laplacian eigenmaps and bayesian clustering based layout ...byu/papers/...sample-slides.pdf ·...
TRANSCRIPT
![Page 1: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/1.jpg)
Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications to
Hotspot Detection and OPC
Tetsuaki Matsunawa1, Bei Yu2 and David Z. Pan3
1Toshiba Corporation2The Chinese University of Hong Kong3The University of Texas at Austin
![Page 2: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/2.jpg)
2
Outline• Background• Pattern Sampling in Physical Verification• Overall flow• Laplacian Eigenmaps• Bayesian Clustering• Applications
– Lithography Hotspot Detection– OPC (Optical Proximity Correction)
• Conclusion
![Page 3: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/3.jpg)
3
Background• Issue: Systematic method for pattern sampling is not established• Goal: Pattern sampling automation for process optimization
A A AAA BB
B B B
CC
C
CC
A B
C
Grouping
Representative patterns
1D patterns 2D patterns
Based on engineer’s knowledge
Test patterns for : Simulation model calibrationSource mask optimizationWafer verification, etc.
![Page 4: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/4.jpg)
4
Pattern SamplingInput Layout
x1 = (0, 1, 0, 1.5, …)x2 = (2, 0.5, 1, -1, …)x3 = (1, -1, 0, 0.3, …)
……
Feature extraction
dimension 1
dim
ensi
on 2
dimension 1
dim
ensi
on 2
Dimension ReductionSampling
![Page 5: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/5.jpg)
5
Pattern Sampling in Physical Verification• Key techniques: Dimension reduction and Clustering
I. W. C. Tam, et al., “Systematic Defect Identification through Layout Snippet Clustering,” ITC, 2010
II. S. Shim, et al., “Synthesis of Lithography Test Patterns through Topology-Oriented Pattern Extraction and Classification,” SPIE, 2014
III. V. Dai, et al., “Systematic Physical Verification with Topological Patterns,” SPIE, 2014
group group[I] W. C. Tam
[II] S. Shim
Examples of clustering results
Classification flow
![Page 6: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/6.jpg)
6
Open Questions• Undefined similarity
• A criterion for defining pattern similarity to evaluate essential characteristics in real layouts is unclear
• Manual parameter tuning• Most clustering algorithms require several
preliminary experiments (total number of clusters)
![Page 7: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/7.jpg)
7
Laplacian Eigenmaps and Bayesian Clustering• We develop
– An efficient feature comparison method• With nonlinear dimensionality reduction / kernel
parameter optimization
– An automated pattern sampling using Bayesian model based clustering• Without manual parameter tuning
![Page 8: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/8.jpg)
8
• Problem: Given layout data, a classification model is trained to extract representative patterns
• Goal: To classify the layout patterns into a set of classes minimizing the Bayes error
Problem formulation: Layout Pattern Sampling
𝒚 = 𝒇 𝒙
Unique pattern setPattern ID
Freq
uenc
y…
Classification model Output (y)Input (x)
Layout data
![Page 9: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/9.jpg)
9
Bayes Error (BE)
𝐵𝐸 = 'min 1 − 𝑝 𝜔/|𝑥 𝑝 𝑥 𝑑𝑥
Bayes Error = 0.02, WCS/BCS = 0.07 Bayes Error = 1.68, WCS/BCS = 0.07
Comparison between BE and Within-Class Scatter/Between-Class Scatter
• To quantify the clustering performance– Define a quality of clustering distributions based on Bayes’
theorem 𝑃 𝜔|𝑥 : conditional probability in class𝜔𝑃 𝑥 : prior probability of data 𝑥
![Page 10: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/10.jpg)
10
Overall Flow(1) Sampling phase
(2) Application phase
Model training for lHotspot detection,lMask Optimization,lProcess Simulation,lWafer Inspection, etc.
Sample Plan Application
GDSIILayout
Feature A
Feature B
Feature C
⁞
Low-dimensionalvectors A
⁞
Layout Feature Extraction
Dimensionality Reduction
Clustering
Layoutdataset A
Low-dimensionalvectors B
Low-dimensionalvectors C
Layoutdataset B
Layoutdataset C
⁞Locating
Feature Points
DRC
Ranked dataset(Feature A, B or C)
Ranking
Ranked dataset(Feature A, B or C)
Ranked dataset(Feature A, B or C)
⁞
![Page 11: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/11.jpg)
11
Feature Point Generation & Feature Extraction
0.0 0.3 0.0 0.3 0.0
0.0 0.3 0.0 0.3 0.0
0.0 0.4 0.3 0.4 0.0
0.0 0.3 0.0 0.3 0.0
0.0 0.3 0.0 0.3 0.0
Feature pointUnique pattern
TLineEnd l
GDS
Locating feature points
Feature extraction
Density based encoding Diffraction order distribution
![Page 12: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/12.jpg)
12
Why dimension reduction and Bayesian clustering?
