Feedback Network Research Articles

Progressive residual feedback network with additional constraints for single image super-resolution

Progressive residual feedback network with additional constraints for single image super-resolution

A CMOS Hall Sensors Array With Integrated Readout Circuit Resilient to Local Magnetic Interference From Current-Carrying Traces

This paper proposes a CMOS Hall sensors array with readout circuitry to counteract the magnetic interference from nearby conductors carrying high currents. The 3×3 Hall sensors array with current-spinning collects the spatial magnetic field information in a time-interleaved manner while the readout circuitry, formed by a capacitively-coupled instrumentation amplifier (CCIA) with T-capacitor feedback network, amplifies the transduced signals for subsequent processing. The calibration mechanism analyzes the magnetic field distribution on the array and separates the magnetic field component due to the environment and local perturbation from adjacent traces, enabling accurate detection of the environmental magnetic field. Prototyped in a 0.35-μm CMOS process, the Hall sensors array and readout circuit occupy an area of 2.2 × 2.2 mm2 and consume 3.5 mA of current under a 3.3-V supply. The hall sensors achieve an average sensitivity of 100 V/(A·T). The input-referred noise of the entire circuit is 0.128 μT/√Hz. With the interference cancellation scheme, the error in the magnetic field readings caused by a conductor 5 mm from the center of the chip carrying 3 A of current reduced from 120 to 5 μT, corresponding to a 24× improvement. Overall, the average error of the captured magnetic field after calibration reaches 0.9%.

Describing complex interactions of social-ecological systems for tipping point assessments: an analytical framework

Humans play an interconnecting role in social-ecological systems (SES), they are part of these systems and act as agents of their destruction and regulation. This study aims to provide an analytical framework, which combines the concept of SES with the concept of tipping dynamics. As a result, we propose an analytical framework describing relevant dynamics and feedbacks within SES based on two matrixes: the “tipping matrix” and the “cross-impact matrix.” We take the Southwestern Amazon as an example for tropical regions at large and apply the proposed analytical framework to identify key underlying sub-systems within the study region: the soil ecosystem, the household livelihood system, the regional social system, and the regional climate system, which are interconnected through a network of feedbacks. We consider these sub-systems as tipping elements (TE), which when put under stress, can cross a tipping point (TP), resulting in a qualitative and potentially irreversible change of the respective TE. By systematically assessing linkages and feedbacks within and between TEs, our proposed analytical framework can provide an entry point for empirically assessing tipping point dynamics such as “tipping cascades,” which means that the crossing of a TP in one TE may force the tipping of another TE. Policy implications: The proposed joint description of the structure and dynamics within and across SES in respect to characteristics of tipping point dynamics promotes a better understanding of human-nature interactions and critical linkages within regional SES that may be used for effectively informing and directing empirical tipping point assessments, monitoring or intervention purposes. Thereby, the framework can inform policy-making for enhancing the resilience of regional SES.

Dynamic cooperation and mutual feedback network for shield machine

A shield machine (SM) is a complex mechanical device used for tunneling. Traditional methods of monitoring SM working conditions relied on artificial experience, which was prone to hidden mechanical failures, human operator errors, and sensor anomalies. Faced with this situation, many scholars have studied intelligent control and decision-making methods for SM. However, most of these methods do not consider the relationship between SM and the operating environment, which makes it difficult for the model to apply to the actual engineering. In order to overcome these limitations, this paper proposes a dynamic cooperation and mutual feedback network based on the Internet of Things (IoT). The network integrates SM, geological information, and control terminals. Then, based on this network, for the two important tasks of the SM field, rate prediction and anomaly detection has established SM Geology Convolutional with Attention (SM-GCA) and SM Geology Variational Autoencoder with Long Short-Term Memory (SM-GVL) models respectively. Our experimental results demonstrate that our proposed two models outperform baseline models in terms of performance.

Fixed-Weight Difference Target Propagation

Target Propagation (TP) is a biologically more plausible algorithm than the error backpropagation (BP) to train deep networks, and improving practicality of TP is an open issue. TP methods require the feedforward and feedback networks to form layer-wise autoencoders for propagating the target values generated at the output layer. However, this causes certain drawbacks; e.g., careful hyperparameter tuning is required to synchronize the feedforward and feedback training, and frequent updates of the feedback path are usually required than that of the feedforward path. Learning of the feedforward and feedback networks is sufficient to make TP methods capable of training, but is having these layer-wise autoencoders a necessary condition for TP to work? We answer this question by presenting Fixed-Weight Difference Target Propagation (FW-DTP) that keeps the feedback weights constant during training. We confirmed that this simple method, which naturally resolves the abovementioned problems of TP, can still deliver informative target values to hidden layers for a given task; indeed, FW-DTP consistently achieves higher test performance than a baseline, the Difference Target Propagation (DTP), on four classification datasets. We also present a novel propagation architecture that explains the exact form of the feedback function of DTP to analyze FW-DTP. Our code is available at https://github.com/TatsukichiShibuya/Fixed-Weight-Difference-Target-Propagation.

