EUVIMEDEUVIMED entdecken
Crossref

Hybrid Neural-Symbolic Natural Language Understanding with Deterministic Routing for Multi-Intent Command Processing

K Hari, S D Tenith Raj, Dr. N. Devakirubai

International Journal of Engineering & Extended Technologies Research · 28.03.2026 · 08 (02)

DOI: 10.15662/ijeetr.2026.0802233 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Engineering & Extended Technologies Research, online erschienen am 28.03.2026. Verfasst von K Hari und 2 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Natural language understanding (NLU) systems used in command-driven interfaces must process both simple and multi-intent queries while maintaining high accuracy and low latency. Conventional approaches typically rely on a single model architecture, which can lead to inefficiencies when queries vary significantly in complexity. To address this limitation, this paper proposes a hybrid neural–symbolic NLU framework that integrates deterministic routing, machine learning (ML), deep learning (DL), slot extraction, and knowledge graph validation The proposed system employs a deterministic complexity-aware routing mechanism that directs low-complexity queries to a lightweight ML ensemble classifier, while complex or multi-intent queries are processed by a transformer-based DL model for deeper semantic analysis. The predicted intents are then passed to a sequence-to-structure slot extraction module based on the T5 architecture to identify structured parameters required for command execution. A symbolic knowledge graph layer further performs intent compatibility checks, dependency validation, and slot constraint verification to ensure logical consistency in the final interpretation Experimental evaluation demonstrates that the proposed hybrid architecture outperforms ML-only and DL-only baselines. The system achieves 98.0% intent accuracy and 95.9% slot accuracy while maintaining an average end-to-end latency of 333.4 ms through efficient routing. These results indicate that integrating neural learning with symbolic validation and deterministic routing provides an effective and scalable solution for multi-intent command understanding in offline intelligent agents

Zitieren

APA

Hari, K., Raj, S. D. T., & Devakirubai, D. N. (2026). Hybrid Neural-Symbolic Natural Language Understanding with Deterministic Routing for Multi-Intent Command Processing. International Journal of Engineering & Extended Technologies Research, 08(02). https://doi.org/10.15662/ijeetr.2026.0802233

Vancouver

Hari K, Raj SDT, Devakirubai DN. Hybrid Neural-Symbolic Natural Language Understanding with Deterministic Routing for Multi-Intent Command Processing. International Journal of Engineering & Extended Technologies Research. 2026;08(02). doi:10.15662/ijeetr.2026.0802233

BibTeX
@article{Hari2026,
  author = {Hari, K and Raj, S D Tenith and Devakirubai, Dr. N.},
  title = {Hybrid Neural-Symbolic Natural Language Understanding with Deterministic Routing for Multi-Intent Command Processing},
  journal = {International Journal of Engineering & Extended Technologies Research},
  year = {2026},
  volume = {08},
  number = {02},
  doi = {10.15662/ijeetr.2026.0802233}
}
Kostenlos starten →