AIVEX Deep Learning Research Group Presents Accepted Paper at ICML 2026

Jaewoo Park, a researcher from AIVEX's Deep Learning Research Group, presented an accepted paper during the poster session at the International Conference on Machine Learning (ICML) 2026, held at COEX in Seoul on July 8.

ICML is one of the world's most prestigious conferences on artificial intelligence and machine learning, alongside NeurIPS and ICLR. It brings together AI researchers and industry experts from around the world to share the latest advances in machine learning research.

At the conference, AIVEX presented its accepted paper, "Memory-Distilled Selection for Noise-Robust Anomaly Detection," and had the opportunity to share its research with attendees from various countries while engaging in valuable technical discussions.

The paper proposes a robust anomaly detection method that can reliably detect defects even when training data contains noisy or contaminated samples—a common challenge in real-world manufacturing environments. Because obtaining perfectly clean training data is often difficult in industrial settings, AI technologies that address this issue have significant potential for practical manufacturing applications.

During the poster session, researchers and practitioners interested in manufacturing AI and anomaly detection visited the AIVEX poster to discuss the proposed methodology, exchange technical insights, and explore its potential for real-world industrial deployment.

AIVEX remains committed to advancing core AI technologies that solve real industrial challenges and transforming its research achievements into Industrial Physical AI solutions that deliver practical value in manufacturing environments.

Moving forward, AIVEX will continue strengthening its research capabilities through active engagement with the global AI research community while developing AI technologies that create meaningful value across industrial applications.