Improving the accuracy and robustness of RRAM-based in-memory computing against RRAM hardware noise and adversarial attacks

We present a novel deep neural network (DNN) training scheme and RRAM in-memory computing (IMC) hardware evaluation towards achieving high robustness to the RRAM device/array variations and adversarial input attacks. We present improved IMC inference accuracy results evaluated on state-of-the-art DNNs including ResNet-18, AlexNet, and VGG with binary, 2-bit, and 4-bit activation/weight precision for the CIFAR-10 dataset. These DNNs are evaluated with measured noise data obtained from three different RRAM-based IMC prototype chips. Across these various DNNs and IMC chip measurements, we show that our proposed hardware noise-aware DNN training consistently improves DNN inference accuracy for actual IMC hardware, up to 8% accuracy improvement for the CIFAR-10 dataset. We also analyze the impact of our proposed noise injection scheme on the adversarial robustness of ResNet-18 DNNs with 1-bit, 2-bit, and 4-bit activation/weight precision. Our results show up to 6% improvement in the robustness to black-box adversarial input attacks.

Progress Guidance Representation for Robust Interactive Extraction of Buildings from Remotely Sensed Images

Accurate building extraction from remotely sensed images is essential for topographic mapping, cadastral surveying and many other applications. Fully automatic segmentation methods still remain a great challenge due to the poor generalization ability and the inaccurate segmentation results. In this work, we are committed to robust click-based interactive building extraction in remote sensing imagery. We argue that stability is vital to an interactive segmentation system, and we observe that the distance of the newly added click to the boundaries of the previous segmentation mask contains progress guidance information of the interactive segmentation process. To promote the robustness of the interactive segmentation, we exploit this information with the previous segmentation mask, positive and negative clicks to form a progress guidance map, and feed it to a convolutional neural network (CNN) with the original RGB image, we name the network as PGR-Net. In addition, an adaptive zoom-in strategy and an iterative training scheme are proposed to further promote the stability of PGR-Net. Compared with the latest methods FCA and f-BRS, the proposed PGR-Net basically requires 1–2 fewer clicks to achieve the same segmentation results. Comprehensive experiments have demonstrated that the PGR-Net outperforms related state-of-the-art methods on five natural image datasets and three building datasets of remote sensing images.

Generalizability and Quality Control of Deep Learning-Based 2D Echocardiography Segmentation Models in a Large Clinical Dataset

Use of machine learning for automated annotation of heart structures from echocardiographic videos is an active research area, but understanding of comparative, generalizable performance among models is lacking. This study aimed to 1) assess the generalizability of five state-of-the-art machine learning-based echocardiography segmentation models within a large clinical dataset, and 2) test the hypothesis that a quality control (QC) method based on segmentation uncertainty can further improve segmentation results. Five models were applied to 47,431 echocardiography studies that were independent from any training samples. Chamber volume and mass from model segmentations were compared to clinically-reported values. The median absolute errors (MAE) in left ventricular (LV) volumes and ejection fraction exhibited by all five models were comparable to reported inter-observer errors (IOE). MAE for left atrial volume and LV mass were similarly favorable to respective IOE for models trained for those tasks. A single model consistently exhibited the lowest MAE in all five clinically-reported measures. We leveraged the 10-fold cross-validation training scheme of this best-performing model to quantify segmentation uncertainty for potential application as QC. We observed that filtering segmentations with high uncertainty improved segmentation results, leading to decreased volume/mass estimation errors. The addition of contour-convexity filters further improved QC efficiency. In conclusion, five previously published echocardiography segmentation models generalized to a large, independent clinical dataset—segmenting one or multiple cardiac structures with overall accuracy comparable to manual analyses—with variable performance. Convexity-reinforced uncertainty QC efficiently improved segmentation performance and may further facilitate the translation of such models.

Reservoir Computing with Delayed Input for Fast and Easy Optimisation

Reservoir computing is a machine learning method that solves tasks using the response of a dynamical system to a certain input. As the training scheme only involves optimising the weights of the responses of the dynamical system, this method is particularly suited for hardware implementation. Furthermore, the inherent memory of dynamical systems which are suitable for use as reservoirs mean that this method has the potential to perform well on time series prediction tasks, as well as other tasks with time dependence. However, reservoir computing still requires extensive task-dependent parameter optimisation in order to achieve good performance. We demonstrate that by including a time-delayed version of the input for various time series prediction tasks, good performance can be achieved with an unoptimised reservoir. Furthermore, we show that by including the appropriate time-delayed input, one unaltered reservoir can perform well on six different time series prediction tasks at a very low computational expense. Our approach is of particular relevance to hardware implemented reservoirs, as one does not necessarily have access to pertinent optimisation parameters in physical systems but the inclusion of an additional input is generally possible.

