Data scarcity is a major challenge when training deep learning (DL) models. DL demands a large amount of data to achieve exceptional performance. Unfortunately, many applications have small or inadequate data to train DL frameworks. Usually, manual labeling is needed to provide labeled data, which typically involves human annotators with a vast background of knowledge. This annotation process is costly, time-consuming, and error-prone. Usually, every DL framework is fed by a significant amount of labeled data to automatically learn representations. Ultimately, a larger amount of data would generate a better DL model and its performance is also application dependent. This issue is the main barrier for many applications dismissing the use of DL. Having sufficient data is the first step toward any successful and trustworthy DL application. This paper presents a holistic survey on state-of-the-art techniques to deal with training DL models to overcome three challenges including small, imbalanced datasets, and lack of generalization. This survey starts by listing the learning techniques. Next, the types of DL architectures are introduced. After that, state-of-the-art solutions to address the issue of lack of training data are listed, such as Transfer Learning (TL), Self-Supervised Learning (SSL), Generative Adversarial Networks (GANs), Model Architecture (MA), Physics-Informed Neural Network (PINN), and Deep Synthetic Minority Oversampling Technique (DeepSMOTE). Then, these solutions were followed by some related tips about data acquisition needed prior to training purposes, as well as recommendations for ensuring the trustworthiness of the training dataset. The survey ends with a list of applications that suffer from data scarcity, several alternatives are proposed in order to generate more data in each application including Electromagnetic Imaging (EMI), Civil Structural Health Monitoring, Medical imaging, Meteorology, Wireless Communications, Fluid Mechanics, Microelectromechanical system, and Cybersecurity. To the best of the authors’ knowledge, this is the first review that offers a comprehensive overview on strategies to tackle data scarcity in DL.
This paper reports the design, fabrication, and testing of a microfluidic MEMS biosensor for rapid sensing of low concentration Escherichia coli O157:H7. It consists of a specially designed focusing and sensing region, which enables the biosensor to detect low concentration of bacterial cells. The focusing region consists of a ramped vertical electrode pair made of electroplated gold along with tilted thin film finger pairs (45°) embedded inside a microchannel. The focusing region generates positive dielectrophoresis force, which moves the cells towards the edges of the tilted thin film electrode fingers, located at the center of the microchannel. The fluidic drag force then carries the focused cells to the sensing region, where three interdigitated electrode arrays (IDEAs) with 30, 20, and 10 pairs, respectively, are embedded inside the microchannel. This technique resulted in highly concentrated samples in the sensing region. The sensing IDEAs are functionalized with the anti-E. coli antibody for specific sensing of E. coli 0157:H7. As E. coli binds to the antibody, it results in an impedance change, which is measured across a wide frequency range of 100 Hz–10 MHz. The biosensor was fabricated on a glass substrate using the SU8 epoxy resist to form the microchannel, gold electroplating to form the vertical focusing electrode pair, a thin gold film to form the sensing electrode, the finger electrodes, traces and bonding pads, and polydimethylsiloxane to seal the device. The microfluidic impedance biosensor was tested with various low concentration bacterial samples and was able to detect bacterial concentration, as low as 39 CFU/ml with a total sensing time of 2 h.
This paper presents an impedance-based biosensor for rapid and simultaneous detection of Salmonella serotypes B, D, and E with very low concentration. The biosensor consists of a focusing region, and three detection regions. The cells focusing was achieved using a ramp down electroplated vertical electrode pair along with tilted thin film finger pairs that generate p-DEP forces to focus and concentrate the bacterial cells into the center of the microchannel, and direct them toward the detection region. The detection regions consist of three interdigitated electrode arrays (IDEA), each with 20 pairs of finger coated with a mixture of anti-Salmonella antibody and crosslinker to enhance the adhesion to IDEA. The impedance changes as the target Salmonella binds to the antibody. The biosensor has showed excellent performance as proven by the detection of a single Salmonella serotype B, and simultaneous detection of two Salmonella serotypes B and D with a limit of detection (LOD) of 8 Cells/ml in ready-to-eat turkey samples, the addition of focusing capability improved the measured signal by a factor of between 4–4.5, the total detection time of 45 minutes, selectivity of the sensor on different types of bacterial cells, and the ability to distinguish between dead and live cells.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.