Диалог с потребителем в эпоху медиатизации: ответы на отзывы о медицинских услугах в условиях Интернет-коммуникации

The article analyses online communicative interaction of the authors of reviews about provided services of the companies. The focus is on methods of consumer feedback within the framework of public communication. It can be noted that when comparing an online consumer review with its traditional paper prototype the consequences of expressing claims in the “Book of Feedback and Suggestions” and in social networks feedback differ significantly in their consequences for the reputation of the company. In this regard, the strategy of responding to consumer feedback - both positive and negative - is a topical subject matter in terms of effective communication with the consumer in the digital age. The purpose of the study is a communicative-pragmatic analysis of commentary texts on behalf of companies representing a response to consumer feedback in the context of online public communication. Reviews about doctors and medical services on the websites of the specialized network publication “Prodoctorov” and responses from a medical institution were used as materials for study. A communicative-pragmatic description of the response to a consumer feedback is based on the main strategy of the addresser, which is mainly expressed in building and preservation of the company’s image. The tactics of responding to a positive or negative assessment of a medical service are considered. Structural-semantic models of two types of responses to reviews are pointed out. The pragmatic characteristics of the response commentary to the consumer review, which bring it closer to the advertising text, and manipulative techniques in arranging feedback from the clients of medical services are noted.

Discovery of Potential KRAS‐SOS1 Inhibitors from South African Natural Compounds: An In silico Approach

AbstractFor decades the direct targeting of KRAS as a driver of non‐small cell lung cancer, colorectal and pancreatic cancers as well as the inhibition of the RAF‐MEK‐ERK pathway has shown little success due to feedback networks that keep the pathway in control. Inhibiting SOS1, a KRAS activator, has therefore become a promising escape route to treating KRAS‐driven cancers. The search for SOS1 inhibitors has since gained momentum although no drug has been approved yet. Owing to the time‐consuming and difficult processes that characterize the discovery and approval of drugs, natural products have become useful in addressing the unmet medical needs. In this study we employed computational techniques to screen South African natural products for inhibitors with the potential to inhibit SOS1‐KRAS activation. In this study, eight natural compounds, from plants and marine life, possessing antineoplastic activities with good docking and ADMET properties were identified. These compounds, viz., SANCDB00219, SANCDB0290, SANCDB00369, SANCDB00416, SANCDB00421, SANCDB00749, SANCDB00957 and SANCDB001124 exhibited favorable total free binding energies in complex with SOS1‐KRAScharacterized by conventional and carbon hydrogen bonds, van der Waals, pi‐alkyl and pi‐sigma interactions with the binding site residues. It was further revealed that these compounds induced conformational stability on the structural architecture of SOS1‐KRAS, and decreased the structural flexibility of its individual C‐α atoms as a mechanism of inhibition. Upon experimental validation, these compounds from a natural origin could serve as lead identification of small molecules to address SOS1‐KRAS associated diseases.

Impact of variability in cell cycle periodicity on cell population dynamics.

The cell cycle consists of a series of orchestrated events controlled by molecular sensing and feedback networks that ultimately drive the duplication of total DNA and the subsequent division of a single parent cell into two daughter cells. The ability to block the cell cycle and synchronize cells within the same phase has helped understand factors that control cell cycle progression and the properties of each individual phase. Intriguingly, when cells are released from a synchronized state, they do not maintain synchronized cell division and rapidly become asynchronous. The rate and factors that control cellular desynchronization remain largely unknown. In this study, using a combination of experiments and simulations, we investigate the desynchronization properties in cervical cancer cells (HeLa) starting from the G1/S boundary following double-thymidine block. Propidium iodide (PI) DNA staining was used to perform flow cytometry cell cycle analysis at regular 8 hour intervals, and a custom auto-similarity function to assess the desynchronization and quantify the convergence to an asynchronous state. In parallel, we developed a single-cell phenomenological model the returns the DNA amount across the cell cycle stages and fitted the parameters using experimental data. Simulations of population of cells reveal that the cell cycle desynchronization rate is primarily sensitive to the variability of cell cycle duration within a population. To validate the model prediction, we introduced lipopolysaccharide (LPS) to increase cell cycle noise. Indeed, we observed an increase in cell cycle variability under LPS stimulation in HeLa cells, accompanied with an enhanced rate of cell cycle desynchronization. Our results show that the desynchronization rate of artificially synchronized in-phase cell populations can be used a proxy of the degree of variance in cell cycle periodicity, an underexplored axis in cell cycle research.

