Minimal Performance Degradation Research Articles

Anti-fouling rotating polymer-based heat exchanger for zero liquid discharge humidification-dehumidification desalination

The objective of this investigation is to assess the performance of various heat exchangers for application in a novel solar-powered zero liquid discharge humidification-dehumidification desalination system. In this study four heat exchangers (HX) having two different flow configurations namely counter flow (cf), cross flow (cr) made up of three different materials namely high-density polyethylene (HDPE), aluminum (Al), and Polypropelyne (PP) were compared in terms of their effectiveness and overall heat transfer coefficient under varying salinity levels (up to 10%) and mixing ratios (0.22–0.45). At a mixing ratio of 0.22 and 0% salinity, PP-HXcr and Al-HXcf exhibited similar effectiveness (∼85%), surpassing that of HDPE-HXcf (∼65%). Despite PP-HXcr's lower thermal conductivity in comparison to Al-HXcf, comparable effectiveness was achieved due to the superior flow distribution in PP-HXcr. Further investigations focused on the impact of salinity on heat exchanger performance. At 3.5% salinity, all heat exchangers experienced a decline in effectiveness and heat transfer coefficient (HTC), with Al-HXcf experiencing a more pronounced decrease compared to PP-HXcr. The higher thermal conductivity of Al-HXcf led to greater salt accumulation, while PP-HXcr demonstrated minimal fouling. As the experiment progressed, fouling increased for all heat exchangers, with the Al-HXcf being practically ineffective at 10% salinity with an effectiveness below 10%. To address the issue of fouling, a rotating cross-flow heat exchanger (RPP-HXcr) was introduced. While the effectiveness of the PP-HXcr drops from 85% to approximately 60% with increasing salinity from 0% to 10%, the RPP-HXcr demonstrates only a marginal decline in effectiveness with increasing salinity. For instance, at mixing ratio of 0.22 when the salinity is increased from 0% to 10%, the effectiveness of RPP-HXcr only drops from 83% to 77%. This exceptional performance was attributed to the continuous contact between the rotating tubes and the incoming feed, effectively preventing fouling and ensuring sustained efficiency. Rotating cross-flow heat exchanger (RPP-HXcr) is introduced and validated as a potentially reliable solution for mitigating fouling, as it demonstrates sustained efficiency and minimal performance degradation across different salinity conditions.

Soft independence guided filter pruning

Filter pruning (FP) is an effective method for reducing the computational costs of convolutional neural networks, and herein, the most critical task involves evaluating the significance of each convolutional filter and eliminating the less important ones while minimizing performance degradation. Most existing FP methods consider only local information, which may prevent them from accurately recognizing the most important filters. To address this limitation, we propose the soft filter independence (SFI) method, which leverages global information to identify the most important filters using their magnitude and correlation information in different functional layers. The SFI criterion measures the replaceability of filters from a global perspective in a network. Filters with low independence can be represented effectively by others, so their information can be accurately conveyed by other filters. In addition, we introduce a novel SFI-based asymptotic pruning ratio, which improves training and pruning stability. Compared to the most advanced FP methods, our method enables CNNs to achieve higher pruning rates and better classification performance.

An Agile Super-Resolution Network via Intelligent Path Selection

In edge computing environments, limited storage and computational resources pose significant challenges to complex super-resolution network models. To address these challenges, we propose an agile super-resolution network via intelligent path selection (ASRN) that utilizes a policy network for dynamic path selection, thereby optimizing the inference process of super-resolution network models. Its primary objective is to substantially reduce the computational burden while maximally maintaining the super-resolution quality. To achieve this goal, a unique reward function is proposed to guide the policy network towards identifying optimal policies. The proposed ASRN not only streamlines the inference process but also significantly boosts inference speed on edge devices without compromising the quality of super-resolution images. Extensive experiments across multiple datasets confirm ASRN’s remarkable ability to accelerate inference speeds while maintaining minimal performance degradation. Additionally, we explore the broad applicability and practical value of ASRN in various edge computing scenarios, indicating its widespread potential in this rapidly evolving domain.

