Towards spike-based machine intelligence with neuromorphic computing.

Guided by brain-like 'spiking' computational frameworks, neuromorphic computing-brain-inspired computing for machine intelligence-promises to realize artificial intelligence while reducing the energy requirements of computing platforms. This interdisciplinary field began with the implementation of silicon circuits for biological neural routines, but has evolved to encompass the hardware implementation of algorithms with spike-based encoding and event-driven representations. Here we provide an overview of the developments in neuromorphic computing for both algorithms and hardware and highlight the fundamentals of learning and hardware frameworks. We discuss the main challenges and the future prospects of neuromorphic computing, with emphasis on algorithm-hardware codesign.

Neuromorphic learning with Mott insulator NiO.

Habituation and sensitization (nonassociative learning) are among the most fundamental forms of learning and memory behavior present in organisms that enable adaptation and learning in dynamic environments. Emulating such features of intelligence found in nature in the solid state can serve as inspiration for algorithmic simulations in artificial neural networks and potential use in neuromorphic computing. Here, we demonstrate nonassociative learning with a prototypical Mott insulator, nickel oxide (NiO), under a variety of external stimuli at and above room temperature. Similar to biological species such as Aplysia, habituation and sensitization of NiO possess time-dependent plasticity relying on both strength and time interval between stimuli. A combination of experimental approaches and first-principles calculations reveals that such learning behavior of NiO results from dynamic modulation of its defect and electronic structure. An artificial neural network model inspired by such nonassociative learning is simulated to show advantages for an unsupervised clustering task in accuracy and reducing catastrophic interference, which could help mitigate the stability-plasticity dilemma. Mott insulators can therefore serve as building blocks to examine learning behavior noted in biology and inspire new learning algorithms for artificial intelligence.

RI2AP: Robust and Interpretable 2D Anomaly Prediction in Assembly Pipelines.

Predicting anomalies in manufacturing assembly lines is crucial for reducing time and labor costs and improving processes. For instance, in rocket assembly, premature part failures can lead to significant financial losses and labor inefficiencies. With the abundance of sensor data in the Industry 4.0 era, machine learning (ML) offers potential for early anomaly detection. However, current ML methods for anomaly prediction have limitations, with F1 measure scores of only 50% and 66% for prediction and detection, respectively. This is due to challenges like the rarity of anomalous events, scarcity of high-fidelity simulation data (actual data are expensive), and the complex relationships between anomalies not easily captured using traditional ML approaches. Specifically, these challenges relate to two dimensions of anomaly prediction: predicting when anomalies will occur and understanding the dependencies between them. This paper introduces a new method called Robust and Interpretable 2D Anomaly Prediction (RI2AP) designed to address both dimensions effectively. RI2AP is demonstrated on a rocket assembly simulation, showing up to a 30-point improvement in F1 measure compared to current ML methods. This highlights its potential to enhance automated anomaly prediction in manufacturing. Additionally, RI2AP includes a novel interpretation mechanism inspired by a causal-influence framework, providing domain experts with valuable insights into sensor readings and their impact on predictions. Finally, the RI2AP model was deployed in a real manufacturing setting for assembling rocket parts. Results and insights from this deployment demonstrate the promise of RI2AP for anomaly prediction in manufacturing assembly pipelines.

Roadmap on emerging hardware and technology for machine learning.

Recent progress in artificial intelligence is largely attributed to the rapid development of machine learning, especially in the algorithm and neural network models. However, it is the performance of the hardware, in particular the energy efficiency of a computing system that sets the fundamental limit of the capability of machine learning. Data-centric computing requires a revolution in hardware systems, since traditional digital computers based on transistors and the von Neumann architecture were not purposely designed for neuromorphic computing. A hardware platform based on emerging devices and new architecture is the hope for future computing with dramatically improved throughput and energy efficiency. Building such a system, nevertheless, faces a number of challenges, ranging from materials selection, device optimization, circuit fabrication and system integration, to name a few. The aim of this Roadmap is to present a snapshot of emerging hardware technologies that are potentially beneficial for machine learning, providing the Nanotechnology readers with a perspective of challenges and opportunities in this burgeoning field.

Using Embedded Feature Selection and CNN for Classification on CCD-INID-V1-A New IoT Dataset.

