Analog IC placement generation via neural networks from unlabeled data. (2020)

• Record Type: Book
• Further Information: Note: António Gusmão, Nuno Horta, Nuno Lourenço, Ricardo Martins.
• Other Names: Gusmão, António; Horta, Nuno C. G; Lourenço, Nuno; Martins, Ricardo
• Contents: Intro -- Preface -- Contents -- Acronyms -- 1 Introduction -- 1.1 Mixed-Signal Systems-on-Chip Paradigm -- 1.2 Mixed-Signal Design Flow -- 1.3 Analog IC Placement Automation by Machine Learning -- 2 Artificial Neural Network Overview -- 2.1 General Architecture and Behavior -- 2.2 Optimizers -- 2.3 Batch Size -- 2.4 Backpropagation Algorithm -- 2.5 Regularization -- 2.6 Hidden Layers and Number of Neurons -- 2.7 Activation Function -- 2.8 Feature Engineering -- 2.9 Conclusions -- 3 State-of-the-Art in Analog Integrated Circuit Placement -- 3.1 Placement Constraints 3.2 Placement Representation -- 3.3 Automatic Layout Generation Approaches -- 3.3.1 Placement Optimization Considering Topological Constrains -- 3.3.2 Layout Migration and Retargeting -- 3.3.3 Layout Synthesis Through Knowledge Mining -- 3.3.4 Machine Learning Applications Towards Layout Production -- 3.4 Conclusions -- 4 ANN Models for Analog Placement Automation -- 4.1 Problem Definition and Feature Description -- 4.2 Constraint Descriptive Vector -- 4.2.1 Current Flow Encoding -- 4.2.2 Symmetry -- 4.2.3 Sizing -- 4.2.4 Zero Padding for Multiple Circuit Design -- 4.2.5 Device Scrambling 4.3 Topological Loss Function -- 4.3.1 Wasted Area and Overlap Between Devices -- 4.3.2 Summed Symmetry Axis' Deviation -- 4.3.3 Current Flow Consistency Error -- 4.4 Model Structure and Training -- 4.5 Conclusions -- 5 Results -- 5.1 MSE Models -- 5.1.1 Structure and Training -- 5.1.2 Non-polynomial/Polynomial Features -- 5.1.3 Device Scrambling -- 5.2 Topological Models -- 5.2.1 Structure and Training -- 5.2.2 General Results -- 5.2.3 TLF Components Correlation Analysis -- 5.2.4 Test Dataset Augmentation -- 5.2.5 Generation of Novel Layouts -- 5.2.6 Folded Single Stage Amplifier Circuit Test 5.3 Conclusions -- 6 Conclusions and Future Work -- 6.1 Conclusions -- 6.2 Future Work -- References
• Publisher Details: Cham : Springer
• Publication Date: 2020
• Extent: 1 online resource (96 p.)
• Subjects: 621.3815; Integrated circuits; Integrated circuits -- Design and construction; Neural networks (Computer science); Machine learning; Computers -- Intelligence (AI) & Semantics; Integrated circuits; Integrated circuits -- Design and construction; Neural networks (Computer science); Electronic books; Electronic books
• Languages: English
• ISBNs: 9783030500610; 3030500616
• Related ISBNs: 9783030500603
• Notes: Note: Includes bibliographical references.
• Access Rights: Legal Deposit; Only available on premises controlled by the deposit library and to one user at any one time; The Legal Deposit Libraries (Non-Print Works) Regulations (UK).
• Access Usage: Restricted: Printing from this resource is governed by The Legal Deposit Libraries (Non-Print Works) Regulations (UK) and UK copyright law currently in force.