Efficient Aggregation Research Articles

Optimal product aggregation for sales and operations planning in mass customisation context

Sales and operations planning (S&OP) is a multi-actor tactical planning process that balances aggregate demand and capacities to deliver realistic production plans. The way demand is aggregated into product families (PFs) is crucial for the alignment sought by S&OP. The issue of gathering individual products into PFs is particularly challenging in the mass customisation (MC) context where products are depicted through combinations of attributes derived from a set of features. This paper introduces the aggregation scheme concept that refers to the subset of features used to constitute PFs. Several aggregation schemes may be considered with different performances from the perspective of the actors involved in the S&OP. A multi-objective model is designed to identify efficient aggregation schemes considering both sales and operations expectations. The model is implemented on numerical instances from automotive industry. The results highlight the existence of more efficient and balanced aggregation schemes than those currently used.

Olive: Oblivious Federated Learning on Trusted Execution Environment against the Risk of Sparsification

Combining Federated Learning (FL) with a Trusted Execution Environment (TEE) is a promising approach for realizing privacy-preserving FL, which has garnered significant academic attention in recent years. Implementing the TEE on the server side enables each round of FL to proceed without exposing the client's gradient information to untrusted servers. This addresses usability gaps in existing secure aggregation schemes as well as utility gaps in differentially private FL. However, to address the issue using a TEE, the vulnerabilities of server-side TEEs need to be considered---this has not been sufficiently investigated in the context of FL. The main technical contribution of this study is the analysis of the vulnerabilities of TEE in FL and the defense. First, we theoretically analyze the leakage of memory access patterns, revealing the risk of sparsified gradients, which are commonly used in FL to enhance communication efficiency and model accuracy. Second, we devise an inference attack to link memory access patterns to sensitive information in the training dataset. Finally, we propose an oblivious yet efficient aggregation algorithm to prevent memory access pattern leakage. Our experiments on real-world data demonstrate that the proposed method functions efficiently in practical scales.

URTSegNet: A real-time segmentation network of unstructured road at night based on thermal infrared images for autonomous robot system

When it comes to poor lighting conditions such as night, backlight, haze, etc., the detection method of unstructured roads based on visible light cameras often fails. Therefore, this paper aims to improve the detection performance of unstructured roads based on thermal infrared images for autonomous robot systems. For this purpose, we propose a real-time network for unstructured road segmentation at night based on thermal infrared images (URTSegNet) in this paper. Like BiSeNet V2, we adopt a two-branch structure, including Detail Branch and Semantic Branch, to extract spatial detail and contextual features. The Detail Branch maintains a small receptive field and high capacity and adjusts the computational cost by employing depth-wise separable convolutions. In the Semantic Branch, we modify Stem Block to obtain more information because of the characteristics of infrared images. Dilated convolution expands the receptive field while keeping the number of parameters unchanged. Finally, we design a more concise and efficient Aggregation Block using a gate mechanism. An infrared unstructured road dataset, including total darkness at night, low light, backlight, haze, and other scenes, is collected to get the real-world dataset and effectively evaluate performance. We got 97.5% Intersection over Union (IoU) and reached an inference speed of 129 FPS on a single NVIDIA GTX 1660S card on our dataset. In addition, extensive evaluation experiments are conducted on two public datasets to verify the generalization ability of the model and the role of each module. This research addresses the problem of road detection for unmanned platforms in unstructured environments at night or under poor lighting conditions. The algorithm has high accuracy and a fast inference speed so that it can be implemented on various embedded devices.