HR-MPF: high-resolution representation network with multi-scale progressive fusion for pulmonary nodule segmentation and classification

Accurate segmentation and classification of pulmonary nodules are of great significance to early detection and diagnosis of lung diseases, which can reduce the risk of developing lung cancer and improve patient survival rate. In this paper, we propose an effective network for pulmonary nodule segmentation and classification at one time based on adversarial training scheme. The segmentation network consists of a High-Resolution network with Multi-scale Progressive Fusion (HR-MPF) and a proposed Progressive Decoding Module (PDM) recovering final pixel-wise prediction results. Specifically, the proposed HR-MPF firstly incorporates boosted module to High-Resolution Network (HRNet) in a progressive feature fusion manner. In this case, feature communication is augmented among all levels in this high-resolution network. Then, downstream classification module would identify benign and malignant pulmonary nodules based on feature map from PDM. In the adversarial training scheme, a discriminator is set to optimize HR-MPF and PDM through back propagation. Meanwhile, a reasonably designed multi-task loss function optimizes performance of segmentation and classification overall. To improve the accuracy of boundary prediction crucial to nodule segmentation, a boundary consistency constraint is designed and incorporated in the segmentation loss function. Experiments on publicly available LUNA16 dataset show that the framework outperforms relevant advanced methods in quantitative evaluation and visual perception.

Treadmill Training Effect on the Myokines Content in Skeletal Muscles of Mice With a Metabolic Disorder Model

The effect of treadmill training loads on the content of cytokines in mice skeletal muscles with metabolic disorders induced by a 16 week high fat diet (HFD) was studied. The study included accounting the age and biorhythmological aspects. In the experiment, mice were used at the age of 4 and 32 weeks, by the end of the experiment—respectively 20 and 48 weeks. HFD feeding lasted 16 weeks. Treadmill training were carried out for last 4 weeks six times a week, the duration 60 min and the speed from 15 to 18 m/min. Three modes of loading were applied. The first subgroup was subjected to stress in the morning hours (light phase); the second subgroup was subjected to stress in the evening hours (dark phase); the third subgroup was subjected to loads in the shift mode (the first- and third-weeks treadmill training was used in the morning hours, the second and fourth treadmill training was used in the evening hours). In 20-week-old animals, the exercise effect does not depend on the training regime, however, in 48-week-old animals, the decrease in body weight in mice with the shift training regime was more profound. HFD affected muscle myokine levels. The content of all myokines, except for LIF, decreased, while the concentration of CLCX1 decreased only in young animals in response to HFD. The treadmill training caused multidirectional changes in the concentration of myokines in muscle tissue. The IL-6 content changed most profoundly. These changes were observed in all groups of animals. The changes depended to the greatest extent on the training time scheme. The effect of physical activity on the content of IL-15 in the skeletal muscle tissue was observed mostly in 48-week-old mice. In 20-week-old animals, physical activity led to an increase in the concentration of LIF in muscle tissue when applied under the training during the dark phase or shift training scheme. In the HFD group, this effect was significantly more pronounced. The content of CXCL1 did not change with the use of treadmill training in almost all groups of animals. Physical activity, introduced considering circadian rhythms, is a promising way of influencing metabolic processes both at the cellular and systemic levels, which is important for the search for new ways of correcting metabolic disorders.

Reservoir Computing with Delayed Input for Fast and Easy Optimization

Reservoir computing is a machine learning method that uses the response of a dynamical system to a certain input in order to solve a task. As the training scheme only involves optimising the weights of the responses of the dynamical system, this method is particularly suited for hardware implementation. Furthermore, the inherent memory of dynamical systems which are suitable for use as reservoirs mean that this method has the potential to perform well on time series prediction tasks, as well as other tasks with time dependence. However, reservoir computing still requires extensive task dependent parameter optimisation in order to achieve good performance. We demonstrate that by including a time-delayed version of the input for various time series prediction tasks, good performance can be achieved with an unoptimised reservoir. Furthermore, we show that by including the appropriate time-delayed input, one unaltered reservoir can perform well on six different time series prediction tasks at a very low computational expense. Our approach is of particular relevance to hardware implemented reservoirs, as one does not necessarily have access to pertinent optimisation parameters in physical systems but the inclusion of an additional input is generally possible.

Improving Language Learners’ Use of Self-Regulated Writing Strategies Through Screencast Feedback

The study aims to explore the impacts of an online training scheme developed for higher education learners that integrates self-regulated learning (SRL) writing strategies into screencast feedback in line with the cyclical model of SRL (i.e., forethought, performance, and reflection on performance phases). During each phase, cognitive, metacognitive, affective and socio-interactional SRL writing strategies were introduced through screencast feedback given to the learners’ writing assignments. The participants were undergraduate English Language Teaching (ELT) students at two state universities ( n = 135) in Turkey. Following a mixed-method research design, previous to and after the 6-week training sessions, both quantitative and qualitative data was gathered and analyzed statistically. The results regarding the learners’ reported use of SRL writing strategies indicate a significant increase in the use of SRL writing strategy after the training. Also, the learners’ opinions on receiving screencast feedback and on the SRL training were considerably positive. The findings are meant to contribute to both online education and teacher education pedagogy.