Search for: All records

Abstract Deep learning techniques have been increasingly applied to the natural sciences, e.g., for property prediction and optimization or material discovery. A fundamental ingredient of such approaches is the vast quantity of labeled data needed to train the model. This poses severe challenges in data-scarce settings where obtaining labels requires substantial computational or labor resources. Noting that problems in natural sciences often benefit from easily obtainable auxiliary information sources, we introduce surrogate- and invariance-boosted contrastive learning (SIB-CL), a deep learning framework which incorporates three inexpensive and easily obtainable auxiliary information sources to overcome data scarcity. Specifically, these are: abundant unlabeled data, prior knowledge of symmetries or invariances, and surrogate data obtained at near-zero cost. We demonstrate SIB-CL’s effectiveness and generality on various scientific problems, e.g., predicting the density-of-states of 2D photonic crystals and solving the 3D time-independent Schrödinger equation. SIB-CL consistently results in orders of magnitude reduction in the number of labels needed to achieve the same network accuracies.

In this study, we compared the transient self-heating behavior of a homoepitaxial β-Ga2O3 MOSFET and a GaN-on-Si HEMT using nanoparticle-assisted Raman thermometry and thermoreflectance thermal imaging. The effectiveness of bottom-side and double-side cooling schemes using a polycrystalline diamond substrate and a diamond passivation layer were studied via transient thermal modeling. Because of the low thermal diffusivity of β-Ga2O3, the use of a β-Ga2O3 composite substrate (bottom-side cooling) must be augmented by a diamond passivation layer (top-side cooling) to effectively cool the device active region under both steady-state and transient operating conditions. Without no proper cooling applied, the steady-state device-to-package thermal resistance of a homoepitaxial β-Ga2O3 MOSFET is 2.6 times higher than that for a GaN-on-Si HEMT. Replacing the substrate with polycrystalline diamond (under a 6.5 μm-thick β-Ga2O3 layer) could reduce the steady-state temperature rise by 65% compared to that for a homoepitaxial β-Ga2O3 MOSFET. However, for high frequency power switching applications beyond the ~102 kHz range, bottom-side cooling (integration with a high thermal conductivity substrate) does not improve the transient thermal response of the device. Adding a diamond passivation over layer diamond not only suppresses the steadystate temperature rise, but also drastically reduces the transient temperature rise under high frequencymore »operating conditions.« less

Poor design choices, bad coding practices, or the need to produce software quickly can stand behind technical debt. Unfortunately, manually identifying and managing technical debt gets more difficult as the software matures. Recent research offers various techniques to automate the process of detecting and managing technical debt to address these challenges. This manuscript presents a mapping study of the many aspects of technical debt that have been discovered in this field of study. This includes looking at the various forms of technical debt, as well as detection methods, the financial implications, and mitigation strategies. The findings and outcomes of this study are applicable to a wide range of software development life-cycle decisions.