Effective Aggregate Data Collection and Enhanced Network Lifetime Using Energy Efficient Aggregation Data Convening Routing in Wireless Sensor Network

Wireless Technology is an integral part of today's applications. It encompasses commercial and human-oriented environmental monitoring, military applications, robotics, and healthcare monitoring systems, home or factory automation, and logistics. A wireless sensor network is a network system that uses wireless sensor nodes to monitor physical or environmental conditions such as voice, temperature, and spatial dispersive movements. Each node can locally sense its environment, process information and data, and send the data to one or more collection points within the WSN. The existing solution is categorized into member nodes and group/cluster heads (CH). The CH election process increases the overhead of the network and reduces the network lifetime. The processing and energy limitations of the nodes are considered for the CH election process. In this cluster formation methods aiming at the Cluster head selection process and providing trust in hierarchical WSN are proposed. The Energy Efficient Aggregation Data Convening Routing (E2ADCR) is used to estimate the routing path, and aggregate data collection to improve the network lifetime. The major advantage of this technique is to avoid the malicious or selfish node from becoming a dominant cluster in a group of clusters. Initially sink node selection forwards the Configuration Message (CM) to every node on the network to construct the performing node. In this, cluster selection is based on connection density, degree of the node angle, and residual energy (Quality Factor) that is evaluated from the link robustness, energy, and degree of the node. Multi-hop link transmission support path optimization technique is estimated in the path when the obstacle is present in the WSN. To introduce an Aggregated Support-based Data Collection to evaluate each packet flow monitor on the network to if any unrelated packet will eliminate to forward to the sink node. The proposed routing path of computational simplicity is achieved by a simple method. The new routing protocols, which were developed during this research, have better energy efficiency.

Decomposition and aggregation of tone efficiencies

In this paper we investigate the decomposition and aggregation of cost efficiencies when input prices vary across decision-making units. In particular, we compare and contrast four alternative decompositions, three of which are compatible with non-homogeneous (in terms of quality) inputs. We also examine the relations among the alternative cost efficiency measures and their components. Our theoretical results indicate that, when input prices vary across decision-making units, the Färe and Grosskopf cost-based efficiency measure is equal to (i) the Tone cost efficiency measure; (ii) the Tone and Tsutsui cost efficiency measure if there is no technical inefficiency; and (iii) the input spending efficiency measure if there is no allocative inefficiency in the sense of Tone. In addition, based on the denominator rule, we derive appropriate weights for aggregating consistently these cost efficiencies and their components across decision-making units. All these provide helpful guidelines for the appropriate applications of these decompositions at both the individual and the aggregate level. We illustrate this by a simple study case for hospitals.

An efficient local multi‐energy systems planning method with long‐term storage

AbstractLong‐term storage will play a crucial role in future local multi‐energy systems (MES) with high penetration renewable energy integration for demand balancing. Local MES planning with long‐term energy storage is essentially a very large‐scale program because numerous decision variables, including binary variables, should be used to model long‐term energy dependencies for accurate operational cost estimation. How to largely reduce decision variables as well as guarantee the planning model accuracy becomes one main concern. To this end, this paper proposes a novel efficient aggregation and modeling method for local MES planning. The aggregation method first decomposes input time series data (renewable energy output and energy demand) into hourly and daily components, based on which more accurate aggregation results with a few typical scenarios can be derived. By incorporating similar decomposition into the operation model of energy devices, the planning model can describe the long‐term energy cycle and the hourly operation characteristic at the same time and yield accurate optimization results with limited complexity. Experimental results show that the proposed method can considerably decrease the complexity of the problem while maintaining agreement with the results based on the optimization of the full‐time series.

Monkeypox detection from skin lesion images using an amalgamation of CNN models aided with Beta function-based normalization scheme.

We have recently been witnessing that our society is starting to heal from the impacts of COVID-19. The economic, social and cultural impacts of a pandemic cannot be ignored and we should be properly equipped to deal with similar situations in future. Recently, Monkeypox has been concerning the international health community with its lethal impacts for a probable pandemic. In such situations, having appropriate protocols and methodologies to deal with the outbreak efficiently is of paramount interest to the world. Early diagnosis and treatment stand as the only viable option to tackle such problems. To this end, in this paper, we propose an ensemble learning-based framework to detect the presence of the Monkeypox virus from skin lesion images. We first consider three pre-trained base learners, namely Inception V3, Xception and DenseNet169 to fine-tune on a target Monkeypox dataset. Further, we extract probabilities from these deep models to feed into the ensemble framework. To combine the outcomes, we propose a Beta function-based normalization scheme of probabilities to learn an efficient aggregation of complementary information obtained from the base learners followed by the sum rule-based ensemble. The framework is extensively evaluated on a publicly available Monkeypox skin lesion dataset using a five-fold cross-validation setup to evaluate its effectiveness. The model achieves an average of 93.39%, 88.91%, 96.78% and 92.35% accuracy, precision, recall and F1 scores, respectively. The supporting source codes are presented in https://github.com/BihanBanerjee/MonkeyPox.

