Metraipolis › archive › dc_72f97e13cdce

Iterated Integrals as Lexical Fingerprints: Path Signature Embeddings for Lexical Similarity — arXiv:2610.41855

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.41855
License (as archived)
CC0-1.0
Archived
2026-11-02T07:42:37.632Z
SHA-256 of the archived text
04f1a23dcfd9880f8d9fb60a5fdb1fd8f1404e523f2eb9965c1ee0b259afa808
Stable URL of this page
https://metraipolis.com/documento/dc_72f97e13cdce

How to cite

Iterated Integrals as Lexical Fingerprints: Path Signature Embeddings for Lexical Similarity — arXiv:2610.41855 — arXiv (en), https://arxiv.org/abs/2610.41855 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_72f97e13cdce (Metraipolis archive dc_72f97e13cdce, sha256 04f1a23dcfd9880f8d9fb60a5fdb1fd8f1404e523f2eb9965c1ee0b259afa808)
BibTeX
@misc{dc72f97e13cdce,
  title = {Iterated Integrals as Lexical Fingerprints: Path Signature Embeddings for Lexical Similarity — arXiv:2610.41855},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.41855},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_72f97e13cdce; sha256 04f1a23dcfd9880f8d9fb60a5fdb1fd8f1404e523f2eb9965c1ee0b259afa808},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41855. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

While sentence embeddings are primarily designed for semantic generalization, their latent geometry implicitly encodes surface-level lexical properties, a relationship that remains poorly understood. Shallow pooling methods discard word order, while deep Transformer models obscure these lexical alignments behind computational complexity and uninterpretable spaces. We propose constructing sentence embeddings via the path signature of Word2Vec trajectories, leveraging the iterated integrals of rough path theory to create representations that explicitly capture sequential and lexical structure in a geometrically tractable manner. Through extensive ablations spanning Word2Vec dimensions, signature depths, and PCA compression, we correlate embedding distances against five lexical string metrics: Levenshtein, Jaro-Winkler, word Jaccard, phoneme Levenshtein, and character TF-IDF. We reveal that while average pooling almost entirely fails to capture character-level edit distances, path signatures strongly align with them, with depth 2 signatures optimally balancing character and word-level sensitivities. Furthermore, we demonstrate that the critical lexical variance is concentrated in a highly compact PCA subspace, and unlike pooling, signatures structurally preserve and quantify word order. These findings establish path signatures not as semantic generalizers, but as rigorous mathematical lenses that explicitly encode discrete lexical structures and word order, offering an interpretable alternative to opaque deep learning for lexical tracking.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.41855 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-02T07:37:15.448Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).