CLASCO Presents Three Ai-Driven Contributions at AI MSE 2025 Conference

11/26/2025

At the 2^nd Conference on Artificial Intelligence in Materials Science and Engineering (AI MSE 2025), held on 18–19 November 2025 in Bochum and online, CLASCO presented three contributions demonstrating the project’s central role in advancing data-driven monitoring for laser-based surface structuring.

Contributions:

Marcelo Sallese. Sensor Signal Processing Using Machine Learning for Reliable Surface Quality in Direct Laser Interference Patterning. This study demonstrated that raw photodiode signals combined with laser parameters enable the detection of structural deviations with an accuracy of 86 percent using a dual-input convolutional neural network.

Marcelo Sallese. Inverse Modelling of Laser Processes. Predicting Process Parameters and Surface Roughness from In-Situ Sensor Signals. The work showed that roughness metrics and DLIP process parameters can be inferred directly from photodiode signal data, supporting reliable real-time quality monitoring.

Wei Wang. Autoencoder-Based Anomaly Detection in Real-Time Monitoring of Direct Laser Interference Patterning Using Plasma Signals. This contribution presented an autoencoder approach that distinguishes healthy from non-healthy process signals based on reconstruction error, achieving an AUC of 0.80.

Together, these results underline CLASCO’s contribution to enabling robust, sensor-supported and AI-enhanced process monitoring for industrial laser surface manufacturing.

Explore the picture gallery: https://dgm.de/aimse/2025/picture-gallery