AAIA: an efficient aggregation scheme against inverting attack for federated learning

AAIA: an efficient aggregation scheme against inverting attack for federated learning

An Efficient Aggregation Method for the Symbolic Representation of Temporal Data

Symbolic representations are a useful tool for the dimension reduction of temporal data, allowing for the efficient storage of and information retrieval from time series. They can also enhance the training of machine learning algorithms on time series data through noise reduction and reduced sensitivity to hyperparameters. The adaptive Brownian bridge-based aggregation (ABBA) method is one such effective and robust symbolic representation, demonstrated to accurately capture important trends and shapes in time series. However, in its current form, the method struggles to process very large time series. Here, we present a new variant of the ABBA method, called fABBA. This variant utilizes a new aggregation approach tailored to the piecewise representation of time series. By replacing the k-means clustering used in ABBA with a sorting-based aggregation technique, and thereby avoiding repeated sum-of-squares error computations, the computational complexity is significantly reduced. In contrast to the original method, the new approach does not require the number of time series symbols to be specified in advance. Through extensive tests, we demonstrate that the new method significantly outperforms ABBA with a considerable reduction in runtime while also outperforming the popular SAX and 1d-SAX representations in terms of reconstruction accuracy. We further demonstrate that fABBA can compress other data types such as images.

Blockchain-Enabled Secure and Privacy-Preserving Data Aggregation for Fog-Based ITS

As an essential component of intelligent transportation systems (ITS), electric vehicles (EVs) can store massive amounts of electric power in their batteries and send power back to a charging station (CS) at peak hours to balance the power supply and generate profits. However, when the system collects the corresponding power data, several severe security and privacy issues are encountered. The identity and private injection data may be maliciously intercepted by network attackers and be tampered with to damage the services of ITS and smart grids. Existing approaches requiring high computational overhead render them unsuitable for the resource-constrained Internet of Things (IoT) environment. To address above problems, this paper proposes a blockchain-enabled secure and privacy-preserving data aggregation scheme for fog-based ITS. First, a fog computing and blockchain co-aware aggregation framework of power injection data is designed, which provides strong support for ITS to achieve secure and efficient power injection. Second, Paillier homomorphic encryption, the batch aggregation signature mechanism and a Bloom filter are effectively integrated with efficient aggregation of power injection data with security and privacy guarantees. In addition, the fine-grained homomorphic aggregation is designed for power injection data generated by all EVs, which provides solid data support for accurate power dispatching and supply management in ITS. Experiments show that the total computational cost is significantly reduced in the proposed scheme while providing security and privacy guarantees. The proposed scheme is more suitable for ITS with latency-sensitive applications and is also adapted to deploying devices with limited resources.

Motion robust magnetic resonance imaging via efficient Fourier aggregation.

We present a method for suppressing motion artifacts in anatomical magnetic resonance acquisitions. Our proposed technique, termed MOTOR-MRI, can recover and salvage images which are otherwise heavily corrupted by motion induced artifacts and blur which renders them unusable. Contrary to other techniques, MOTOR-MRI operates on the reconstructed images and not on k-space data. It relies on breaking the standard acquisition protocol into several shorter ones (while maintaining the same total acquisition time) and subsequent efficient aggregation in Fourier space of locally sharp and consistent information among them, producing a sharp and motion mitigated image. We demonstrate the efficacy of the technique on T2-weighted turbo spin echo magnetic resonance brain scans with severe motion corruption from both 3T and 7T scanners and show significant qualitative and quantitative improvement in image quality. MOTOR-MRI can operate independently, or in conjunction with additional motion correction methods.