Semi-supervised aortic dissections segmentation: A time-dependent weighted feedback fusion framework.

The segmentation of true lumen (TL) and false lumen (FL) plays an important role in the diagnosis and treatment of aortic dissection (AD). Although the deep learning methods have achieved remarkable performance for this task, a large number of labeled data are required for training. In order to alleviate the burden of doctors' labeling, in this paper, a novel time-dependent weighted feedback fusion based semi-supervised aortic dissections segmentation framework is proposed by effectively leveraging the unlabeled data. A feedback network is additionally extended to encode the predicted output from the backbone network into high-level feature space, which is then fused with the original feature information of the image to fix previous potential mistakes, thereby segmentation accuracy is improved iteratively. To utilize both labeled data and unlabeled data, the fused feature space flows into the network again to generate the second feedback and make sure consistency with the previous one. The utilization of image feature space provides better robustness and accuracy for the proposed structure. Experiments show that our method outperforms five existing state-of-the-art semi-supervised segmentation methods on both a type-B AD dataset and a public dataset.

Dual-Domain Triggered Iterative Learning Control for Networked Switched Systems Against Denial-of-Service Attacks

In this article, dual-domain triggered iterative learning control (ILC) is studied for networked switched systems against denial-of-service (DoS) attacks. Unlike existing research on ILC focusing only on updating between iterations, a novel dual-domain (time and iteration domains) triggered ILC is proposed. The updates between iterations and updates of network feedback information for each iteration on the time scale are performed, reducing the iteration steps and network data transmission frequency within each iteration. In each iteration, the system output is transmitted via a network, which is assumed to be vulnerable to DoS attacks. An attack detection mechanism and buffer-based compensation mechanism are proposed. Then, the boundedness of the iterative tracking error of the networked switched systems is ensured through a switching stability analysis based on the Lyapunov theory. Finally, the advantages of the proposed methods are substantiated by a simulation example consisting of three parts.

Better Lightweight Network for Free: Codeword Mimic Learning for Massive MIMO CSI Feedback

The channel state information (CSI) needs to be fed back from the user equipment (UE) to the base station (BS) in frequency division duplexing (FDD) multiple-input multiple-output (MIMO) system. Recently, neural networks are widely applied to CSI compressed feedback since the original overhead is too large for the massive MIMO system. Notably, lightweight feedback networks attract special attention due to their practicality of deployment. However, the feedback accuracy is likely to be harmed by the network compression. In this paper, a cost free distillation technique named codeword mimic (CM) is proposed to train better feedback networks with the practical lightweight encoder. A mimic-explore training strategy with a special distillation scheduler is designed to enhance the CM learning. Experiments show that the proposed CM learning outperforms the previous state-of-the-art feedback distillation method, boosting the performance of the lightweight feedback network without any extra inference cost.

A Charge Sensitive Amplifier in a 28 nm CMOS Technology for Pixel Detectors at Future Particle Colliders

This paper is concerned with the design of a Charge Sensitive Amplifier (CSA) in a 28 nm CMOS technology. The CSA discussed in this work is conceived for High Energy Physics (HEP) experiments at next-generation colliders, where pixel detectors will be read out by specific front-end chips, typically including a CSA exploited for charge-to-voltage conversion of the signal delivered by the sensor. The main analog performance parameters of the CSA, also referred to as the pre-amplifier, are assessed here by means of specific Spectre simulations, which are meant to evaluate the behavior of the analog processor in terms of noise, linearity and capability to compensate for very large detector leakage currents. Noise simulations revealed an equivalent noise charge close to 75 electrons rms for typical operating conditions. Up to 50 nA sensor leakage current can be compensated for thanks to the CSA Keummenacher feedback network. The total current consumption of the CSA is close to 2.2 µA, which, together with a power supply of 0.9 V, translates to a power consumption of 2.0 µW.

FNSAM: Image super-resolution using a feedback network with self-attention mechanism.