Feature A
Feature B
Feature C
Comparable data
BayesModel
High dimension Low dimension
Dimension Reduction
Required feature comparison for optimal feature selectionØ The optimal characteristics for layout representation vary in different applications
How to compare diverse layout feature types?Ø #of dimensions differs with different types of features
Hard to achieve completely automatic clusteringØ Hypothetical parameters are required for typical clustering task
Automatic Clustering
![Page 13: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/13.jpg)
13
Laplacian Eigenmaps
Original data (3D) Linear(2D)Principal Component Analysis
Nonlinear(2D)Laplacian Eigenmaps
Comparison with linear/nonlinear algorithm
𝐿𝜓 = 𝛾𝐷𝜓
𝐷 = diag <𝑊>,>@
A
>@BC𝑊>,>@ = D
1 if𝑥> ∈𝑘𝑁𝑁 𝑥>@𝑜𝑟𝑥>@ ∈ 𝑘𝑁𝑁 𝑥>
0 otherwise
𝐿 = 𝐷 − 𝑊
Solve an eigenvalue problem:
Laplacian matrix Diagonal matrix Kernel : k-nearest neighbors
To reduce dimensions while preserving complicated structure
![Page 14: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/14.jpg)
14
Kernel Parameter Optimization• Optimization through estimating density-ratio �̂� 𝐱 = 𝐰𝚽 𝐱
between given feature vectors 𝑃 𝑥 and embedded feature vectors 𝑃′ 𝑥
maxY
∑ log 𝑤]𝜙 𝑥>_A_>BC
Subjectto∑ 𝑤]𝜙 𝑥> = 𝑛and𝑤 ≥ 0A>BC
This is convex optimization, so repeating gradient ascent and constraint satisfaction converges to global solution
�̂� 𝑥
𝑟 𝑥 =𝑃′ 𝑥𝑃 𝑥
𝑃′ 𝑥𝑃 𝑥
𝑛′: #of test samples𝑛 : #of training samples
![Page 15: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/15.jpg)
15
Bayesian Clustering
Data
k1 k2 k3 …4
𝛼 + 𝑛 − 12
𝛼 + 𝑛 − 1𝛼
𝛼 + 𝑛 − 1
Clusters
Prior probability :
x1 x2x3
x4
x5
x6
xn…x1
𝑝 𝐱|𝛼, 𝑝 𝛉 =< 𝜋/𝒩 𝜇/,𝜎/
k
/BC
mixture ratioGaussian distribution
𝑝 𝑧A = 𝑘|𝑥A,𝑧C,… , 𝑧AnC ∝𝑝 𝑥A|𝑘
𝑛/𝛼 + 𝑛 − 1 𝑘 = 1⋯𝐾
𝑝 𝑥A|𝑘ArY𝛼
𝛼 + 𝑛 − 1 𝑘 = 𝐾 + 1
Centroid Similarity
• Clustering automation without arbitrary parameter tuning• Bayesian based method: express a parameter distribution
as an infinite dimensional distribution
![Page 16: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/16.jpg)
16
Experiments• Pattern sampling
– Comparison of conventional methods• Dimensionality reduction
– Principal Component Analysis (PCA) vs. LaplacianEigenmaps (LE)
• Clustering– K-means (Km) vs. Bayesian clustering (BC)
• Applications to – Lithography Hotspot Detection– OPC
![Page 17: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/17.jpg)
17
Effectiveness of Pattern Sampling• Representative patterns could be automatically selected
Clustering results: #of extracted patterns
Misclassification error rate:Bayes Error
■PCA+Km ■LE+Km■PCA+BC ■LE+BC
#of e
xtra
cted
pat
tern
s
BE
■PCA+Km ■LE+Km■PCA+BC ■LE+BC
Ratio: ■PCA+Km: 1.0 ■LE+Km: 5.6■PCA+BC: 0.7 ■LE+BC: 0.5
![Page 18: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/18.jpg)
18
Application to Lithography Hotspot Detection• To detect hotspot in short runtime
• Experiments– Detection model training with different patterns• PCA+Km, LE+Km, PCA+BC, LE+BC• Learning algorithm is fixed to Adaptive Boosting
(AdaBoost)– Metrics: detection accuracy and false alarm
![Page 19: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/19.jpg)
19
Effectiveness of Hotspot Detection• Comparison with conventional clustering method• Result: Proposed framework achieved the best false-alarm
Detection accuracy:#correctly detected hotspots / #total
hotspots
False alarm:#correctly detected hotspots /
#falsely detected hotspots■PCA+Km ■LE+Km■PCA+BC ■LE+BC
■PCA+Km ■LE+Km■PCA+BC ■LE+BC
![Page 20: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/20.jpg)
20
Application to Regression-based OPC• To predict edge movements in short runtime
• Experiments– Prediction model training with different patterns• PCA+Km, LE+Km, PCA+BC, LE+BC• Learning algorithm is fixed to Linear regression
– Metric: RMSPE (Root Mean Square Prediction Error)
Iteration : 0 Iteration : 5
Printed imageMask imagePredictededgemovements
Regression based method Conventional model-based OPC(time consuming)
![Page 21: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/21.jpg)
21
Effectiveness of OPC regression• Proposed framework achieved the best prediction accuracy
■PCA+Km ■LE+Km■PCA+BC ■LE+BC
RM
SPE(
nm)
Ratio: ■PCA+Km: 1.0 ■LE+Km: 1.1■PCA+BC: 0.9 ■LE+BC: 0.8
Prediction accuracy:RMSPE: Root mean square
prediction error
![Page 22: Laplacian Eigenmaps and Bayesian Clustering Based Layout ...byu/papers/...Sample-slides.pdf · Laplacian Eigenmaps and Bayesian Clustering Based Layout Pattern Sampling and Its Applications](https://reader035.vdocuments.us/reader035/viewer/2022062505/5ee134c5ad6a402d666c2ac6/html5/thumbnails/22.jpg)
22
Conclusion
ØWe have introduced a new method to sample unique patterns.ØBy applying our dimension reduction technique,
dimensionality- and type-independent layout feature can be used in accordance with applications.
ØThe Bayesian clustering is able to classify layout data without manual parameter tuning.
ØThe experimental results show that our proposed method can effectively sample layout patterns that represent characteristics of whole chip layout.