SIM-FED: Secure IoT malware detection model with federated learning

Many IoT devices are presently in use without sufficient security measures. The vulnerability of these devices to malware highlights the necessity for effective methods to identify malicious activity while protecting data privacy. Federated Learning is a collaborative learning technique that can create a fair model while keeping the local data private. To this end, this paper presents the SIM-FED, a novel model for detecting malware in IoT devices. This model utilizes both deep learning and federated learning algorithms. The SIM-FED offers a privacy-preserving solution for IoT malware detection, without sharing data, enhancing security in distributed environments. The model utilizes a lightweight one-dimensional CNN with optimized hyperparameters, reducing preprocessing, and computational overhead. Various federated aggregation strategies are evaluated, and the FedAvg strategy is selected to integrate the results of local models in the proposed model. Moreover, the model's resilience against white-box and black box cyber-attacks is assessed, revealing minimal performance degradation and highlighting its robustness, which is often overlooked in previous studies. The performance of the SIM-FED is assessed by conducting a series of experiments using the IoT-23 dataset. The results show that the SIM-FED outperforms other deep learning and federated learning models in terms of all evaluation metrics, achieving a remarkable 99.52 % accuracy.

What Makes Quantization for Large Language Model Hard? An Empirical Study from the Lens of Perturbation

Quantization has emerged as a promising technique for improving the memory and computational efficiency of large language models (LLMs). Though the trade-off between performance and efficiency is well-known, there is still much to be learned about the relationship between quantization and LLM performance. To shed light on this relationship, we propose a new perspective on quantization, viewing it as perturbations added to the weights and activations of LLMs. We call this approach ``the lens of perturbation". Using this lens, we conduct experiments with various artificial perturbations to explore their impact on LLM performance. Our findings reveal several connections between the properties of perturbations and LLM performance, providing insights into the failure cases of uniform quantization and suggesting potential solutions to improve the robustness of LLM quantization. To demonstrate the significance of our findings, we implement a simple non-uniform quantization approach based on our insights. Our experiments show that this approach achieves minimal performance degradation on both 4-bit weight quantization and 8-bit quantization for weights and activations. These results validate the correctness of our approach and highlight its potential to improve the efficiency of LLMs without sacrificing performance.

A lightweight multi-task learning network based on key area guidance for counterfeit detection

Counterfeit detection traditionally relies on manual efforts, but manual detection efficiency is notably low. The accuracy of deep learning methods is challenging because of the insufficient samples, so it is crucial to allow the model to learn effective representation at a lower training cost. Given the above problems, we proposed a lightweight multi-task learning method that employs an uncomplicated auxiliary task to enhance the main task’s attention and reduce the training sample requirements. A key area guidance algorithm is designed to construct the auxiliary task, disturbing key image areas to generate new samples and training the auxiliary task to recognize the disturbance. This guides the main task in discerning authenticity from these key areas. Additionally, a tailored data preprocessing strategy was designed to improve the method’s performance further. Achieving an impressive 98.8% accuracy in identifying various counterfeiting points, our method outperforms existing advanced methods. Importantly, the method significantly reduces training costs. Even with an 80% reduction in the sample size, the method maintains a 92.1% accuracy, demonstrating minimal performance degradation compared to alternative methods.

Structural effects of asymmetric magnet shape on performance of surface permanent magnet synchronous motors

This study proposes a novel surface permanent magnet synchronous motor (N-SPMSM) structure, which features asymmetric magnets attached to the rotor surface. The N-SPMSM structure exhibits reduced structural complexity and minimal electromagnetic performance degradation. The properties of N-SPMSM are analyzed by comparing its structural complexity (in terms of the shape) and manufacturing complexity and electromagnetic performance [in terms of the cogging–mutual torque ratio and back-electromotive force (EMF) values], with the corresponding values of a ring-type SPMSM (R-SPMSM) and step-skew-type SPMSM (T-SPMSM). The analysis results demonstrate that N-SPMSM has lower shape complexity than T-SPMSM and lower manufacturing complexity than both R-SPMSM and T-SPMSM. The cogging torque reduction and back-EMF performances of N-SPMSM are similar to that of R-SPMSM and T-SPMSM.

