A document by Yi Jin. Click on the document to view its contents.

Supporting InformationYee-Fun Lim,* Chee Koon Ng, US Vaitesswar, Kedar Hippalgaonkar*

Identifying Clostridioides difficile infection (CDI) patients at risk of mortality or recurrence will facilitate prevention, timely treatment and improve clinical outcomes. We aim to establish an open-access web-based prediction system, which estimates CDI patients’ mortality and recurrence outcomes, and explains the machine learning prediction with patients’ characteristics. Prognostic models were developed using four various types of machine learning algorithms and statistical logistics regression model utilizing over 15,000 CDI patients from 41 hospitals in Hong Kong. The boosting-based machine learning algorithm Gradient Boosting Machine (Mortality AUC: 0.7878; Recurrence AUC: 0.7076) outperformed statistical models (Mortality AUC: 0.7573; Recurrence AUC: 0.6927) and other machine learning algorithms. The open-access prediction system for clinicians to assess and interpret the risk factors of CDI patients is now available at https://www.cdiml.care/. In this article, we explain the development of machine learning models and illustrate how to apply hyperparameter tuning with cross-validation to optimize the model accuracy.

Abstract:There is a lack of reliable prognostic biomarkers for hypoxic-ischemic (HI) brain injury in preterm infants. Herein, spectrally detailed wavelet scalograms (WSs), derived from the 1024 Hz sampled electroencephalograms (EEG) of preterm fetal sheep after HI (n ¼ 7), are infused into a high-performance deep convolutional neural network (CNN) pattern classifier to identify high-frequency spike transient biomarkers. The deep WS-CNN pattern classifier identifies EEG spikes with remarkable accuracy of 99.81 ¼ 0.15% (area under curve, AUC ¼ 1.000), cross-validated across 5010 EEG waveforms, during the first 6 h post-HI (42 h total), an important clinical period for diagnosis of HI brain injury. Further, a feature-fusion strategy is introduced to extract the spectrally dominant features of the raw EEG epochs to form robust 3D input matrix sets to be infused into the deep 2D-CNNs for pattern classification. The results show that the proposed WS-CNN approach is less sensitive to the potential morphological variations of spikes across all subjects compared to other deep CNNs and spectral-fuzzy classifiers, allowing the user to flexibly choose an approach depending on their computational requirements. Collectively, the data provide a reliable framework that could help support well-timed diagnosis of at-risk neonates in clinical practice.

This Supporting Information includes:[Use this section to include additional explanatory text such as an extended technical description of results, full details of mathematical models, etc.]  Corresponding author Email:  xxxxx@xxxx.xxx

This Supporting Information includes the extended description of the superposition state of the asymmetric double-well system in vacuum system and in solution, truth tables for the residue pairs and their corresponding quantum logic gates, and figures for the double well potential energy surfaces and transmission spectra of the residue pairs.Corresponding Authors Email: sysheu@ym.edu.tw and dyyang@pub.iams.sinica.edu.tw   

This supporting information includes details of the computational model, supplementary figures, interactive 3D CAD model of DraBot and supplementary movies.  Corresponding author Email:  shyni.varghese@duke.edu

This Supporting information includes:1. Component Selection and Performance of SFA2. Actuator Manufacturing and Preparation of Conductive Ink 3. Average Thickness of Conductive Coating layer on PU Foam4. Time Response of the Actuator in Different Modes 5. Characterization and Experimental Setup6. Measurement and Data Analysis7. Design Specifications of Soft Robotic Applications8. Supporting Video Corresponding author Email:  saravanaprashanth@outlook.com, barbara.mazzolai@iit.it 