Optimizing Distributed Face Recognition Systems through Efficient Aggregation of Facial Embeddings

Biometrics are one of the most privacy-sensitive data. Ubiquitous authentication systems with a focus on privacy favor decentralized approaches as they reduce potential attack vectors, both on a technical and organizational level. The gold standard is to let the user be in control of where their own data is stored, which consequently leads to a high variety of devices used. Moreover, in comparison with a centralized system, designs with higher end-user freedom often incur additional network overhead. Therefore, when using face recognition for biometric authentication, an efficient way to compare faces is important in practical deployments, because it reduces both network and hardware requirements that are essential to encourage device diversity. This paper proposes an efficient way to aggregate embeddings used for face recognition based on an extensive analysis on different datasets and the use of different aggregation strategies. As part of this analysis, a new dataset has been collected, which is available for research purposes. Our proposed method supports the construction of massively scalable, decentralized face recognition systems with a focus on both privacy and long-term usability.

Efficiency Evaluation of a Forestry Green Economy under a Multi-Dimensional Output Benefit in China—Based on Evidential Reasoning and the Cross Efficiency Model

The efficiency evaluation of forestry green economy development is related to the direction of forestry development and plays an important role in balancing the economic and environmental issues within that forestry development. The existing research faces three challenges: first, the output indicator is singular; second, the perspective of a self-assessment is extremely limited; and third, the multi perspective fusion method is not in line with the mechanism of the cross efficiency evaluation model. To address these challenges and the characteristics of forestry development output, we constructed multi-level output indicators from four aspects: ecology, economy, society, and sustainability and used evidence reasoning to combine the output indicators. Based on the perspective of a cross evaluation among peers, four different cross efficiency values are defined from the evaluation relationship between the different decision-making units to obtain economic–aggressive, social–neutral, ecological–benevolent, sustainable–neutral, and comprehensive–neutral cross efficiencies. According to the relationship between self- and cross evaluation, an order conditional entropy cross efficiency aggregation model has been proposed and used to analyze the development efficiency of the forestry green economy in 31 Chinese provinces in 2019. Considering the uneven distribution of the forestry resources in China, the development in the 31 provinces and cities is divided into four types by discussing the relationship between the output indicators and efficiency, while the reasons for the unbalanced development and the poor comprehensive development are discussed according to five cross efficiencies.

Using a Au/pitted a-plane GaN substrate to aggregate polar molecules for highly efficient surface-enhanced Raman scattering.

Searching for efficient surface-enhanced Raman scattering (SERS) substrates remains a challenge. In this study, we used metal-organic chemical vapor deposition to directly grow a pitted a-plane GaN thin film, subsequently covered by a thin Au layer (∼25nm), for use as a SERS substrate, without the need for any additional etching or lithography process. The SERS substrate containing these micrometer-sized pits provided a low limit of detection (∼10-9 M) for Rhodamine 6G (R6G), with a high enhancement factor (4.27 × 108) relative to normal Raman spectroscopy. Furthermore, Raman spectral mapping indicated that most of the R6G molecules were concentrated in the pits, enhancing the localization of the probe molecules for further analysis. The same molecular localization phenomenon was also effective for polar methylene blue but not for nonpolar paraffin. The molecular aggregation became more ambiguous upon increasing the thickness of the Au layer, suggesting that the polarity of the Ga and N atoms in the pits was responsible for the efficient aggregation of the polar R6G molecules, which could be potentially beneficial for biomedical detection.

Hydrogen-Induced Aggregation of Au@Pd Nanoparticles for Eye-Readable Plasmonic Hydrogen Sensors.