Composite aerogel incorporating low temperature phase change microcapsules for enhanced thermal insulation

Cryogenic transportation and storage confront significant challenges from harsh weather conditions, heightened energy consumption, and epidemic situations, compelling the need for the creation of exceptionally efficient thermal insulation materials. To address this demand, a composite phase change aerogel was designed in this study through incorporating low-temperature microencapsulated phase change microcapsules (MPCM) into a cellulose nanofiber/polyvinyl alcohol (CNF/PVA) system. The MPCM, consisting of a polyurethane-acrylate (PUA) shell and an n-tetradecane core, exhibited excellent encapsulation performance with leak-proof capability. Remarkable low-temperature phase change energy storage properties were observed, including a phase change temperature of approximately 6 °C and an impressive phase change enthalpy of 112 J/g. The MPCM also demonstrated stability during successive heating–cooling repetitions, maintaining its heat storage capacity and morphology for at least 300 cycles. These exceptional thermal characteristics endowed the resulting aerogel with effective thermal insulation and temperature retardation abilities. Meanwhile, the integration of CNF/PVA as the matrix in the composite aerogel led to minimal degradation of thermal storage performance compared to pure MPCM. Moreover, the addition of MPCM significantly enhanced the compressive strength, reaching 5.6 times that of the neat CNF/PVA aerogel. The composite aerogel showed a notably low density of 0.165 g/cm3 and could be reshaped through heating. This work provides a simple yet effective idea for designing bulk materials with low-temperature phase change capabilities, offering promising prospects in the field of thermal insulation.

Improved chaos grasshopper optimizer and its application to HRES techno-economic evaluation

Current political and economic trends are moving more and more toward the use of renewable and clean energy as a result of rising energy use and diminishing fossil fuel supplies. In this paper, an improved chaos-based grasshopper optimizer used for techno-economic evaluation in integrated green power systems is investigated. The integrated system consists of a fuel cell system, a wind farm, and solar energy. The integrated solar, wind, and hydrogen fuel cell architectures increase the effectiveness and electrical output of the system while needing less energy storage in structures that are unconnected from the grid. The grasshopper optimization technique and chaos theory have been combined to create the suggested chaotic grasshopper optimizer in this study. The performance, precision, and robustness of this optimization were then assessed, using four benchmark tasks. The ICGO model is utilized to assign suitable ratings to all system devices, thereby guaranteeing the attainment of optimal performance and efficiency. The Net Present Cost (NPC) analysis revealed that the ICGO algorithm attained the lowest minimum NPC value of 274.541E4 USD and the highest maximum NPC value of 311.94E4 USD. The average NPC value of the ICGO algorithm (289.176E4 USD) was found to be comparable to the other algorithms examined in the study. These findings indicate that the ICGO algorithm outperformed other optimization algorithms in minimizing the cost of the renewable energy system. The chaotic grasshopper optimizer can handle several targets, restrictions, and variables with ease, and the results demonstrate that it is substantially more efficient and precise than standard optimization techniques. It is also quite durable, with minimal performance degradation as compared to the benchmark solutions. This study demonstrates the effectiveness of the chaos grasshopper optimizer as an HRES technique.

