- This event has passed.
October 13 - October 15
OneSource Presented COP-PILOT-Related Research at IFIP ICTSS 2026
One Source participated in the 38th International Conference on Testing Software and Systems (IFIP ICTSS 2026), where it presented the paper “Evaluation of LLM Moderators Through Multi-Judge Council Scoring”.
The paper explored approaches to evaluating the reliability, policy alignment, and explainability of decisions generated by Large Language Models (LLMs). Its findings highlighted the importance of robust evaluation methods and human oversight, particularly when multiple AI models are involved in decision-making.
This work is relevant to COP-PILOT’s development of AI-assisted service onboarding and orchestration, where trustworthy and reliable AI outputs are essential. The proposed evaluation framework offers insights that could support the validation of AI-driven processes, helping identify inconsistencies, assess decision quality, and flag cases requiring human review.
Through this contribution, OneSource supported COP-PILOT’s efforts to advance reliable, transparent, and trustworthy AI technologies across the Cloud-Edge-IoT continuum.
OneSource’s presentation is available here (Day 3 – Industry Track II: AI and LLMs in Industrial Testing)
