Samad Azimi Abriz, Mansoor Fateh, Fatemeh Jafarinejad
et al.
Deep learning faces challenges like limited data, vanishing gradients, high parameter counts, and long training times. This article addresses two key issues: 1) data scarcity in ophthalmology and 2) vanishing gradients in deep networks. To overcome data limitations, an image processing‐based data generation method is proposed, expanding the dataset size by 12x. This approach enhances model training and prevents overfitting. For vanishing gradients, a deep neural network is introduced with optimized weight updates in initial layers, enabling the use of more and deeper layers. The proposed methods are validated using the retinal fundus multi‐disease image database dataset, a limited and imbalanced ophthalmology dataset available on the Grand Challenge website. Results show a 10% improvement in model accuracy compared to the original dataset and a 5% improvement over the benchmark reported on the website.
Computer engineering. Computer hardware, Control engineering systems. Automatic machinery (General)
Password leakage incidents often involve the leakage of user passwords and identity information. Because users are accustomed to reusing passwords across multiple network services, attackers can tweak leaked passwords to accurately attack user accounts. This is called a credential tweaking attack. By analyzing large-scale leaked passwords and the corresponding user identity information, this study finds that user strategies for creating passwords are often associated with user identity information. However, current research on credential tweaking attacks relies only on leaked password structures and ignores leaked user identity information when predicting password tweaking strategies. To improve the accuracy of credential tweaking attacks, this study designs a credential tweaking attack optimization method based on user identity information. In the preprocessing phase, username and regional information is extracted from the user identity information and the probability of users' different password creation strategies in different regions is statistically calculated. In the training phase, regional information is combined to learn users' character-level editing operations on leaked passwords. In the password generation phase, a password generation method that integrates character-level editing operations, structure-level editing operations, and username information is designed. The experimental results show that in an attack with 10<sup>3</sup> guesses, the cracking rate of this method has a maximum improvement of 41.8% compared to the existing best method (PassBERT), highlighting the threat posed by credential tweaking attacks based on user identity information to password security.
Abstract Models based on MLP-Mixer architecture are becoming popular, but they still suffer from adversarial examples. Although it has been shown that MLP-Mixer is more robust to adversarial attacks compared to convolutional neural networks (CNNs), there has been no research on adversarial attacks tailored to its architecture. In this paper, we fill this gap. We propose a dedicated attack framework called Maxwell’s demon Attack (MA). Specifically, we break the channel-mixing and token-mixing mechanisms of the MLP-Mixer by perturbing inputs of each Mixer layer to achieve high transferability. We demonstrate that disrupting the MLP-Mixer’s capture of the main information of images by masking its inputs can generate adversarial examples with cross-architectural transferability. Extensive evaluations show the effectiveness and superior performance of MA. Perturbations generated based on masked inputs obtain a higher success rate of black-box attacks than existing transfer attacks. Moreover, our approach can be easily combined with existing methods to improve the transferability both within MLP-Mixer based models and to models with different architectures. We achieve up to 55.9% attack performance improvement. Our work exploits the true generalization potential of the MLP-Mixer adversarial space and helps make it more robust for future deployments.
Luis J. Lopez-Giraldo, Hernando Amaya, Johana A. Alvarez
et al.
Soursop (Anona Muricata Lin.), a tropical fruit native to Central and South America, is recognized for its rich composition of water, sugars, proteins, vitamins, minerals, and dietary fibre. Despite its potential for food and dietary supplement production, approximately 30% of the Colombian soursop production is discarded due to non-compliance with regulated standards for this highly perishable fruit. This has prompted research into utilizing these discarded residues for ethanol production through fermentation processes.
In this study, the feasibility of ethanol production from soursop leachate was assessed on a laboratory pilot scale. The investigation focused on the use of low-cost inorganic nitrogen sources, namely ammonium sulphate, ammonium chloride, and urea. Additionally, three distinct scale-up methodologies Reynolds number, volumetric power, and impeller tip speed were evaluated in working volumes of 0.5 and 5 L. A comprehensive rheological and hydrodynamic analysis was conducted using native yeasts.