Amine-functionalized Schiff base covalent organic frameworks supported PdAuIr nanoparticles as high-performance catalysts for formic acid dehydrogenation and hexavalent chromium reduction

Formic acid (FA) holds significant potential as a liquid hydrogen storage medium. However, it is important to improve the reaction rates and extend the practical applications of FA dehydrogenation and Cr(VI) reduction through the development of efficient heterogeneous catalysts. This study reports the synthesis of a uniformly dispersed PdAuIr nanoparticles (NPs) catalyst loaded with amine groups covalent organic frameworks (COFs). The alloyed NPs demonstrated exceptional effectiveness in FA dehydrogenation rate and Cr(VI) reduction. The initial turnover of frequency (TOF) value for FA dehydrogenation without additives was 9970 h−1 at 298 K, the apparent activation energy (Ea) was 30.3 kJ/mol and the rate constant (k) for Cr(VI) reduction was 0.742 min−1. Additionally, it showcased the ability to undergo recycling up to six times with minimal degradation in performance. The results indicate that its remarkable catalytic performance can be attributed primarily to the favorable mass transfer attributes of the aminated COFs supports, the strong metal-support interaction (SMSI), and the synergistic effects among the metals. This study offers a novel perspective on the advancement of efficient and durable heterogeneous catalysts with diverse capabilities, thereby making significant contributions to the fields of energy and environmental preservation.

Numerical roll motion control by using fins based on the linear quadratic regulator and dynamic mode decomposition

Reducing the roll motion and acceleration of a ship is a critical concern for the safety and comfort requirement in a seaway. In this paper, the uncontrolled roll motion time series is obtained under the harsh beam sea simulation for a naval vessel using a single degree of freedom nonlinear equation of motion. Then the linear quadratic controller (LQR) controller is applied by utilizing one pair of fins (port and starboard) that were positioned amidship. The amplitude and rate saturations of the fins are taken into consideration in the simulations. The LQR controller is designed based on an analytical mathematical model, with the selection of appropriate controller parameters. Then, an estimation is made based on the dynamic mode decomposition method for the ship model by taking measurements from online states as if those are derived from the experimental measurements. Finally, the ‘estimated mathematical model’ is used for the LQR controller design. The results demonstrate the successfulness of the estimation process, with only minimal degradation in performance observed.

Impact of Oxygen Storage Components in Prototype Pd-Based Three-Way Catalysts under Exhaust Conditions Relevant to Propane Engines

With increasing concerns about global warming, the push for sustainable and eco-friendly fuels is accelerating. Propane, recognized as liquefied petroleum gas or LPG, has garnered research interest as an alternative fuel due to its notable advantages, including a high-octane rating, reduced greenhouse gas emissions, and potential cost-effectiveness. However, to realize its full potential as an alternative fuel it is essential to develop catalysts that efficiently handle emissions at low temperatures. In our research, we investigated three distinct palladium (Pd)-based three-way catalyst (TWC) formulations (PdRh, Pd-only, and Pd-OSC) to investigate the influence of typical TWC components rhodium (Rh) and oxygen storage components (OSC) in exhaust scenarios relevant to propane-fueled engines. Among these, the formulation containing oxygen storage components (Pd-OSC) showed the highest reactivity for both NO and C3H8 while minimizing performance degradation from hydrothermal aging (HTA). Notably, the temperature of 50% conversion (T50) for propane in the Pd-OSC fresh and HTA sample was lower by 30 °C and 13 °C, respectively, compared to the Pd-only sample, highlighting the role of oxygen storage materials in enhancing catalyst performance, even without dithering. Additionally, N2 physisorption showed that the Pd-OSC sample has a higher surface area and increased pore volume. This underscores the idea that OSC materials not only augment the catalyst’s porosity but also optimize reactant accessibility to active sites, thus elevating catalytic efficiency. In addition to evaluating performance, we further explored the performance and characteristics of the catalysts using catalytic probe reactions, such as water–gas shift and steam reforming reactions.

