Savvas Karipidis
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Savvas Karipidis has a total of 13 co-authors. Here are the top co-authors by number of shared publications:
- Thomas Noulis
Electronics Laboratory, Department of Physics, Aristotle University of Thessaloniki, 54124 Thessaloniki, Greece
3 shared publications Georg Pelz
2 shared publicationsA. Buzo
1 shared publication- Anastasios Michailidis
Electronics Laboratory, Department of Physics, Aristotle University of Thessaloniki, 54124 Thessaloniki, Greece
1 shared publication Andi Buzo
1 shared publication- Andi Buzo
Infineon Technologies AG Neubiberg, Neubiberg, Germany
1 shared publication
Latest Publications
- Open Access
- 26 August 2026
- Published by Elsevier in SSRN Electronic Journal
Generative Deep Learning Surrogate Framework for Continuous Time-Domain Waveform Synthesis of Analog Circuit Transient Response
A novel generative deep learning surrogate modeling framework is proposed for predicting the full continuous time-domain dynamic transient response of non-linear analog circuits, thus eliminating the large signal analysis from the custom design flow. The circuit vehicle used is a 12-parameter cold-start Dickson frequency-boosting charge pump in a 180 nm CMOS technology. To overcome the exponential simulation burden of high-dimensional grid sampling and the limitations of simplified closed-form equations, an automated data generation pipeline is introduced. The pipeline integrates dual-decoupled scrambled Sobol sequence sampling, SipHash duplicate filtering, and parallelized Ramer-Douglas-Peucker (RDP) waveform compression. The core surrogate model employs a hybrid deep architecture consisting of a fully connected feature-expansion stage and an 1D transposed convolutional decoder stack. Automated Bayesian hyperparameter optimization using a tree-structured Parzen estimator (TPE) identifies the optimal network structure and training dynamics. The surrogate model synthesizes full continuous transient waveforms with a mean normalized dynamic voltage error of 10.30 % (standard deviation of 4.75 %) and an inference execution time of 25 ms, providing high-fidelity transient prediction and massive design cycle acceleration.
- 26 August 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessGenerative Deep Learning Surrogate Framework for Continuous Time-Domain Waveform Synthesis of Analog Circuit Transient Response
A novel generative deep learning surrogate modeling framework is proposed for predicting the full continuous time-domain dynamic transient response of non-linear analog circuits, thus eliminating the large signal analysis from the custom design flow. The circuit vehicle used is a 12-parameter cold-start Dickson frequency-boosting charge pump in a 180 nm CMOS technology. To overcome the exponential simulation burden of high-dimensional grid sampling and the limitations of simplified closed-form equations, an automated data generation pipeline is introduced. The pipeline integrates dual-decoupled scrambled Sobol sequence sampling, SipHash duplicate filtering, and parallelized Ramer-Douglas-Peucker (RDP) waveform compression. The core surrogate model employs a hybrid deep architecture consisting of a fully connected feature-expansion stage and an 1D transposed convolutional decoder stack. Automated Bayesian hyperparameter optimization using a tree-structured Parzen estimator (TPE) identifies the optimal network structure and training dynamics. The surrogate model synthesizes full continuous transient waveforms with a mean normalized dynamic voltage error of 10.30 % (standard deviation of 4.75 %) and an inference execution time of 25 ms, providing high-fidelity transient prediction and massive design cycle acceleration.
- Open Access
- 2 February 2026
- Published by Elsevier in AEU - International Journal of Electronics and Communications
- Vol. 207
Multi-task neural metamodel and NSGA-II for ultra-low-power OTA design in tunable vowel filters
- 2 February 2026
- Published by Elsevier in AEU - International Journal of Electronics and Communications
- Vol. 207
Open AccessMulti-task neural metamodel and NSGA-II for ultra-low-power OTA design in tunable vowel filters
- Open Access
- 5 January 2026
- Published by Elsevier in Integration
- Vol. 108
Hybrid algorithm based optimization strategies for analog circuit sizing in low dropout regulators
- 5 January 2026
- Published by Elsevier in Integration
- Vol. 108
Open AccessHybrid algorithm based optimization strategies for analog circuit sizing in low dropout regulators
- Open Access
- 27 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 93869-93882
Radiation-Hardened—AI-Accelerated Custom IC Design Methodology
In this work, a radiation hardened - AI accelerated custom IC design methodology is proposed. The methodology employs reinforcement learning (RL) to optimize the IC design process, integrating radiation dosage performance degradation assessments and radiation-hardened-by-design (RHBD) MOSFET cell implementations at both schematic and layout levels. This approach streamlines the development of radiation-immune silicon products by embedding artificial intelligence to accelerate the design process and advanced radiation hardening strategies directly into the standard design flow. To validate the proposed framework, a charge-sensitive amplifier (CSA) based on a folded-cascode architecture is designed and assessed in a 180 nm standard CMOS process, demonstrating the efficiency of the methodology in radiation-resilient analog front-end systems.
- 27 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 93869-93882
Open AccessRadiation-Hardened—AI-Accelerated Custom IC Design Methodology
In this work, a radiation hardened - AI accelerated custom IC design methodology is proposed. The methodology employs reinforcement learning (RL) to optimize the IC design process, integrating radiation dosage performance degradation assessments and radiation-hardened-by-design (RHBD) MOSFET cell implementations at both schematic and layout levels. This approach streamlines the development of radiation-immune silicon products by embedding artificial intelligence to accelerate the design process and advanced radiation hardening strategies directly into the standard design flow. To validate the proposed framework, a charge-sensitive amplifier (CSA) based on a folded-cascode architecture is designed and assessed in a 180 nm standard CMOS process, demonstrating the efficiency of the methodology in radiation-resilient analog front-end systems.