The results revealed that ammonium chloride, at a carbon/nitrogen ratio of 15/1, demonstrated the highest yields, reaching 0.33 g/g. Furthermore, the study indicated that employing volumetric power with increased agitation enhanced biomass production in 4.85 g/L, while impeller tip speed with moderate agitation resulted in higher ethanol production, yielding 0.28 g/g. These findings underscore the versatility of soursop in ethanol production and emphasize the significance of selecting an appropriate scale-up methodology aligned with production objectives. The implications of this choice on the achieved yields are substantial, emphasizing the importance of tailored approaches in optimizing ethanol production from soursop leachate.
Chemical engineering, Computer engineering. Computer hardware
This editorial presents innovative research at the intersection of edge computing and various disciplines, demonstrating the transformative potential of edge computing technologies. The issue includes a study on a digital Proportional-Integral-Derivative (PID) regulator model for controlling unmanned aerial vehicles, using digital filtering methods and a genetic algorithm; discusses an autonomous Internet of Things (IoT) system for monitoring classroom microclimates, contributing to the understanding of how microclimate parameters influence the physiological state of students; details the design and implementation of an educational model for a smart home, integrating various subsystems and renewable energy sources. This issue aims to inspire further exploration and innovation in edge computing, driving the field forward and opening up new possibilities for technology and society.
In real time, the speech signal received contains noise produced in the background and reverberations. These disturbances reduce the quality of speech; therefore, it is important to eliminate the noise and increase the intelligibility and quality of speech signal. Speech enhancement is the primary task in any real-time application that handles speech signals. In the proposed method, the most effective and challenging noise, i.e., babble noise, is removed, and the clean speech is recovered. The enhancement of the corrupted speech signal is done by applying a deep neural network-based denoising algorithm in which the ideal ratio mask is used to mask the noisy speech and separate the clean speech signal. In the proposed system, the speech signal corrupted by noise is enhanced. Evaluation of enhanced speech signal by performance metrics such as short time objective intelligibility and signal to noise ratio of the denoised speech show that the speech intelligibility and speech quality are improved by the proposed method.
Computer engineering. Computer hardware, Mechanics of engineering. Applied mechanics
Abstract Cloud-native database systems have started to gain broad support and popularity due to more and more applications and systems moving to the cloud. Various cloud-native databases have been emerging in recent years, but their developments are still in the primary stage. At this stage, database developers are generally confused about improving the performance of the database by applying AI technologies. The maturity model can help database developers formulate the measures and clarify the improvement path during development. However, the current maturity models are unsuitable for cloud-native databases since their architecture and resource management differ from traditional databases. Hence, we propose a maturity model for AI-empowered cloud-native databases from the perspective of resource management. We employ a systematic literature review and expert interviews to conduct the maturity model. Also, we develop an assessment tool based on the maturity model to help developers assess cloud-native databases. And we provide an assessment case to prove our maturity model. The assessment case results show that the database’s development direction conforms to the maturity model. It proves the effectiveness of the maturity model.
Abstract Microelectronic computers have encountered challenges in meeting all of today’s demands for information processing. Meeting these demands will require the development of unconventional computers employing alternative processing models and new device physics. Neural network models have come to dominate modern machine learning algorithms, and specialized electronic hardware has been developed to implement them more efficiently. A silicon photonic integration industry promises to bring manufacturing ecosystems normally reserved for microelectronics to photonics. Photonic devices have already found simple analog signal processing niches where electronics cannot provide sufficient bandwidth and reconfigurability. In order to solve more complex information processing problems, they will have to adopt a processing model that generalizes and scales. Neuromorphic photonics aims to map physical models of optoelectronic systems to abstract models of neural networks. It represents a new opportunity for machine information processing on sub-nanosecond timescales, with application to mathematical programming, intelligent radio frequency signal processing, and real-time control. The strategy of neuromorphic engineering is to externalize the risk of developing computational theory alongside hardware. The strategy of remaining compatible with silicon photonics externalizes the risk of platform development. In this perspective article, we provide a rationale for a neuromorphic photonics processor, envisioning its architecture and a compiler. We also discuss how it can be interfaced with a general purpose computer, i.e. a CPU, as a coprocessor to target specific applications. This paper is intended for a wide audience and provides a roadmap for expanding research in the direction of transforming neuromorphic photonics into a viable and useful candidate for accelerating neuromorphic computing.