An implicit barotropic mode solver for MPAS-ocean using a modern Fortran solver interface

We demonstrate use of a modern Fortran solver interface to manage solver algorithms for an implicit barotropic mode solver in the Model for Predictions Across Scales-Ocean (MPAS-O). ForTrilinos, a Fortran interface to Trilinos that contains a large collection of solver capabilities written in C++, has been implemented in MPAS-O to provide access to a suite of linear solver options. By virtue of the simplified wrapper and interface generator (SWIG) automation tool that generates modern Fortran interfaces to C++ code, we were able to implement the Fortran solver interface in MPAS-O using a familiar Fortran coding style while minimizing performance degradation. The ForTrilinos solver interface is written within MPAS-O’s time stepping modules as a subroutine in conjunction with MPAS-O code. Applied to an idealized ocean and a high-resolution realistic ocean test case, parallel performance of ForTrilinos solvers is examined. It is found that parallel scalability of the ForTrilinos solvers is highly dependent on the number of global synchronization points per solver iteration in each iterative solver algorithm. ForTrilinos solvers perform best compared to the Fortran hand-crafted (FHC) solver when the amount of work per processor is large enough. However, parallel scalability is better with the FHC solver and so when the work per core is modest FHC outperforms ForTrilinos. The intercomparison between the ForTrilinos and FHC solvers reveals that this performance hit in the ForTrilinos solver mostly comes from the global synchronization process, while suggesting that the matrix-vector multiplication process in the FHC solver needs to be optimized for better performance.

Enhance membership inference attacks in federated learning

In Federated learning, models in training will unintentionally memorize detailed information about private data, and the aggregation process on the central server requires users to upload their model parameters, making the models still susceptible to membership inference attacks. However, the existing membership inference attacks in federated learning have been less effective. This paper proposes a new membership inference attack method in federated learning that utilizes data poisoning and sequence prediction confidence. By injecting toxic data, the model can remember the detailed information of specific classes in the target private dataset to the maximum extent. The privacy data detailed information of the target clients contained in the model will be represented through the output confidence vector. Afterward, we aggregate the confidence information obtained from multiple epochs in federated learning and utilize the AdaBoost classifier to learn the details from it. Finally, we use different thresholds to partition the predicted confidence scores output by the AdaBoost classifier obtaining the membership information. We conducted experiments on multiple datasets and models to validate the effectiveness of our attack method. The results showed a high attack effectiveness with minimal overall model performance degradation.

Optimized PID Controller of DC-DC Buck Converter based on Archimedes Optimization Algorithm

This research assesses the suitability of the Archimedes Optimization Algorithm (AOA) as a metaheuristic technique to fine-tune a PID controller in a closed-loop DC-DC buck converter. The converter's core function is to regulate output voltage, ensuring stability despite load fluctuations and input voltage changes. The operational effectiveness of the converter hinges significantly on the gain settings of the PID controller and determining the optimal gain setting for the PID controller is a non-trivial task. For robust performance, the PID controller necessitates optimal gain settings, attainable through metaheuristic optimization. The algorithm aids in identifying ideal proportional, integral, and derivative gains based on varying load conditions. Leveraging the metaheuristic algorithm, the PID controller is optimized to minimize voltage errors, reduce overshoot, and enhance response time. The proposed PID controller, optimized using AOA, is contrasted with PID controllers tuned via alternative algorithms including the hybrid Nelder-Mead method (AEONM), artificial ecosystem-based optimization (AEO), differential evolution (DE), and particle swarm optimizer (PSO). Performance evaluation involves injecting a voltage disturbance into the buck converter with load changes of up to 20%. Results demonstrate the superiority of the AOA-optimized PID controller in voltage recovery. It demonstrates a faster response time and outstanding voltage regulation performance, while also exhibiting minimal performance degradation during load changes. This study concludes that the AOA optimization algorithm surpasses other methods in tuning the PID controller for closed-loop DC-DC buck converters.

Extreme γ-Ray Radiation Tolerance of Spectrometer-Grade CsPbBr3 Perovskite Detectors.

