EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Towards Self-Evolving Agents: A Dual-Process Framework for Continual Context Refinement

Liangyu Teng, Wei Ni, Liang Song, Jun Shi, Yanfei Li

Electronics · 16.03.2026 · 15 (6) · 1232

DOI: 10.3390/electronics15061232 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Electronics, online erschienen am 16.03.2026. Verfasst von Liangyu Teng und 4 weiteren Autor:innen. Die Arbeit nennt 52 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Key Research and Development Program of China, Specific Research Fund of the Innovation Platform for Academicians of Hainan Province, Shanghai Key Research Laboratory of NSAI.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Large Language Models (LLMs) have become essential for interactive AI systems, yet they remain fundamentally static after deployment: they cannot update their parameters from interaction feedback and often repeat the same mistakes across long interaction streams. We propose Dual-Process Agent (DPA), a framework for continual context refinement that enables learning without modifying a frozen model backbone. Inspired by dual-process theory from cognitive science, DPA decomposes each interaction episode into two complementary processes: a fast System 1 that retrieves compact, relevant context from an explicit long-term memory and generates responses, and a slow System 2 that reflects on outcomes and writes curated updates back into memory. To prevent memory degradation over extended interactions, DPA maintains bulletized memory entries with utility statistics and employs a conservative curator gate that filters generic, redundant, or conflicting insertions. Experiments on six diverse benchmarks demonstrate that DPA consistently outperforms vanilla prompting and competitive baselines on both GPT-5.1 and Llama-3.1-8B backbones, achieving the best overall performance across multiple reasoning and knowledge-intensive tasks.

Förderung

  • National Key Research and Development Program of China
  • Specific Research Fund of the Innovation Platform for Academicians of Hainan Province
  • Shanghai Key Research Laboratory of NSAI

Zitieren

APA

Teng, L., Ni, W., Song, L., Shi, J., & Li, Y. (2026). Towards Self-Evolving Agents: A Dual-Process Framework for Continual Context Refinement. Electronics, 15(6), 1232. https://doi.org/10.3390/electronics15061232

Vancouver

Teng L, Ni W, Song L, Shi J, Li Y. Towards Self-Evolving Agents: A Dual-Process Framework for Continual Context Refinement. Electronics. 2026;15(6):1232. doi:10.3390/electronics15061232

BibTeX
@article{Teng2026,
  author = {Teng, Liangyu and Ni, Wei and Song, Liang and Shi, Jun and Li, Yanfei},
  title = {Towards Self-Evolving Agents: A Dual-Process Framework for Continual Context Refinement},
  journal = {Electronics},
  year = {2026},
  volume = {15},
  number = {6},
  pages = {1232},
  doi = {10.3390/electronics15061232}
}
Kostenlos starten →