High-resolution (HR) magnetic resonance imaging (MRI) provides rich pathological information which is of great significance in diagnosis and treatment of brain lesions. However, obtaining HR brain MRI images comes at the cost of extending scan time and using sophisticated expensive instruments. This study aims to reconstruct HR MRI images from low-resolution (LR) images by developing a deep learning based super-resolution (SR) method. We propose a feedback network with self-attention mechanism (FNSAM) for SR reconstruction of brain MRI images. Specifically, a feedback network is built to correct shallow features by using a recurrent neural network (RNN) and the self-attention mechanism (SAM) is integrated into the feedback network for extraction of important information as the feedback signal, which promotes image hierarchy. Experimental results show that the proposed FNSAM obtains more reasonable SR reconstruction of brain MRI images both in peak signal to noise ratio (PSNR) and structural similarity index measure (SSIM) than some state-of-the-arts. Our proposed method is suitable for SR reconstruction of MRI images.

Stability Compensation Design and Analysis of a Piezoelectric Ceramic Driver with an Emitter Follower Stage.

Piezoelectric ceramic has been widely applied in many fields because of its characteristics, and the performance of piezoelectric ceramic is determined strongly by its driver. In this study, an approach to analyzing the stability of a piezoelectric ceramic driver with an emitter follower stage was presented, and a compensation was proposed. First of all, using the method of modified nodal analysis and loop gain analysis, the transfer function for the feedback network was analytically deduced, and the cause of the instability of the driver was found to be the pole composed of the effective capacitance from the piezoelectric ceramic and the transconductance from the emitter follower. Then, a compensation involving a novel delta topology composed of an isolation resistor and a second feedback path was proposed, and its function principle was discussed. Simulations showed a correspondence between the analysis and the effectiveness of the compensation. Finally, an experiment was set up with two prototypes, one with compensation, and the other without compensation. Measurements showed the elimination of oscillation in the compensated driver.

Stable frequency transfer for clock synchronization for telecom networks

A stable and reliable clock synchronization system that transfers reference clocks to nodes is essential for telecom networks. Mainly, the clock synchronization of telecom networks is carried out through satellite links. However, for the evolving ubiquitous wireless communications, especially in some harsh environments, such as underground and trenches, the satellite-based clock synchronization scheme bears inevitable obstacles, which may cause strategic risks for the telecom networks, since no control and manipulation can be imposed on the satellite systems. Therefore, to ensure the robust operation of telecom networks with performance guarantees, some auxiliary means are desirable to assist clock synchronization that is solely based on satellite links. In this article, we report a phase-stabilized frequency signal transmission on branching optical fiber for clock synchronization for telecom networks. Particularly, the phase fluctuation due to optical carriers separation link and fiber link is compensated by a feedback network, which employs a high-precision voltage-controlled oscillator in the phase-locked loop to drive the acousto-optic frequency shifter for fast phase correction. Furtherly, the factors that cause performance limitations of the frequency distribution system are analyzed. Eventually, the experiment results obtained show that a stabilized fiber-optic frequency transfer scheme can be used as a reliable method for clock synchronization with high accuracy.

TMS-evoked responses are driven by recurrent large-scale network dynamics.

A compelling way to disentangle the complexity of the brain is to measure the effects of spatially and temporally synchronized systematic perturbations. In humans, this can be non-invasively achieved by combining transcranial magnetic stimulation (TMS) and electroencephalography (EEG). Spatiotemporally complex and long-lasting TMS-EEG evoked potential (TEP) waveforms are believed to result from recurrent, re-entrant activity that propagates broadly across multiple cortical and subcortical regions, dispersing from and later re-converging on, the primary stimulation site. However, if we loosely understand the TEP of a TMS-stimulated region as the impulse response function of a noisy underdamped harmonic oscillator, then multiple later activity components (waveform peaks) should be expected even for an isolated network node in the complete absence of recurrent inputs. Thus emerges a critically important question for basic and clinical research on human brain dynamics: what parts of the TEP are due to purely local dynamics, what parts are due to reverberant, re-entrant network activity, and how can we distinguish between the two? To disentangle this, we used source-localized TMS-EEG analyses and whole-brain connectome-based computational modelling. Results indicated that recurrent network feedback begins to drive TEP responses from 100 ms post-stimulation, with earlier TEP components being attributable to local reverberatory activity within the stimulated region. Subject-specific estimation of neurophysiological parameters additionally indicated an important role for inhibitory GABAergic neural populations in scaling cortical excitability levels, as reflected in TEP waveform characteristics. The novel discoveries and new software technologies introduced here should be of broad utility in basic and clinical neuroscience research.