The perovskite compound CsPbBr3 has recently been discovered as a promising room-temperature semiconductor radiation detector, offering an inexpensive and easy-to-manufacture alternative to the current benchmark material Cd1-x Znx Te (CZT). The performance of CsPbBr3 sensors is evaluated under harsh conditions, such as high radiation doses often found in industrial settings and extreme radiation in space. Results show minimal degradation in detector performance after exposure to 1Mrad of Co-60 gamma radiation, with no significant change to energy resolution or hole mobility and lifetime. Additionally, many of the devices are still functional after being exposed to a 10Mrad dose over 3 days, and those that do not survive can still be refabricated into working detectors. These results suggest that the failure mode in these devices is likely related to the interface between the electrode and material and their reaction, or the electrode itself and not the material itself. Overall, the study suggests that CsPbBr3 has high potential as a reliable and efficient radiation detector in various applications, including those involving extreme fluxes and energies of gamma-ray radiation.

Comparative study of minutiae selection methods for digital fingerprints.

Biometric systems are more and more used for many applications (physical access control, e-payment, etc.). Digital fingerprint is an interesting biometric modality as it can easily be used for embedded systems (smartcard, smartphone, and smartwatch). A fingerprint template is composed of a set of minutiae used for their comparison. In embedded systems, a secure element is in general used to store and compare fingerprint templates to meet security and privacy requirements. Nevertheless, it is necessary to select a subset of minutiae from a template due to storage and computation constraints. In this study, we present, a comparative study of the main minutiae selection methods from the literature. The considered methods require no further information like the raw image. Experimental results show their relative performance when using different matching algorithms and datasets. We identified that some methods can be used within different contexts (enrollment or verification) with minimal degradation of performance.

Neural-network preconditioners for solving the Dirac equation in lattice gauge theory

This work develops neural-network--based preconditioners to accelerate solution of the Wilson-Dirac normal equation in lattice quantum field theories. The approach is implemented for the two-flavor lattice Schwinger model near the critical point. In this system, neural-network preconditioners are found to accelerate the convergence of the conjugate gradient solver compared with the solution of unpreconditioned systems or those preconditioned with conventional approaches based on even-odd or incomplete Cholesky decompositions, as measured by reductions in the number of iterations and/or complex operations required for convergence. It is also shown that a preconditioner trained on ensembles with small lattice volumes can be used to construct preconditioners for ensembles with many times larger lattice volumes, with minimal degradation of performance. This volume-transferring technique amortizes the training cost and presents a pathway towards scaling such preconditioners to lattice field theory calculations with larger lattice volumes and in four dimensions.

Communication-efficient decentralized elastic-net broad learning system based on quantized and censored communications

As an efficient learning model, the broad learning system (BLS) has achieved great success in machine learning and various applications due to its outstanding performance. Differing from typical deep learning based models, the BLS has a more straightforward structure and few parameters. In this study, we focus on developing a decentralized version of the elastic-net BLS (D-ENBLS) with communication efficiency. The scenario we considered is that the training data is distributed throughout a network of interconnected agents, which prohibits sharing the raw data owing to resource limitations or privacy concerns. In such a distributed paradigm, the communication between agents is an important issue to investigate. From the perspective of saving communication resources, we introduce quantization and communication censoring strategies on the D-ENBLS algorithm to improve the communication procedure and minimize the communication cost with minimal performance degradation. The novel algorithm refers to as DQC-ENBLS, in which quantization reduces the number of bits for each transmission, whereas communication censoring reduces the total number of transmissions. By formulating the training problem as a finite-sum minimization, the alternating multiplication method (ADMM) is employed to solve the optimization problem in a decentralized manner. The experimental results verify that the DQC-ENBLS algorithm can reduce communication cost while maintaining similar performance on the test datasets.

Discovering Symbolic Policy for Building Control using Reinforcement Learning

We propose a learning framework for interpretable HVAC control in buildings using deep reinforcement learning (DRL). Our framework includes a data-driven surrogate environment to emulate building dynamics and a Deep Symbolic Policy for discovering interpretable control policies. We focus on maintaining the temperature within the desired range for occupant comfort. Our results show that the discovered symbolic policies are interpretable and perform well compared to standard DRL algorithms. Additionally, the discovered policies in surrogate models exhibit transferability to physics-based environments with minimal performance degradation.