As Internet of Things (IoT) networks expand globally with an annual increase of active devices, providing better safeguards to threats is becoming more prominent. An intrusion detection system (IDS) is the most viable solution that mitigates the threats of cyberattacks. Given the many constraints of the ever-changing network environment of IoT devices, an effective yet lightweight IDS is required to detect cyber anomalies and categorize various cyberattacks. Additionally, most publicly available datasets used for research do not reflect the recent network behaviors, nor are they made from IoT networks. To address these issues, in this paper, we have the following contributions: (1) we create a dataset from IoT networks, namely, the Center for Cyber Defense (CCD) IoT Network Intrusion Dataset V1 (CCD-INID-V1); (2) we propose a hybrid lightweight form of IDS-an embedded model (EM) for feature selection and a convolutional neural network (CNN) for attack detection and classification. The proposed method has two models: (a) RCNN: Random Forest (RF) is combined with CNN and (b) XCNN: eXtreme Gradient Boosting (XGBoost) is combined with CNN. RF and XGBoost are the embedded models to reduce less impactful features. (3) We attempt anomaly (binary) classifications and attack-based (multiclass) classifications on CCD-INID-V1 and two other IoT datasets, the detection_of_IoT_botnet_attacks_N_BaIoT dataset (Balot) and the CIRA-CIC-DoHBrw-2020 dataset (DoH20), to explore the effectiveness of these learning-based security models. Using RCNN, we achieved an Area under the Receiver Characteristic Operator (ROC) Curve (AUC) score of 0.956 with a runtime of 32.28 s on CCD-INID-V1, 0.999 with a runtime of 71.46 s on Balot, and 0.986 with a runtime of 35.45 s on DoH20. Using XCNN, we achieved an AUC score of 0.998 with a runtime of 51.38 s for CCD-INID-V1, 0.999 with a runtime of 72.12 s for Balot, and 0.999 with a runtime of 72.91 s for DoH20. Compared to KNN, XCNN required 86.98% less computational time, and RCNN required 91.74% less computational time to achieve equal or better accurate anomaly detections. We find XCNN and RCNN are consistently efficient and handle scalability well; in particular, 1000 times faster than KNN when dealing with a relatively larger dataset-Balot. Finally, we highlight RCNN and XCNN's ability to accurately detect anomalies with a significant reduction in computational time. This advantage grants flexibility for the IDS placement strategy. Our IDS can be placed at a central server as well as resource-constrained edge devices. Our lightweight IDS requires low train time and hence decreases reaction time to zero-day attacks.

Roadmap of spin-orbit torques.

Spin-orbit torque (SOT) is an emerging technology that enables the efficient manipulation of spintronic devices. The initial processes of interest in SOTs involved electric fields, spin-orbit coupling, conduction electron spins and magnetization. More recently interest has grown to include a variety of other processes that include phonons, magnons, or heat. Over the past decade, many materials have been explored to achieve a larger SOT efficiency. Recently, holistic design to maximize the performance of SOT devices has extended material research from a nonmagnetic layer to a magnetic layer. The rapid development of SOT has spurred a variety of SOT-based applications. In this Roadmap paper, we first review the theories of SOTs by introducing the various mechanisms thought to generate or control SOTs, such as the spin Hall effect, the Rashba-Edelstein effect, the orbital Hall effect, thermal gradients, magnons, and strain effects. Then, we discuss the materials that enable these effects, including metals, metallic alloys, topological insulators, two-dimensional materials, and complex oxides. We also discuss the important roles in SOT devices of different types of magnetic layers, such as magnetic insulators, antiferromagnets, and ferrimagnets. Afterward, we discuss device applications utilizing SOTs. We discuss and compare three-terminal and two-terminal SOT-magnetoresistive random-access memories (MRAMs); we mention various schemes to eliminate the need for an external field. We provide technological application considerations for SOT-MRAM and give perspectives on SOT-based neuromorphic devices and circuits. In addition to SOT-MRAM, we present SOT-based spintronic terahertz generators, nano-oscillators, and domain wall and skyrmion racetrack memories. This paper aims to achieve a comprehensive review of SOT theory, materials, and applications, guiding future SOT development in both the academic and industrial sectors.

Automatic Lung Health Screening Using Respiratory Sounds.