A Target Imaging and Recognition Method Based on Raptor Vision

It is a big challenge to quickly and accurately recognize targets in a complex background. The mutual constraints between a wide field of vision (FOV) and high resolution affect the optical tracking and imaging ability in a wide area. In nature, raptors possess unique imaging structures and optic nerve systems that can accurately recognize prey. This paper proposes an imaging system combined with a deep learning algorithm based on the visual characteristics of raptors, aiming to achieve wide FOV, high spatial resolution, and accurate recognition ability. As for the imaging system, two sub-optical systems with different focal lengths and various-size photoreceptor cells jointly simulate the deep fovea of a raptor’s eye. The one simulating the peripheral region has a wide FOV and high sensitivity for capturing the target quickly by means of short focal length and large-size photoreceptor cells, and the other imitating the central region has high resolution for recognizing the target accurately through the long focal length and small-size photoreceptor cells. Furthermore, the proposed algorithm with an attention and feedback network based on octave convolution (AOCNet) simulates the mechanism of the optic nerve pathway by adding it into the convolutional neural network (CNN), thereby enhancing the ability of feature extraction and target recognition. Experimental results show that the target imaging and recognition system eliminates the limitation between wide FOV and high spatial resolution, and effectively improves the accuracy of target recognition in complex backgrounds.

A Dynamic Fusion of Local and Non-Local Features-Based Feedback Network on Super-Resolution

Many Symmetry blocks were proposed in the Single Image Super-Resolution (SISR) task. The Attention-based block is powerful but costly on non-local features, while the Convolutional-based block is good at efficiently handling the local features. However, assembling two different Symmetry blocks will generate an Asymmetry block, making the classic Symmetry-block-based Super-Resolution (SR) architecture fail to deal with these Asymmetry blocks. In this paper, we proposed a new Dynamic fusion of Local and Non-local features-based Feedback Network (DLNFN) for SR, which focus on optimizing the traditional Symmetry-block-based SR architecture to hold two Symmetry blocks in parallel, making two Symmetry-blocks working on what they do best. (1) We introduce the Convolutional-based block for the local features and Attention-based network block for non-local features and propose the Delivery–Adjust–Fusion framework to hold these blocks. (2) we propose a Dynamic Weight block (DW block) which can generate different weight values to fuse the outputs on different feedback iterations. (3) We introduce the MAConv layer to optimize the In block, which is critical for our two blocks-based feedback algorithm. Experiments show our proposed DLNFN can take full advantage of two different blocks and outperform other state-of-the-art algorithms.

Regulation of morphogen pathways by a Drosophila chondroitin sulfate proteoglycan Windpipe.

Morphogens provide quantitative and robust signaling systems to achieve stereotypic patterning and morphogenesis. Heparan sulfate (HS) proteoglycans (HSPGs) are key components of such regulatory feedback networks. In Drosophila, HSPGs serve as co-receptors for a number of morphogens, including Hedgehog (Hh), Wingless (Wg), Decapentaplegic (Dpp) and Unpaired (Upd, or Upd1). Recently, Windpipe (Wdp), a chondroitin sulfate (CS) proteoglycan (CSPG), was found to negatively regulate Upd and Hh signaling. However, the roles of Wdp, and CSPGs in general, in morphogen signaling networks are poorly understood. We found that Wdp is a major CSPG with 4-O-sulfated CS in Drosophila. Overexpression of wdp modulates Dpp and Wg signaling, showing that it is a general regulator of HS-dependent pathways. Although wdp mutant phenotypes are mild in the presence of morphogen signaling buffering systems, this mutant in the absence of Sulf1 or Dally, molecular hubs of the feedback networks, produces high levels of synthetic lethality and various severe morphological phenotypes. Our study indicates a close functional relationship between HS and CS, and identifies the CSPG Wdp as a novel component in morphogen feedback pathways.

Utilization of deep learning and semantic analysis for opinion mining in information extraction: a review

In concern of the increasing availability and popularity of the opinion information sources at a different platform like individual blogs, online feedback, and social network are proliferating and gaining new opportunities and challenges that can be actively exploited using information technology to seek and comprehend people's opinions. In today's field people or entrepreneurs before taking any decision they must be considering the opinion of peoples or information networks. Most of them express the view or opinion through social media platform like Tweeter, Facebook, or blogs on the internet. Therefore, it is essential to analyze to automatically analyze the immense amount of social data available on the internet. Deep learning (DL) has appeared as an influential machine learning (ML) method that studies the properties of different layers or data and provides more advanced predictive results. The study of DL, along with success in many other practical fields, has been widely used in opinion and emotion analysis in recent years. This review explores new challenges in brainstorming and facilitates DL study and the use of semantic analysis. The analysis focuses mainly on methods that try to solve new challenges identified by empathy programs, compared to those already in place. It also conducts extensive research on its current uses in emotion analysis.