Plasmonic materials provide a promising platform for optical hydrogen detection, but their sensitivities remain limited. Herein, a new type of eye-readable H2 sensor based on Au@Pd core-shell nanoparticle arrays (NAs) is reported. After exposed to 2% H2, Au@Pd (16/2) NAs demonstrate a dramatic decrease in the optical extinction intensity, along with an obvious color change from turquoise to gray. Experimental results and theoretical calculations prove that the huge optical change resulted from the H2-induced aggregation of Au@Pd nanoparticles (NPs), which remarkably alters the plasmon coupling effect between NPs. Moreover, we optimize the sensing behavior from two aspects. The first is selecting appropriate substrates (either rigid glass substrate or flexible polyethylene terephthalate substrate) to offer moderate adhesion force to NAs, ensuring an efficient aggregation of Au@Pd NPs upon H2 exposure. The second is adjusting the Pd shell thickness to control the extent of NP aggregation and thus the detection range of the as-prepared sensors. This work highlights the advantage of designing eye-readable plasmonic H2 sensors from the aspect of tuning the interparticle plasmonic coupling in NP assemblies. Au@Pd NAs presented here have several advantages in terms of simple fabrication method, eye-readability in air background at room temperature, tunable detection range, and high cost-effectiveness.

SEAR: Secure and Efficient Aggregation for Byzantine-Robust Federated Learning

Federated learning facilitates the collaborative training of a global model among distributed clients without sharing their training data. Secure aggregation, a new security primitive for federated learning, aims to preserve the confidentiality of both local models and training data. Unfortunately, existing secure aggregation solutions fail to defend against Byzantine failures that are common in distributed computing systems. In this work, we propose a new secure and efficient aggregation framework, SEAR, for Byzantine-robust federated learning. Relying on the trusted execution environment, i.e., Intel SGX, SEAR protects clients’ private models while enabling Byzantine resilience. Considering the limitation of the current Intel SGX's architecture (i.e., the limited trusted memory), we propose two data storage modes to efficiently implement aggregation algorithms efficiently in SGX. Moreover, to balance the efficiency and performance of aggregation, we propose a sampling-based method to efficiently detect Byzantine failures without degrading the global model's performance. We implement and evaluate SEAR in a LAN environment, and the experiment results show that SEAR is computationally efficient and robust to Byzantine adversaries. Compared to the previous practical secure aggregation framework, SEAR improves aggregation efficiency by 4-6 times while supporting Byzantine resilience at the same time.

Two-step aggregation of gold nanoparticles based on charge neutralization for detection of melamine by colorimetric and surface-enhanced Raman spectroscopy platform.

A colorimetric and surface-enhanced Raman scattering (SERS) signal amplification platform based on 2-step aggregation of gold nanoparticles (AuNP) was constructed for the sensitive detection of melamine. In this study, the positively charged SYBR Green I was used for the first step of aggregation of AuNP, via charge neutralization, to obtain small-sized AuNP aggregates. The positively charged SYBR Green I decreased the negative charges of the surface of AuNP, which was beneficial to the aggregation of AuNP. In addition, the melamine could aggregate AuNP by decreasing the negative charges of the surface of AuNP and self-assemble with each other on the surface of AuNP by hydrogen bonds. Therefore, the second efficient aggregation of small-sized AuNP aggregates could be achieved with melamine at low concentration, resulting in significant signal changes of color and SERS. The sensitivity of a colorimetric (0.60 mg/L) and SERS (0.089 mg/L) platform, based on 2-step aggregation of AuNP, was 15 and 2.2 times higher than that based on 1-step aggregation of AuNP for detecting melamine.

CO2 Emission Efficiency Analysis of Rail-Water Intermodal Transport: A Novel Network DEA Model

