EUVIMEDEUVIMED entdecken
Crossref

Adaptive Compute Efficient Learning via Conceptual-Criticality (Student Abstract)

Iñigo Parra, Mano Bharathi M, Mayank Kumar, Pushpa Kumar Balan, Priyadarsi Mishra

Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (48) · 41349-41351

DOI: 10.1609/aaai.v40i48.42265 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Iñigo Parra und 4 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The computational cost of large language models (LLMs) is a primary obstacle to sustainable deployment. Static resource allocation is inefficient, as not all inputs require the same depth of processing. We propose a framework for adaptive, compute-efficient learning via conceptual criticality, which dynamically tailors computation to the assessed difficulty of an input. A lightweight criticality prediction module es- timates conceptual complexity on a continuous scale, and this score governs the LLM’s inference pathway, selectively activating token pruning, layer skipping, and quantization. Simple inputs are processed with minimal FLOPs and la- tency, while complex inputs use the model’s full capacity to preserve accuracy. We benchmark our framework and in- troduce metrics to quantify sensitivity to input criticality and per-sample computational savings. Results demonstrate an improved accuracy-efficiency trade-off, paving the way for more resource-aware systems.

Zitieren

APA

Parra, I., M, M. B., Kumar, M., Balan, P. K., & Mishra, P. (2026). Adaptive Compute Efficient Learning via Conceptual-Criticality (Student Abstract). Proceedings of the AAAI Conference on Artificial Intelligence, 40(48), 41349-41351. https://doi.org/10.1609/aaai.v40i48.42265

Vancouver

Parra I, M MB, Kumar M, Balan PK, Mishra P. Adaptive Compute Efficient Learning via Conceptual-Criticality (Student Abstract). Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(48):41349-41351. doi:10.1609/aaai.v40i48.42265

BibTeX
@article{Parra2026,
  author = {Parra, Iñigo and M, Mano Bharathi and Kumar, Mayank and Balan, Pushpa Kumar and Mishra, Priyadarsi},
  title = {Adaptive Compute Efficient Learning via Conceptual-Criticality (Student Abstract)},
  journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
  year = {2026},
  volume = {40},
  number = {48},
  pages = {41349--41351},
  doi = {10.1609/aaai.v40i48.42265}
}
Kostenlos starten →