Showing 1–6 of 6 results for author: Honarvar, H

Semiconductor thermal and electrical properties decoupled by localized phonon resonances

Authors: Bryan T. Spann, Joel C. Weber, Matt D. Brubaker, Todd E. Harvey, Lina Yang, Hossein Honarvar, Chia-Nien Tsai, Andrew C. Treglia, M. Lee, Mahmoud I. Hussein, Kris A. Bertness

Abstract: Thermoelectric materials convert heat into electricity through thermally driven charge transport in solids, or vice versa for cooling. To be competitive with conventional energy-generation technologies, a thermoelectric material must possess the properties of both an electrical conductor and a thermal insulator. However, these properties are normally mutually exclusive because of the interconnecti… ▽ More Thermoelectric materials convert heat into electricity through thermally driven charge transport in solids, or vice versa for cooling. To be competitive with conventional energy-generation technologies, a thermoelectric material must possess the properties of both an electrical conductor and a thermal insulator. However, these properties are normally mutually exclusive because of the interconnection of the scattering mechanisms for charge carriers and phonons. Recent theoretical investigations on sub-device scales have revealed that silicon membranes covered by nanopillars exhibit a multitude of local phonon resonances, spanning the full spectrum, that couple with the heat-carrying phonons in the membrane and collectively cause a reduction in the in-plane thermal conductivity$-$while, in principle, not affecting the electrical properties because the nanopillars are external to the pathway of voltage generation and charge transport. Here this effect is demonstrated experimentally for the first time by investigating device-scale suspended silicon membranes with GaN nanopillars grown on the surface. The nanopillars cause up to 21 % reduction in the thermal conductivity while the electrical conductivity and the Seebeck coefficient remain unaffected, thus demonstrating an unprecedented decoupling in the semiconductor's thermoelectric properties. The measured thermal conductivity behavior for coalesced nanopillars and corresponding lattice-dynamics calculations provide further evidence that the reductions are mechanistically tied to the phonon resonances. This finding breaks a longstanding trade-off between competing properties in thermoelectricity and paves the way for engineered high-efficiency solid-state energy recovery and cooling. △ Less

Submitted 11 January, 2023; originally announced January 2023.

Journal ref: Advanced Materials 35, 2209779 (2023)

Contrastive Learning Improves Critical Event Prediction in COVID-19 Patients

Authors: Tingyi Wanyan, Hossein Honarvar, Suraj K. Jaladanki, Chengxi Zang, Nidhi Naik, Sulaiman Somani, Jessica K. De Freitas, Ishan Paranjpe, Akhil Vaid, Riccardo Miotto, Girish N. Nadkarni, Marinka Zitnik, ArifulAzad, Fei Wang, Ying Ding, Benjamin S. Glicksberg

Abstract: Machine Learning (ML) models typically require large-scale, balanced training data to be robust, generalizable, and effective in the context of healthcare. This has been a major issue for develo** ML models for the coronavirus-disease 2019 (COVID-19) pandemic where data is highly imbalanced, particularly within electronic health records (EHR) research. Conventional approaches in ML use cross-ent… ▽ More Machine Learning (ML) models typically require large-scale, balanced training data to be robust, generalizable, and effective in the context of healthcare. This has been a major issue for develo** ML models for the coronavirus-disease 2019 (COVID-19) pandemic where data is highly imbalanced, particularly within electronic health records (EHR) research. Conventional approaches in ML use cross-entropy loss (CEL) that often suffers from poor margin classification. For the first time, we show that contrastive loss (CL) improves the performance of CEL especially for imbalanced EHR data and the related COVID-19 analyses. This study has been approved by the Institutional Review Board at the Icahn School of Medicine at Mount Sinai. We use EHR data from five hospitals within the Mount Sinai Health System (MSHS) to predict mortality, intubation, and intensive care unit (ICU) transfer in hospitalized COVID-19 patients over 24 and 48 hour time windows. We train two sequential architectures (RNN and RETAIN) using two loss functions (CEL and CL). Models are tested on full sample data set which contain all available data and restricted data set to emulate higher class imbalance.CL models consistently outperform CEL models with the restricted data set on these tasks with differences ranging from 0.04 to 0.15 for AUPRC and 0.05 to 0.1 for AUROC. For the restricted sample, only the CL model maintains proper clustering and is able to identify important features, such as pulse oximetry. CL outperforms CEL in instances of severe class imbalance, on three EHR outcomes with respect to three performance metrics: predictive power, clustering, and feature importance. We believe that the developed CL framework can be expanded and used for EHR ML work in general. △ Less

Submitted 11 January, 2021; originally announced January 2021.

