EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

EchoTwin Privacy-Aware Personalized Email Reply Generation with Low Rank Adaptation

Thomas iskander, Mohammed Abdelwahab, Hanaa Atwa, Anasimon Samir, Carol Faried

Computational Discovery and Intelligent Systems (CDIS) · 2026

DOI: 10.66279/g71jd323 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Computational Discovery and Intelligent Systems (CDIS), 2026. Verfasst von Thomas iskander und 4 weiteren Autor:innen. Die Arbeit nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Email continues to serve as a primary medium for both professional and personal communication. Although contemporary assistance systems, including Smart Reply and Smart Compose, enhance writing efficiency, they generally do not explicitly model or adapt to an individual’s unique communication patterns. This paper presents EchoTwin, an AI-assisted email-reply generation system that produces personalized draft replies aligned with a user’s historical communication patterns. EchoTwin builds on a decoder-only Transformer large language model and applies Low-Rank Adaptation (LoRA) for parameter-efficient personalization: a lightweight adapter is trained for each user while a single base model is shared across the population, avoiding the cost of an independent model per user. The system integrates with Gmail through Google OAuth 2.0 and the Gmail API, keeping a human reviewer in the loop to approve, edit, or reject every draft before it is sent. The approach was evaluated on a curated subset of the Enron Email-Reply dataset, comprising 2,221 email-reply pairs from ten users, comparing personalized generation against the base model on 223 held-out samples using lexical similarity (ROUGE), semantic similarity, corpus-level style similarity, and stylometric analysis. The personalized adapters achieved statistically significant gains of 8.10% in ROUGE-1, 35.65% in ROUGE-L, and 5.64% in corpus-level style similarity, plus a 25.26% reduction in mean stylometric distance (Wilcoxon signed-rank test, ???? < 0.05). Sentence-BERT similarity to the reference reply decreased by 10.29%, reflecting the expected trade-off between reference-based similarity and individual style adaptation. These prototype-level results indicate that user-specific LoRA adapters can improve measured stylistic alignment in personalized email-reply generation while keeping the user in control of the final output.

Zitieren

APA

iskander, T., Abdelwahab, M., Atwa, H., Samir, A., & Faried, C. (2026). EchoTwin Privacy-Aware Personalized Email Reply Generation with Low Rank Adaptation. Computational Discovery and Intelligent Systems (CDIS). https://doi.org/10.66279/g71jd323

Vancouver

iskander T, Abdelwahab M, Atwa H, Samir A, Faried C. EchoTwin Privacy-Aware Personalized Email Reply Generation with Low Rank Adaptation. Computational Discovery and Intelligent Systems (CDIS). 2026. doi:10.66279/g71jd323

BibTeX
@article{iskander2026,
  author = {iskander, Thomas and Abdelwahab, Mohammed and Atwa, Hanaa and Samir, Anasimon and Faried, Carol},
  title = {EchoTwin Privacy-Aware Personalized Email Reply Generation with Low Rank Adaptation},
  journal = {Computational Discovery and Intelligent Systems (CDIS)},
  year = {2026},
  doi = {10.66279/g71jd323}
}
Kostenlos starten →