Significant changes have been made on audio-based technologies over years in several different fields. Healthcare is no exception. One of such avenues is health screening based on respiratory sounds. In this paper, we developed a tool to detect respiratory sounds that come from respiratory infection carrying patients. Linear Predictive Cepstral Coefficient (LPCC)-based features were used to characterize such audio clips. With Multilayer Perceptron (MLP)-based classifier, in our experiment, we achieved the highest possible accuracy of 99.22% that was tested on a publicly available respiratory sounds dataset (ICBHI17) (Rocha et al. Physiol. Meas. 40(3):035,001 20) of size 6800+ clips. In addition to other popular machine learning classifiers, our results outperformed common works that exist in the literature.

Deep neural network to detect COVID-19: one architecture for both CT Scans and Chest X-rays.

Since December 2019, the novel COVID-19's spread rate is exponential, and AI-driven tools are used to prevent further spreading [1]. They can help predict, screen, and diagnose COVID-19 positive cases. Within this scope, imaging with Computed Tomography (CT) scans and Chest X-rays (CXRs) are widely used in mass triage situations. In the literature, AI-driven tools are limited to one data type either CT scan or CXR to detect COVID-19 positive cases. Integrating multiple data types could possibly provide more information in detecting anomaly patterns due to COVID-19. Therefore, in this paper, we engineered a Convolutional Neural Network (CNN) -tailored Deep Neural Network (DNN) that can collectively train/test both CT scans and CXRs. In our experiments, we achieved an overall accuracy of 96.28% (AUC = 0.9808 and false negative rate = 0.0208). Further, major existing DNNs provided coherent results while integrating CT scans and CXRs to detect COVID-19 positive cases.

DeepSuccinylSite: a deep learning based approach for protein succinylation site prediction.

BACKGROUND: Protein succinylation has recently emerged as an important and common post-translation modification (PTM) that occurs on lysine residues. Succinylation is notable both in its size (e.g., at 100 Da, it is one of the larger chemical PTMs) and in its ability to modify the net charge of the modified lysine residue from + 1 to - 1 at physiological pH. The gross local changes that occur in proteins upon succinylation have been shown to correspond with changes in gene activity and to be perturbed by defects in the citric acid cycle. These observations, together with the fact that succinate is generated as a metabolic intermediate during cellular respiration, have led to suggestions that protein succinylation may play a role in the interaction between cellular metabolism and important cellular functions. For instance, succinylation likely represents an important aspect of genomic regulation and repair and may have important consequences in the etiology of a number of disease states. In this study, we developed DeepSuccinylSite, a novel prediction tool that uses deep learning methodology along with embedding to identify succinylation sites in proteins based on their primary structure. RESULTS: Using an independent test set of experimentally identified succinylation sites, our method achieved efficiency scores of 79%, 68.7% and 0.48 for sensitivity, specificity and MCC respectively, with an area under the receiver operator characteristic (ROC) curve of 0.8. In side-by-side comparisons with previously described succinylation predictors, DeepSuccinylSite represents a significant improvement in overall accuracy for prediction of succinylation sites. CONCLUSION: Together, these results suggest that our method represents a robust and complementary technique for advanced exploration of protein succinylation.

In situ unsupervised learning using stochastic switching in magneto-electric magnetic tunnel junctions.

Spiking neural networks (SNNs) offer a bio-plausible and potentially power-efficient alternative to conventional deep learning. Although there has been progress towards implementing SNN functionalities in custom CMOS-based hardware using beyond Von Neumann architectures, the power-efficiency of the human brain has remained elusive. This has necessitated investigations of novel material systems which can efficiently mimic the functional units of SNNs, such as neurons and synapses. In this paper, we present a magnetoelectric-magnetic tunnel junction (ME-MTJ) device as a synapse. We arrange these synapses in a crossbar fashion and perform in situ unsupervised learning. We leverage the capacitive nature of write-ports in ME-MTJs, wherein by applying appropriately shaped voltage pulses across the write-port, the ME-MTJ can be switched in a probabilistic manner. We further exploit the sigmoidal switching characteristics of ME-MTJ to tune the synapses to follow the well-known spike timing-dependent plasticity (STDP) rule in a stochastic fashion. Finally, we use the stochastic STDP rule in ME-MTJ synapses to simulate a two-layered SNN to perform image classification tasks on a handwritten digit dataset. Thus, the capacitive write-port and the decoupled-nature of read-write path of ME-MTJs allow us to construct a transistor-less crossbar, suitable for energy-efficient implementation of in situ learning in SNNs. This article is part of the theme issue 'Harmonizing energy-autonomous computing and intelligence'.