Deep Learning with Heterogeneous Graph Embeddings for Mortality Prediction from Electronic Health Records

Authors: Tingyi Wanyan, Hossein Honarvar, Ariful Azad, Ying Ding, Benjamin S. Glicksberg

Abstract: Computational prediction of in-hospital mortality in the setting of an intensive care unit can help clinical practitioners to guide care and make early decisions for interventions. As clinical data are complex and varied in their structure and components, continued innovation of modeling strategies is required to identify architectures that can best model outcomes. In this work, we train a Heterog… ▽ More Computational prediction of in-hospital mortality in the setting of an intensive care unit can help clinical practitioners to guide care and make early decisions for interventions. As clinical data are complex and varied in their structure and components, continued innovation of modeling strategies is required to identify architectures that can best model outcomes. In this work, we train a Heterogeneous Graph Model (HGM) on Electronic Health Record data and use the resulting embedding vector as additional information added to a Convolutional Neural Network (CNN) model for predicting in-hospital mortality. We show that the additional information provided by including time as a vector in the embedding captures the relationships between medical concepts, lab tests, and diagnoses, which enhances predictive performance. We find that adding HGM to a CNN model increases the mortality prediction accuracy up to 4\%. This framework serves as a foundation for future experiments involving different EHR data types on important healthcare prediction tasks. △ Less

Submitted 27 December, 2020; originally announced December 2020.

Two orders of magnitude reduction in silicon membrane thermal conductivity by resonance hybridizations

Authors: Hossein Honarvar, Mahmoud I. Hussein

Abstract: The thermal conductivity of a freestanding single-crystal silicon membrane may be reduced significantly by attaching nanoscale pillars on one or both surfaces. Atomic resonances of the nanopillars locally and intrinsically couple with the base membrane phonon modes causing these modes to hybridize and flatten at each coupling location in the phonon band structure. The ensuing group velocity reduct… ▽ More The thermal conductivity of a freestanding single-crystal silicon membrane may be reduced significantly by attaching nanoscale pillars on one or both surfaces. Atomic resonances of the nanopillars locally and intrinsically couple with the base membrane phonon modes causing these modes to hybridize and flatten at each coupling location in the phonon band structure. The ensuing group velocity reductions, which in principle may be tuned to take place across silicon's full spectrum, lead to a lowering of the in-plane thermal conductivity in the base membrane. Using equilibrium molecular dynamics simulations, we report a staggering two orders of magnitude reduction in the thermal conductivity at room temperature by this mechanism. △ Less

Submitted 3 July, 2016; v1 submitted 28 June, 2016; originally announced June 2016.

Comments: Labeling arrows corrected in Fig. 3d; manuscript is submitted for journal publication

Journal ref: Physical Review B 97, 195413 (2018)

Thermal transport size effects in silicon membranes featuring nanopillars as local resonators

Authors: Hossein Honarvar, Lina Yang, Mahmoud I. Hussein

Abstract: Silicon membranes patterned by nanometer-scale pillars standing on the surface provide a practical platform for thermal conductivity reduction by resonance hybridization. Using molecular simulations, we investigate the effect of nanopillar size, unit-cell size, and finite-structure size on the net capacity of the local resonators in reducing the thermal conductivity of the base membrane. The resul… ▽ More Silicon membranes patterned by nanometer-scale pillars standing on the surface provide a practical platform for thermal conductivity reduction by resonance hybridization. Using molecular simulations, we investigate the effect of nanopillar size, unit-cell size, and finite-structure size on the net capacity of the local resonators in reducing the thermal conductivity of the base membrane. The results indicate that the thermal conductivity reduction increases as the ratio of the volumetric size of a unit nanopillar to that of the base membrane is increased, and the intensity of this reduction varies with unit-cell size at a rate dependent on the volumetric ratio. Considering sample size, the resonance-induced thermal conductivity drop is shown to increase slightly with the number of unit cells until it would eventually level off. △ Less

Submitted 10 May, 2016; originally announced May 2016.

Comments: Manuscript is submitted for journal publication

Journal ref: Applied Physics Letters 108, 263101, 2016

Spectral energy analysis of locally resonant nanophononic metamaterials by molecular simulations

Authors: Hossein Honarvar, Mahmoud I. Hussein

Abstract: A nanophononic metamaterial is a new type of nanostructured material that features an array, or a forest, of intrinsically distributed resonating substructures. Each substructure exhibits numerous local resonances, each of which may hybridize with the phonon dispersion of the underlying host material causing significant reductions in the group velocities and consequently a reduction in the lattice… ▽ More A nanophononic metamaterial is a new type of nanostructured material that features an array, or a forest, of intrinsically distributed resonating substructures. Each substructure exhibits numerous local resonances, each of which may hybridize with the phonon dispersion of the underlying host material causing significant reductions in the group velocities and consequently a reduction in the lattice thermal conductivity. In this paper, molecular dynamics simulations are utilized to investigate both the dynamics and the thermal transport properties of a nanophononic metamaterial configuration consisting of a freely suspended silicon membrane with an array of silicon nanopillars standing on the surface. The simulations yield results consistent with earlier lattice-dynamics based predictions which showed a reduction in the thermal conductivity due to the presence of the local resonators. Using a spectral energy density approach, in which only simulation data is utilized and no a priori information on the nanostructure resonant phonon modes is provided, we show direct evidence of the existence of resonance hybridizations as an inherent mechanism contributing to the slowing down of thermal transport in this system. △ Less

Submitted 28 November, 2015; originally announced November 2015.

Comments: This work was presented in Paris, France at Phononics 2015 (May 31-June 5, 2015) and a portion of the results are published in the conference proceedings (www.phononics2015.org)

Journal ref: Physical Review B 93, 081412 (2016)