Abstract

Proprioception, the ability to perceive one’s own configuration and movement in space, enables organisms to safely and accurately interact with their environment and each other. The underlying sensory nerves that make this possible are highly dense and use sophisticated communication pathways to propagate signals from nerves in muscle, skin and joints to the central nervous system wherein the organism can process and react to stimuli. In a step forward to realize robots with such perceptive capability, we propose a flexible sensor framework that incorporates a novel hybrid modeling strategy, taking advantage of computational mechanics and machine learning. We implement the sensor framework on a large, thin and flexible sensor that transforms sparsely distributed strains into continuous surface shape. Finite element (FE) analysis is utilized to determine sensor design parameters, while an FE model is built to enrich the morphological data used in the supervised training to achieve continuous surface reconstruction. A mapping between the local strain data and the enriched surface data is subsequently trained using ensemble learning. This hybrid approach enables real-time, robust and high-order surface shape reconstruction. The sensing performance is evaluated in terms of accuracy, repeatability, and feasibility with numerous scenarios, which has not been demonstrated and reported on such a large-scale (A4-paper-size) sensor before.  

This supporting information includes supplemental figures, movies, additional results, and the pseudocode of breadth-first-search target generation algorithm.  Corresponding author Email:   yyang60@jhu.edu (Y.Y.)     libome@tsinghua.edu (B.L.)

This Supporting information includes interactive plots, videos, and data captured while performing evaluation and validation experiments for our paper. 

Authors   Xulei Wu1,2, Bingjie Dang2, Hong Wang5, Xiulong Wu1,*, and Yuchao Yang2,3,4,* 1School of Electronics and Information Engineering, Anhui University, Hefei 230601, China. 2Key Laboratory of Microelectronic Devices and Circuits (MOE), School of Integrated Circuits, Peking University, Beijing 100871, China.

Figure S1Experimental system: (A) overall view and (B) generator part.  

Figure S1. Detailed explanation of microfabrication step of fully integrated NIR-LFC. a) The wafer-level microfabrication of iMLA-AFF involves a thin Cr lift-off, and plasma enhanced chemical vapor deposition (PECVD) of SiO2, and a thick Cr lift-off (6 nm Cr – 135 nm SiO2 – 130 nm Cr), photolithographic patterning of DNR photoresist (DNR L300-D1, Dong-jin Semichem, Co., Ltd, Korea), and thermal reflow. Note that a DNR photoresist exhibits both UV curable (Negative photoresist) and thermoplastic characteristic, suitable for metal lift-off as well as microlens formation. The hydrophobic coating of fluorocarbon (C4F8) effectively prevents the lateral expansion of microlenses on a metal surface during thermal reflow. iMLA-AFF are inversely bonded to an image sensor with a 60 μm gap spacer and packaged to a compact objective lens by using a UV curable adhesive. The NIR-LFC is fully assembled by combining a 8.5 mm × 4.7 mm printed circuit board with two VCSEL sources and VCSEL housing. b) A scanning electron microscope (SEM) of hexagonally arranged iMLA-AFF with 30 μm in microlens diameter and 3 μm in microlens gap. c) A photograph of fully packaged NIR-LFC. The camera module is connected to flexible extension cable and delivers raw image to Raspberry Pi 4(B). The total physical dimension of camera module is 8.5 mm × 14.0 mm × 5.6 mm.

Figure S1. a) The memristive synaptic behavior with an ideally symmetric and linear weight update ability (constant ΔG for identical pulses) but limited conductance levels (N =20). b) Test accuracy for 10,000 images in the MNIST dataset obtained during the training of memristive DBN as a function of the training epoch (CDth=64).

This Supplementary Information includes: Section S1- Fabrication method                                        Section S2- Actuation method                                          Section S3- Analysis of sixth-DOF torque                     Section S4- Experiments                                                     Figures S1-S31                                                                         Supporting Table                                                                   References                                                                                Other supplementary materials for this manuscript include the following:Supporting SI Videos S1-S10 Corresponding author(s) Email:   gzlum@ntu.edu.sg  

 This file includes: Figures S1 to S12Movies S1 to S10 Other Supplementary Materials for this manuscript include the following: Movies S1 to S10