Genetic determinants related to pharmacological induction of foetal haemoglobin in transfusion-dependent HbE-ß thalassaemia.

Thalassaemia are the most common inherited autosomal recessive single gene disorders characterized by chronic hereditary haemolytic anaemia due to the absence or reduced synthesis of one or more of the globin chains. Haemoglobin E-ß thalassaemia is the genotype responsible for approximately one half of all severe beta-thalassaemia worldwide. This study proposes to evaluate the effect of various molecular parameters on the response of hydroxyurea. Hydroxyurea was started at an initial dose of 10 mg/kg of body weight/day on 110 transfusion-dependent HbE-ß thalassaemia patients. HbF level was measured by HPLC analysis. ß-Thalassaemia mutations, XmnI and five other SNPs, and α-globin gene deletions and triplications were detected by ARMS-PCR, RFLP-PCR and Gap-PCR, respectively. Based on the factors for evaluating hydroxyurea-response, 42 patients were responders as they showed an increment of Hb from a mean baseline value of 6.45 g/dl (± 0.70) to 7.78 g/dl (± 0.72) post-therapy. Based on increase in HbF above the median value (14.72%) post-therapy, 78 patients were found to be responders. All the 78 responders showed mean decrease in transfusion of 74.26% (± 8.32) with a maximum decrease of 98.43%. There was a significant correlation between decrease in transfusion and increase in HbF level for all 78 responders. XmnI polymorphism showed the strongest association (p < 0.0001) with increase in HbF levels and Hb levels. Patients with α-globin gene deletions were better responders. It was concluded that hydroxyurea treatment is effective in transfusion-dependent HbE-ß thalassaemia patients and the response is best in patients having both XmnI polymorphism and α-deletion.

Alpha Globin Gene Mutation: A Major Determinant of Hydroxyurea Response in Transfusion-Dependent HbE-ß-Thalassaemia.

Thalassaemias are the most common inherited autosomal recessive single gene disorders characterised by chronic hereditary haemolytic anaemia due to absence or reduced synthesis of one or more of the globin chains. Haemoglobin E (HbE)-ß-thalassaemia is the genotype responsible for approximately one-half of all cases of severe ß-thalassaemia worldwide. This study proposes to evaluate response of hydroxyurea in reducing transfusion requirements of severe HbE-ß-thalassaemia patients, and its correlation with foetal haemoglobin (HbF) level and α-mutation. Hydroxyurea was started at a baseline dose in 82 transfusion-dependent HbE-ß-thalassaemia patients. HbF levels and %F-cells were measured. ß-Thalassaemia mutations and α-globin gene deletions and triplications were detected by amplification refractory mutation system (ARMS)-polymerase chain reaction (PCR) and Gap-PCR, respectively. Patients were categorised as good (41.5%), moderate (31.7%), and poor responders (26.8%) based on their decrease in transfusion requirements. Nine patients were excellent responders who became transfusion independent. The mean increase in HbF levels and %F-cells after therapy was correlated with decrease in transfusion requirements. Patients having a deletion of the α-globin gene were better responders. The response was proportional to the number of α-globin gene deletions. We conclude that hydroxyurea treatment decreases transfusion requirements, and the response correlates with α-globin gene deletions.

Medium-scale carbon nanotube thin-film integrated circuits on flexible plastic substrates.

The ability to form integrated circuits on flexible sheets of plastic enables attributes (for example conformal and flexible formats and lightweight and shock resistant construction) in electronic devices that are difficult or impossible to achieve with technologies that use semiconductor wafers or glass plates as substrates. Organic small-molecule and polymer-based materials represent the most widely explored types of semiconductors for such flexible circuitry. Although these materials and those that use films or nanostructures of inorganics have promise for certain applications, existing demonstrations of them in circuits on plastic indicate modest performance characteristics that might restrict the application possibilities. Here we report implementations of a comparatively high-performance carbon-based semiconductor consisting of sub-monolayer, random networks of single-walled carbon nanotubes to yield small- to medium-scale integrated digital circuits, composed of up to nearly 100 transistors on plastic substrates. Transistors in these integrated circuits have excellent properties: mobilities as high as 80 cm(2) V(-1) s(-1), subthreshold slopes as low as 140 m V dec(-1), operating voltages less than 5 V together with deterministic control over the threshold voltages, on/off ratios as high as 10(5), switching speeds in the kilohertz range even for coarse (approximately 100-microm) device geometries, and good mechanical flexibility-all with levels of uniformity and reproducibility that enable high-yield fabrication of integrated circuits. Theoretical calculations, in contexts ranging from heterogeneous percolative transport through the networks to compact models for the transistors to circuit level simulations, provide quantitative and predictive understanding of these systems. Taken together, these results suggest that sub-monolayer films of single-walled carbon nanotubes are attractive materials for flexible integrated circuits, with many potential areas of application in consumer and other areas of electronics.