How to evaluate the carbon emission efficiency of multimodal transport is an important issue of public concern, and this article attempts to solve it with a network data envelopment analysis (DEA) model. DEA is a method to evaluate the efficiency of homogeneous decision-making units (DMUs). First, this article studies the efficiency decomposition and efficiency aggregation of the general network structure for DEA model. In efficiency decomposition, the relationship between system efficiency and division efficiency is discussed; whereas in efficiency aggregation, the division tendency brought about by the definition of weights is analyzed. Then, a reasonable and single compromise solution to division efficiency scores is investigated while the system efficiency remains optimal. Finally, a two-stage network DEA model of rail-water intermodal transport is established with carbon dioxide (CO2) emissions as an undesirable output. Based on this model, the rail-water intermodal transport efficiencies of 14 ports in China in 2015 are evaluated by the methods of efficiency decomposition, efficiency aggregation, and non-cooperation. The results show that Rizhao Port, Tangshan Port, Nanjing Port, and Zhuhai Port have set an example to other ports. Qinhuangdao Port, Ningbo-Zhoushan Port, Guangzhou Port, and Beiliang Port need to improve the efficiency of railway transportation. Beibu Gulf port, Zhanjiang Port, Dalian Port, Lianyungang Port, Yantai Port, and Yichang Port should optimize their intermodal system. In addition, Yantai Port and Yichang Port urgently need to improve the port efficiency in low-carbon operation. The network DEA model constructed in this article can be further applied to the efficiency evaluation of multi-link supply chains, and the empirical results can provide a reference for the efficiency evaluation of ports in China.

Efficacy of simultaneous hexavalent chromium biosorption and nitrogen removal by the aerobic denitrifying bacterium Pseudomonas stutzeri YC-34 from chromium-rich wastewater

The impact of high concentrations of heavy metals and the loss of functional microorganisms usually affect the nitrogen removal process in wastewater treatment systems. In the study, a unique auto-aggregating aerobic denitrifier (Pseudomonas stutzeri strain YC-34) was isolated with potential applications for Cr(VI) biosorption and reduction. The nitrogen removal efficiency and denitrification pathway of the strain were determined by measuring the concentration changes of inorganic nitrogen during the culture of the strain and amplifying key denitrification functional genes. The changes in auto-aggregation index, hydrophobicity index, and extracellular polymeric substances (EPS) characteristic index were used to evaluate the auto-aggregation capacity of the strain. Further studies on the biosorption ability and mechanism of cadmium in the process of denitrification were carried out. The changes in tolerance and adsorption index of cadmium were measured and the micro-characteristic changes on the cell surface were analyzed. The strain exhibited excellent denitrification ability, achieving 90.58% nitrogen removal efficiency with 54 mg/L nitrate-nitrogen as the initial nitrogen source and no accumulation of ammonia and nitrite-nitrogen. Thirty percentage of the initial nitrate-nitrogen was converted to N2, and only a small amount of N2O was produced. The successful amplification of the denitrification functional genes, norS, norB, norR, and nosZ, further suggested a complete denitrification pathway from nitrate to nitrogen. Furthermore, the strain showed efficient aggregation capacity, with the auto-aggregation and hydrophobicity indices reaching 78.4 and 75.5%, respectively. A large amount of protein-containing EPS was produced. In addition, the strain effectively removed 48.75, 46.67, 44.53, and 39.84% of Cr(VI) with the initial concentrations of 3, 5, 7, and 10 mg/L, respectively, from the nitrogen-containing synthetic wastewater. It also could reduce Cr(VI) to the less toxic Cr(III). FTIR measurements and characteristic peak deconvolution analysis demonstrated that the strain had a robust hydrogen-bonded structure with strong intermolecular forces under the stress of high Cr(VI) concentrations. The current results confirm that the novel denitrifier can simultaneously remove nitrogen and chromium and has potential applications in advanced wastewater treatment for the removal of multiple pollutants from sewage.

Acquisition, Aggregation, and Sharing of Information in Sequential-Move Aggregative Games

We study quadratic payoff common-value sequential-move aggregative games with signaling. Applications include Stackelberg quantity-setting, public good contribution, common resource, gas emission, and manufacturer-retailer games. Whereas efficient aggregation of information occurs in large Stackelberg markets, it does not occur in the four remaining large markets. If all agents share their information with their followers, expected welfare increases in the Stackelberg and supply chain games and it decreases in the remaining three games. Finally, we characterize the conditions for payoffs and welfare to increase with more precise private or public information. We also provide robustness checks when the values are private. (JEL C70, D43, D82, D83, H41, Q21, Q54)