arXiv:2605.26468v1 Announce Type: new Abstract: Latent defect screening is challenged by extremely low failure rates, high-dimensional test data, and absence of labeled anomalies. We propose the first unsupervised anomaly detection framework incorporating a Diffusion Transformer. Raw test measurements are first compressed by an autoencoder, then reshaped into a structured token sequence enriched with sinusoidal and per-device wafer-position embeddings. Anomaly scores are derived from the noise-prediction error over mid-range diffusion timesteps, enabling fast wafer-scale screening without any
The increasing complexity of integrated circuits and the demand for higher yield rates necessitate advanced anomaly detection methods, making this research timely.
This breakthrough provides a new, unsupervised method for identifying latent defects in ICs, which is critical for semiconductor manufacturing efficiency and reliability.
The ability to perform unsupervised, wafer-scale defect screening using generative diffusion models could significantly reduce waste and improve quality control in chip production.
- · Semiconductor manufacturers
- · AI hardware companies
- · Consumer electronics industry
- · Companies relying on traditional defect screening methods
- · Manufacturers with high IC failure rates
Improved semiconductor manufacturing yields and reduced costs due to enhanced defect detection.
Faster development cycles for new and complex chips, accelerating innovation across various tech sectors.
Increased global competition in semiconductor production as the barrier to entry for quality control potentially lowers or shifts.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG