Conventional wafer defect analysis relies on offline, closed-world classifiers that fail to detect novel, unseen defect signatures emerging during continuous semiconductor manufacturing. To address this, we propose WaferHDC, an online open-world wafer defect discovery framework based on hyperdimensional computing (HDC). The memory-centric architecture of WaferHDC enables rapid anomaly detection and grouping of recurring failures into compact hypervector representations. This allows the memory space to dynamically scale with new defect modes without retraining a static classifier or archiving entire raw wafer history. WaferHDC first models normal spatial distributions, then dynamically groups recurring failures into persistent defect memory without labels or predefined class counts. It outperforms the strongest online baseline by 13.3 p.p. in adjusted Rand index and 10.5 p.p. in normalized mutual information on the industry-standard WM-811K dataset. These results demonstrate the effectiveness of WaferHDC for continuous and open-world wafer defect analysis.
Figures & tables
Figure 1. Online open-world wafer defect discovery: As wafers arrive, previously unseen defects may appear, and recurring modes gradually emerge. A monitoring system detects abnormal wafers and organizes recurring patterns without labels. Defect labels are shown only for illustration.
Method
Open-world
Online
Mode discovery
WaferSegClassNet ( Nag et al., 2022 )
No
No
No
One-Class Wafer Map ( Jo and Lee, 2021 )
Yes
No
No
Deep Open-Set Recognition ( Frittoli et al., 2022 )
Yes
No
No
Iterative Cluster ( Pleli et al., 2024 )
Yes
No
Yes
Enhanced Open-Set Recognition ( Shin et al., 2025 )
Yes
No
No
DECOR ( Jothiraj et al., 2025 )
Yes
No
Yes
Table 1. Comparison of representative wafer-map analysis methods. “Open-world” handles patterns outside predefined training classes. “Online” causally processes previously observed wafers. “Mode discovery” organizes unknown wafers into distinct recurring groups.
Figure 2. Overview of WaferHDC. Each wafer is encoded as a distance-sensitive binary hypervector. Stage 1 determines whether the wafer has normal production behavior. Suspicious wafers are forwarded to Stage 2, where recurring abnormal patterns are accumulated in temporary memories and promoted to persistent defect modes after sufficient evidence.
Figure 3. Stage 1 normal-only defect screening on WM-811K. WaferHDC achieves competitive unknown-defect rejection, AUROC, and AUPRC against anomaly-detection baselines.
Figure 4. Stage 2-only online defect-mode discovery. Left: assignment coverage and clustering alignment measured by ARI and NMI. Right: the coverage–quality tradeoff in terms of assigned defect wafers and NMI.
Figure 5. Full-pipeline online open-world defect discovery. Left: defect rejection and clustering alignment for WaferHDC and representative detector–online-clustering baselines. Right: defect assignment rate versus NMI, ARI, with color indicating the Stage 1 detector and marker shape indicating the Stage 2 discovery method.
Figure 6. Normal contamination in discovered defect modes
Online anomaly detection identifies unusual events in data streams as they arrive, adapting to changing conditions while balancing detection speed, accuracy, and false alarms.
Spatial defect signatures on wafer bin maps (WBMs) trace yield loss to specific process faults, yet supervised classifiers recognize only the defect types seen during training, and one-class detectors built on a single mechanism tend to capture either local structural deviations or global distributional violations, but rarely both. This work proposes a hybrid one-class framework that couples a patch-based student-teacher detector (EfficientAD) with a denoising diffusion probabilistic model (DDPM) used for partial-diffusion reconstruction, and fuses their percentile-calibrated scores through a fixed convex combination. Trained on only 700 normal wafers from the WM-38K mixed-type dataset and evaluated on 18,658 held-out wafers, the fused detector reached an AUROC of 0.9985 and reduced misclassifications from 852 (DDPM) and 1,412 (EfficientAD) to 618, with all pairwise differences significant at p < 0.001. Beyond aggregate accuracy, the analysis shows that the gain arises from weakly overlapping errors between the two modules, yet fixed-weight fusion recovers only 40-70% of the correction available to an oracle selector. Under the benchmark's inverted class balance, average precision and F1 saturate, while the Matthews correlation coefficient and negative predictive value expose unreliable normal predictions. Pixel-level maps further show that strong image-level separability does not imply spatial localization, and the diffusion module succeeds as a local density prior rather than through global geometric reasoning. These findings motivate sample-adaptive fusion and imbalance-aware evaluation of hybrid wafer anomaly detectors.
Limon Bin Hossain, Md Sadib Rahman Ananta
Department of Industrial and Production Engineering, Bangladesh University of Engineering and Technology (BUET), Dhaka, 1000, Bangladesh
In semiconductor manufacturing, defect analysis is essential, but manual inspection cannot scale. However, existing automated inspection methods remain insufficient for root-cause analysis and process optimization. To this end, we present SePArate, a weakly supervised wafer defect segmentation method. SePArate enables pixel-level separation of patterns by leveraging only image-level annotations. It consists of a three-phase training: encoder pretraining, knowledge transfer to learn spatial cues, and training on synthetic mixed-defect data for accurate segmentation. Experiments demonstrate that SePArate outperforms the baselines.
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 labeled defects or manual feature engineering. Our approach achieves state-of-the-art performance on industrial 16nm IC test data under extreme class imbalance, offering interpretable failure localization through latent-space reconstruction residuals.
Yuxuan Yin, Chen He, Todd Jacobs +4
Department of Electrical and Computer Engineering, University of California Santa Barbara, CA, USA · Automotive Processing, NXP Semiconductors, TX, USA