Recent advancements in federated learning (FL) have produced models that retain user privacy by training across multiple decentralized devices or systems holding local data samples. However, these ...strategies often neglect the inherent challenges of statistical heterogeneity and vulnerability to adversarial attacks, which can degrade model robustness and fairness. Personalized FL strategies offer some respite by adjusting models to fit individual client profiles, yet they tend to neglect server-side aggregation vulnerabilities. To address these issues, we propose Reinforcement Federated Learning (RFL), a novel framework that leverages deep reinforcement learning to adaptively optimize client contribution during aggregation, thereby enhancing both model robustness against malicious clients and fairness across participants under non-identically distributed settings. To achieve this goal, we propose a meticulous approach involving a Deep Deterministic Policy Gradient-based algorithm for continuous control of aggregation weights, an innovative client selection method based on model parameter distances, and a reward mechanism guided by validation set performance. Empirically, extensive experiments demonstrate that, in terms of robustness, RFL outperforms the state-of-the-art methods, while maintaining comparable levels of fairness, offering a promising solution to build resilient and fair federated systems.
Modulation and templating are two synthetic techniques that have garnered significant attention over the last several years for the preparation of hierarchically porous metal–organic frameworks ...(HP‐MOFs). In this study, by using fatty acids with different lengths and concentrations as dual‐functional modulators/templates, we were able to obtain HP‐MOFs with tunable mesopores that exhibit different pore diameters and locations. We found that the length and concentration of the fatty acids can determine if micelle formation occurs, which in turn dictates the porosity of the resulting HP‐MOFs. The HP‐MOFs with different mesopores differed in their performance in gas uptake and dye adsorption, and the structure–performance relationships were ascribed to the pore diameters and locations. This approach could provide a potentially universal method to efficiently introduce hierarchal mesopores into existing microporous MOF adsorbents with tunable properties.
Das Beste aus beiden Welten: Fettsäuren wurden als duale funktionale Modulatoren/Template verwendet, um hierarchisch poröse Metall‐organische Gerüste (HP‐MOFs) aufzubauen. Die Länge und Konzentration der Fettsäure bestimmt, ob Mizellenbildung und damit die Porosität des resultierenden HP‐MOFs auftreten kann (siehe Bild). HP‐MOFs mit verschiedenen Mesoporen unterschieden sich in ihrer Leistung bei der Gasaufnahme und Farbstoffadsorption.
Heterometallic metal–organic frameworks (MOFs) allow the precise placement of various metals at atomic precision within a porous framework. This new level of control by MOFs promises fascinating ...advances in basic science and application. However, the rational design and synthesis of heterometallic MOFs remains a challenge due to the complexity of the heterometallic systems. Herein, we show that bimetallic MOFs with MX2(INA)4 moieties (INA=isonicotinate; M=Co2+ or Fe2+; X=OH−, Cl−, Br−, I−, NCS−, or NCSe−) can be generated by the sequential modification of a Zr‐based MOF. This multi‐step modification not only replaced the linear organic linker with a square planar MX2(INA)4 unit, but also altered the symmetry, unit cell, and topology of the parent structure. Single‐crystal to single‐crystal transformation is realized so that snapshots for transition process were captured by successive single‐crystal X‐ray diffraction. Furthermore, the installation of Co(NCS)2(INA)4 endows field‐induced slow magnetic relaxation property to the diamagnetic Zr‐MOF.
MOF‐Metamorphose: Heterometall‐organische Gerüstverbindungen wurden durch sequenzielle Modifikation eines Zr‐basierten MOF synthetisiert. Die mehrstufige Modifikation verändert die Symmetrie, Topologie und Elementarzelle, hält aber die Einkristallinität aufrecht. Die nach der Synthese eingeführte Co(NCS)2(Pyridin)4‐Gruppierung verleiht dem diamagnetischen Zr‐MOF eine feldinduzierte langsame magnetische Relaxation.
Chasing Tacit Knowledge: Multi-layered Sensing in Woodworking Nitsche, Michael; Yang, Jialuo; Liu, Yilin Elaine ...
Proceedings of the Eighteenth International Conference on Tangible, Embedded, and Embodied Interaction,
02/2024
Conference Proceeding
We discuss existent approaches to trace tacit knowledge in craft practices through a turn to more-than-human centered design. This leads to a call for a multi-layered sensing system that covers not ...only the human body but also the tools and materials involved. Furthermore, it stresses the necessity of relational data collection and analysis. We present the design and implementation of this concept for a traditional woodworking practice: shaving wood with a block plane. In closing, we show feasibility of the suggested design through an initial pilot test. The argument contributes a theoretical turn to material-based sensing principles in the search for ways to track tacit knowledge, it outlines key design principles of this approach, provides an example implementation, and indicates a first validation through pilot data.