Bio-threat preparedness: Need for a paradigm shift.

India of late has been vulnerable to Chemical, Biological, Radiological and Nuclear (CBRN) threat, on account of its unique geographic position. Biological threat is an imminent threat in the hands of a terrorist. The public health system of our country is overburdened due to its present role and bio-attack response is not a priority area. This paper suggests that as the prime focus is on the CR and N threats in the integrated CBRN preparedness strategy and that specialized and technical forces are needed to deal with a bio-threat; hence there is a need for a paradigm shift in policy. The emerging field of bio-threat needs to be delinked from the joint family of 'CBRN', with consequent structural and functional changes. A separate specialized cadre needs to be formed for dealing with bio-threat, created from the pool of doctors and non-medical scientists from the AFMS and the DRDO. Structural changes are needed in the organization, to bring in the resources of NCDC, New Delhi for enhanced disease surveillance capacity and creation of a bio-threat mitigation node in the AFMC, Pune.

Ferroelectric capacitors and field-effect transistors as in-memory computing elements for machine learning workloads.

This study discusses the feasibility of Ferroelectric Capacitors (FeCaps) and Ferroelectric Field-Effect Transistors (FeFETs) as In-Memory Computing (IMC) elements to accelerate machine learning (ML) workloads. We conducted an exploration of device fabrication and proposed system-algorithm co-design to boost performance. A novel FeCap device, incorporating an interfacial layer (IL) and Hf 0.5 Zr 0.5 O 2 (HZO), ensures a reduction in operating voltage and enhances HZO scaling while being compatible with CMOS circuits. The IL also enriches ferroelectricity and retention properties. When integrated into crossbar arrays, FeCaps and FeFETs demonstrate their effectiveness as IMC components, eliminating sneak paths and enabling selector-less operation, leading to notable improvements in energy efficiency and area utilization. However, it is worth noting that limited capacitance ratios in FeCaps introduced errors in multiply-and-accumulate (MAC) computations. The proposed co-design approach helps in mitigating these errors and achieves high accuracy in classifying the CIFAR-10 dataset, elevating it from a baseline of 10% to 81.7%. FeFETs in crossbars, with a higher on-off ratio, outperform FeCaps, and our proposed charge-based sensing scheme achieved at least an order of magnitude reduction in power consumption, compared to prevalent current-based methods.

Overexpression of 5-lipoxygenase and its relation with cell proliferation and angiogenesis in 7,12-dimethylbenz(α)anthracene-induced rat mammary carcinogenesis.

The present study was performed to investigate the critical role of 5-lipoxygenase (5-LOX) in 7,12-dimethylbenz(α)anthracene (DMBA)-induced rat mammary inflammation associated carcinogenesis. Female Sprague-Dawley rats at 50 days of age were treated with 7,12-dimethylbenz(α)anthracene (DMBA; 0.5 mg/100 g body weight) by a single tail vein injection, followed by administration of zileuton (2000 mg/kg diet) from week 7 until the termination of the study at 31 wk. 5-LOX protein expression, 5-hydroxyeicosatetraenoic acid (5-HETE), and leukotriene B4 (LTB4 ) production in rat mammary tissue were analyzed at 6, 12, and 24 wk post-DMBA injection. Rate of cell proliferation was analyzed by bromodioxyuridine labeling index (BrdU-LI). Microvessel density, level of VEGF, and MMP-2 were also measured. DMBA induces inflammation in rat mammary gland as early as 6 wk. 5-LOX is upregulated in DMBA treated rats right from 6 wk when compared with their normal counterparts. An overexpression of 5-LOX is accompanied with increase in 5-HETE, LTB4 production and high BrdU-LI with an increase of two key angiogenic factors for tumorigenesis; MMP-2 and VEGF. It was found that 5-LOX specific inhibitor brought about substantial protection against DMBA-induced mammary carcinogenesis. Histological findings showed substantial repair of hyperplastic lesions. There was a significant reduction in the rate of cell proliferation and expression of angiogenic factors, MMP-2 and VEGF. 5-LOX plays an important role in DMBA-induced inflammation associated carcinogenesis via activation of MMP-2 and VEGF. 5-LOX expression can be considered as a critical event in controlling the process of mammary tumor development.

