In this work, a deep representation learning method is proposed to build continuous-valued representations of individual integrated circuit (IC) devices. These representations are used to render mixed-variable analog circuit sizing problems as continuous ones and to apply a low-budget black box Bayesian optimization (BO) variant to solve them. By transforming the initial search spaces into continuous-valued ones, the BO’s Gaussian process models (GPs), which typically operate on real-valued spaces, can be used to guide the optimization search towards the global optimum. The proposed Device Representation Learning approach involves using device simulation data and training a composite model of a Variational Autoencoder (VAE) and a dense Neural Network. The latent variables of the trained VAE model serve as the representations of the integrated device and replace the discrete-valued parametrizations of particular devices. A thorough explanation of the proposed methodology’s mathematical formulation is given and example sizing applications on real-world analog circuits and integrated devices underline its efficiency.

Purpose: Total hip arthroplasty is among the most successful orthopedic procedures; however, approximately 10%–20% of patients require revision surgery within 15–20 years. Potential failure mechanisms at the bone–implant interface include stress shielding and excessive micromotion. This experimental work aims at examining the optimal ratio of femur length and stem length with respect to primary stability and stress shielding effect, employing a validated finite element analysis. Methods: One hundred human femora were reconstructed from computed tomography (CT) scans, and for each femur, 13 personalized cementless stems were designed with varying lengths (70–130 mm, in 5 mm increments), resulting in 1300 finite element models. Each model was analyzed under standardized joint and muscle loading conditions. The model was validated experimentally after cadaveric implantation with digital image correlation (DIC) under static loading. Primary outcomes included micromotion at the bone–implant interface and proximal stress shielding. Optimal stem length was determined as the ratio of stem‐to‐femur length that minimized the combined normalized scores of micromotion and stress shielding. The effect of stem‐to‐femur length ratio was analysed with linear regression. Results: The analysis revealed a correlation between femoral length and stem length. Longer stems provided better initial stability and reduced micromotion. On the other hand, shorter stems demonstrated less proximal stress shielding and a better stress distribution resembling that of the intact femur. The optimal stem length was found to be 18.5% ± 2.7% (mean ± SD) of the femur length in terms of the mean of calculated micromotions and stress shielding, indicating that personalized stem length selection can enhance the bone–implant osseointegration. Conclusion: The selection of a stem length that corresponds proportionally to femoral length can improve primary stability and stress shielding upon implantation, highlighting the importance of enhancing initial fixation in order to extend the lifespan of hip replacements. Level of Evidence: N/A.