As a privacy-preserving paradigm of decentralized machine learning, federated learning (FL) has become a hot spot in the field of machine learning. Existing FL approaches generally assume that the ...global model can be deployed and trained on any client. However, in practical applications, the devices participated in FL are often heterogeneous and have different computation capacities, resulting in the difficulty of large neural network model training. The current solutions, such as reducing the scale of the global model to fit all clients or removing weak devices to deploy a larger model, will lead to model accuracy degradation, owing to the limitation of model scale or the loss of data on weak clients. To address the device heterogeneity issue inherent in FL, we propose FedBranch, a heterogeneous FL framework based on multi-branch neural network model. Its core idea is to assign a proper branch model to each client according to their computation capacity. In FedBranch, a layer-wise aggregation method is designed to address aggregation of different branches. Meanwhile, we introduce a model regularization method to improve the convergence efficiency and model performance of FedBranch. Besides, we propose a training task offloading algorithm based on Split Learning to safely and effectively share training tasks among different branch models. Extensive experiments conducted on different datasets demonstrate that our FedBranch method has higher convergence efficiency and model accuracy than existing federated learning methods in various heterogeneous scenarios.
Soil monitoring plays an essential role in agricultural systems. Rather than deploying sensors' antennas above the ground, burying them in the soil is an attractive way to retain a non-intrusive ...aboveground space. Low Power Wide-Area Network (LPWAN) has shown its long-distance and low-power features for aboveground Internet-of-Things (IoT) communication, presenting a potential of extending to underground cross-soil communication over a wide area, which however has not been investigated before. The variation of soil conditions brings significant signal polarization misalignment, degrading communication reliability. In this paper, we propose Demeter, a low-cost low-power programmable antenna design to keep reliable cross-soil communication automatically. First, we propose a hardware architecture to enable polarization adjustment on commercial-off-the-shelf (COTS) single-RF-chain LoRa radio. Moreover, we develop a low-power programmable circuit to obtain polarization adjustment. We further design an energy-efficient heuristic calibration algorithm and an adaptive calibration scheduling method to keep signal polarization alignment automatically. We implement Demeter with a customized PCB circuit and COTS devices. Then, we evaluate its performance in various soil types and environmental conditions. The results show that Demeter can achieve up to 11.6 dB SNR gain indoors and 9.94 dB outdoors, 4× horizontal communication distance, at least 20 cm deeper underground deployment, and up to 82% energy consumption reduction per day compared with the standard LoRa.
Based on laser-induced incandescence (LII) and cavity ring-down spectroscopy (CRDS), a measurement setup was established for the study of flame soot particles, and characterize the performance ...parameters. The measurement results of the path-integrated attenuation coefficient of soot particles show that the two-color LII test system and the CRDS system are independent of each other.
As a vehicle for recording and tracking defects, bug report provides basis for solving software quality problems. However, moving fake or duplication bug report in multi-person and parallel software ...testing project is a labor-intensive job. Therefore, this paper proposes a method based on vector space model for automatic dealing with this problem. We built a matching library according to the test requirements and confirmed bug reports and used vector space model to calculate the similarity between the bug report and the matching library. Then the correctness of the bug report is detected based on this similarity.
By comparing the predicted number of defects with the number found in crowdsourced test in real time, people can dynamically assess the progress of crowdsourced test tasks. In this paper, we propose ...a cross-project dynamic defect prediction model (CPDDPM) for crowdsourced test to predict the number of defects in real time. In the construction of training dataset, we use density-based clustering method to select instances from the multiple source project datasets and build the initial training dataset. In the dynamic correction, CPDDPM iteratively corrects the prediction model using crowdsourced test reports and ability attributes of the crowdsourced testers until the predicted results converge. We collected project defect datasets on the crowdsourced test platform, and evaluated prediction accuracy of CPDDPM by using relative error and prediction at level l. The results show that CPDDPM can greatly improve the prediction performance of defect number.
With widely applied in various fields, deep learning (DL) is becoming the key driving force in industry. Although it has achieved great success in artificial intelligence tasks, similar to ...traditional software, it has defects that, once it failed, unpredictable accidents and losses would be caused. In this paper, we propose a test cases generation technique based on an adversarial samples generation algorithm for image classification deep neural networks (DNNs), which can generate a large number of good test cases for the testing of DNNs, especially in case that test cases are insufficient. We briefly introduce our method, and implement the framework. We conduct experiments on some classic DNN models and datasets. We further evaluate the test set by using a coverage metric based on states of the DNN.