To address the attacks on the communication data by malicious or compromised nodes in Wireless Sensor Network(WSN),this paper proposes a feedback trust model based on double cluster head to ensure the reliability and integrity of data in transmission,perception and fusion.The results of direct interactions between nodes are used to evaluate their direct trust,and the communication,data perception and fusion trust are considered at the same time.The communication trust of neighbor nodes is evaluated by using the Bayesian formula.The historical trust of nodes is used as the supplement of direct trust by using time sliding window,and the weights of direct and indirect trust are adjusted dynamically to make the comprehensive trust evaluation more objective and accurate.On this basis,the monitoring mechanism for interactions between double cluster head and the feedback mechanism for base station trust are introduced.The primary cluster head and the supervisory cluster head independently evaluate the data perception trust of the members according to the spatial correlation of the local data.The base station uses the temporal correlation of the data fusion results of the dual cluster head to evaluate the data fusion trust,and feeds back the final trust results to all nodes.Simulation results show that the model can effectively detect abnormal data and malicious nodes,successfully resist selective forwarding attacks,forgery local data attacks and forgery fusion data attacks,achieving a good balance between network security and energy consumption.
Jens Trautmann, Arthur Beckers, Lennert Wouters
et al.
Locating a cryptographic operation in a side-channel trace, i.e. finding out where it is in the time domain, without having a template, can be a tedious task even for unprotected implementations. The sheer amount of data can be overwhelming. In a simple call to OpenSSL for AES-128 ECB encryption of a single data block, only 0.00028% of the trace relate to the actual AES-128 encryption. The rest is overhead. We introduce the (to our best knowledge) first method to locate a cryptographic operation in a side-channel trace in a largely automated fashion. The method exploits meta information about the cryptographic operation and requires an estimate of its implementation’s execution time.
The method lends itself to parallelization and our implementation in a tool greatly benefits from GPU acceleration. The tool can be used offline for trace segmentation and for generating a template which can then be used online in real-time waveformmatching based triggering systems for trace acquisition or fault injection. We evaluate it in six scenarios involving hardware and software implementations of different cryptographic operations executed on diverse platforms. Two of these scenarios cover realistic protocol level use-cases and demonstrate the real-world applicability of our tool in scenarios where classical leakage-detection techniques would not work. The results highlight the usefulness of the tool because it reliably and efficiently automates the task and therefore frees up time of the analyst.
The method does not work on traces of implementations protected by effective time randomization countermeasures, e.g. random delays and unstable clock frequency, but is not affected by masking, shuffling and similar countermeasures.
Computer engineering. Computer hardware, Information technology
&NA; Marmosets are small New World primates that are posited to become an important preclinical animal model for studying intractable human brain diseases. A critical step in the development of marmosets as a viable model for human brain dysfunction is to characterize brain networks that are homologous with human network topologies. In this regard, the use of functional magnetic resonance imaging (fMRI) holds tremendous potential for functional brain mapping in marmosets. Although possible, implementation of hardware for fMRI in awake marmosets (free of the confounding effects of anesthesia) is not trivial due to the technical challenges associated with developing specialized imaging hardware. Here, we describe the design and implementation of a marmoset holder and head‐fixation system with an integrated receive coil for awake marmoset fMRI. This design minimized head motion, with less than 100 &mgr;m of translation and 0.5 degrees of rotation over 15 consecutive resting state fMRI runs (at 15 min each) across 3 different marmosets. The fMRI data was of sufficient quality to reliably extract 8 resting state networks from each animal with only 60–90 min of resting state fMRI acquisition per animal. The restraint system proved to be an efficient and practical solution for securing an awake marmoset and positioning a receive array within minutes, limiting stress to the animal. This design is also amenable for multimodal imaging, allowing for electrode or lens placement above the skull via the open chamber design. All computer‐aided‐design (CAD) files and engineering drawings are provided as an open resource, with the majority of the parts designed to be 3D printed. HighlightsThe marmoset is a powerful preclinical model for studying human brain diseases.fMRI holds tremendous potential for functional brain mapping in marmosets.Awake marmoset fMRI (task‐based) is not trivial with little available MRIhardware.Here, we provide openly available designs allowing for fully awake marmoset fMRI.
Albert Ho Yuen Lau, Gary Kwok Ki Chik, Zhengyang Zhang
et al.
