under a task-aware loss with fixed weights and uncertainty-based scaling. To reduce surface divergence we add Latin transliteration and IPA reconstruction as auxiliary views. We integrate these views through KL-based consistency and through embedding-level fusion. We evaluate alignment quality using pairwise metrics, specifically ROC-AUC and triplet accuracy, on curated Egyptian-English and intra-Egyptian cognate datasets. Translation yields the strongest gains. IPA with KL consistency improves cross-branch alignment, while early fusion demonstrates limited efficacy. Although the overall alignment remains limited, the findings provide a reproducible baseline and practical guidance for modeling historical languages under real constraints. They also show how normalization and task design shape what counts as alignment in typologically distant settings.