Body composition parameters as correlates of coronary artery disease.

BACKGROUND & OBJECTIVES: Obesity is a major risk factor for cardiovascular disease (CAD). This study was aimed to assess the risk for CAD determined by certain new and conventional body composition parameters such as visceral fat area (VFA), body mass index (BMI), waist to hip ratio (WHR), etc. METHODS: We did an age and sex matched case-control study of acute myocardial infarction with 100 participants in a tertiary care hospital (50 cases and 50 controls) representing the serving army personnel. The relation between VFA, per cent body fat (PBF), BMI, waist and hip circumferences, and WHR to CAD was assessed. RESULTS: The study showed that there was a significantly increased risk for CAD associated with VFA (OR: 5.67; 95% CI: 1.96, 16.95), WHR (7.07; 2.19, 24.27), waist circumference (WC) (2.63; 1.05, 6.66) and BMI (2.53; 1.03, 6.26). INTERPRETATION & CONCLUSIONS: In conclusion, increased VFA, BMI, WHR and WC showed an association with CAD. VFA is a good index for assessing not only visceral fat accumulation but also cardiovascular risk factors.

DIET-SNN: A Low-Latency Spiking Neural Network With Direct Input Encoding and Leakage and Threshold Optimization.

Bioinspired spiking neural networks (SNNs), operating with asynchronous binary signals (or spikes) distributed over time, can potentially lead to greater computational efficiency on event-driven hardware. The state-of-the-art SNNs suffer from high inference latency, resulting from inefficient input encoding and suboptimal settings of the neuron parameters (firing threshold and membrane leak). We propose DIET-SNN, a low-latency deep spiking network trained with gradient descent to optimize the membrane leak and the firing threshold along with other network parameters (weights). The membrane leak and threshold of each layer are optimized with end-to-end backpropagation to achieve competitive accuracy at reduced latency. The input layer directly processes the analog pixel values of an image without converting it to spike train. The first convolutional layer converts analog inputs into spikes where leaky-integrate-and-fire (LIF) neurons integrate the weighted inputs and generate an output spike when the membrane potential crosses the trained firing threshold. The trained membrane leak selectively attenuates the membrane potential, which increases activation sparsity in the network. The reduced latency combined with high activation sparsity provides massive improvements in computational efficiency. We evaluate DIET-SNN on image classification tasks from CIFAR and ImageNet datasets on VGG and ResNet architectures. We achieve top-1 accuracy of 69% with five timesteps (inference latency) on the ImageNet dataset with 12× less compute energy than an equivalent standard artificial neural network (ANN). In addition, DIET-SNN performs 20- 500× faster inference compared to other state-of-the-art SNN models.

RoMIA: a framework for creating Robust Medical Imaging AI models for chest radiographs.

Artificial Intelligence (AI) methods, particularly Deep Neural Networks (DNNs), have shown great promise in a range of medical imaging tasks. However, the susceptibility of DNNs to producing erroneous outputs under the presence of input noise and variations is of great concern and one of the largest challenges to their adoption in medical settings. Towards addressing this challenge, we explore the robustness of DNNs trained for chest radiograph classification under a range of perturbations reflective of clinical settings. We propose RoMIA, a framework for the creation of Robust Medical Imaging AI models. RoMIA adds three key steps to the model training and deployment flow: (i) Noise-added training, wherein a part of the training data is synthetically transformed to represent common noise sources, (ii) Fine-tuning with input mixing, in which the model is refined with inputs formed by mixing data from the original training set with a small number of images from a different source, and (iii) DCT-based denoising, which removes a fraction of high-frequency components of each image before applying the model to classify it. We applied RoMIA to create six different robust models for classifying chest radiographs using the CheXpert dataset. We evaluated the models on the CheXphoto dataset, which consists of naturally and synthetically perturbed images intended to evaluate robustness. Models produced by RoMIA show 3%-5% improvement in robust accuracy, which corresponds to an average reduction of 22.6% in misclassifications. These results suggest that RoMIA can be a useful step towards enabling the adoption of AI models in medical imaging applications.