Thermal ablation has been adopted as one of the most common cancer treatment approaches in medical surgery. By increasing the temperature (>50 °C) on the cells, the cells are destroyed because of denaturation. Herein, an ultrathin Archimedean spiral pattern heater/sensor technology is introduced which can perform ablation by attaching conformally onto the organs for precise heating and temperature sensing. In the heater mode, the heater temperature is linearly proportional to the input joule heating power up to 400 mW. In the sensor mode, the temperature of the conformal metal wire is also linearly related to the resistance by the temperature coefficient of resistance (TCR). The conformal heater to perform ex vivo ablation on the porcine liver is utilized. By further integrating the devices with robotic palm and perform heat‐and‐sense interactions, a human–machine interface (HMI) apparatus is demonstrated which can be potentially applied in surgical robots or other tactile stimulation systems.
Computer engineering. Computer hardware, Control engineering systems. Automatic machinery (General)
Abstract In addressing the challenge of Big Data Analytics, what has been of notable significance is the analysis of online search traffic data in order to analyze and predict human behavior. Over the last decade, since the establishment of the most popular such tool, Google Trends, the use of online data has been proven valuable in various research fields, including -but not limited to- medicine, economics, politics, the environment, and behavior. In the field of politics, given the inability of poll agencies to always well approximate voting intentions and results over the past years, what is imperative is to find new methods of predicting elections and referendum outcomes. This paper aims at presenting a methodology of predicting referendum results using Google Trends; a method applied and verified in six separate occasions: the 2014 Scottish Referendum, the 2015 Greek Referendum, the 2016 UK Referendum, the 2016 Hungarian Referendum, the 2016 Italian Referendum, and the 2017 Turkish Referendum. Said referendums were of importance for the respective country and the EU as well, and received wide international attention. Google Trends has been empirically verified to be a tool that can accurately measure behavioral changes as it takes into account the users’ revealed and not the stated preferences. Thus we argue that, in the time of intelligence excess, Google Trends can well address the analysis of social changes that the internet brings.
Computer engineering. Computer hardware, Information technology
Biofuels have to be produced from biomass under sustainable conditions accomplishing appropriate characteristics and specification for their use as transportation fuels. Nowadays, renewable diesel appears as a more promising option than traditional biodiesel (methyl ester), because its properties are closer to diesel, facilitating its use in current diesel motors. Hydrotreating process for obtaining renewable diesel requires catalysts which are generally supported metals. In these catalysts, the active phase can be a reduced metal, sulphide, carbide or phosphide metallic or bimetallic; while the support can be active carbon, alumina or zeolite. The reduced metals are the simplest to prepare. They do not undergo leaching and have good selectivity. In this work, the behaviour of these catalysts in the production of renewable diesel from palm oil was evaluated. Tests in a STA (simultaneous TGA and DSC) were used to compare the catalytic activity of Pt, Pd, Rh and Ru catalysts, studying the effect on selectivity and conversion of the type and concentration of the active phase, as well as the support. The tests were carried out at 5 MPa of H2 from 20 °C to 290 °C with a temperature ramp of 10 °C/min. IR and GC-MS were also used for identifying the products obtained at the best operating conditions. The best results were obtained for Rh/C, Ru/C and Pt/USY (CBV-780), mainly generation of hydrocarbons such as n-C15 alkanes up to n-C18. Also, the appearance of fatty acids with the same number of carbons was detected.
Chemical engineering, Computer engineering. Computer hardware
Abstract This article presents a security module based on a field programmable gate array (FPGA) to mitigate man-in-the-middle cyber attacks. Nowadays, the FPGA is considered to be the state of the art in nuclear power plants IC it also provides acceptable solutions for embedded computing applications that require cybersecurity. The proposed FPGA-based security module is developed to mitigate information-gathering attacks, which can be made by gaining physical access to the network, e.g., a man-in-the-middle attack, using a cryptographic process to ensure data confidentiality and integrity and prevent injecting malware or malicious data into the critical digital assets of a nuclear power plant data communication system. A model-based system engineering approach is applied. System requirements analysis and enhanced function flow block diagrams are created and simulated using CORE9 to compare the performance of the current and developed systems. Hardware description language code for encryption and serial communication is developed using Vivado Design Suite 2017.2 as a programming tool to run the system synthesis and implementation for performance simulation and design verification. Simple windows are developed using Java for physical testing and communication between a personal computer and the FPGA.