edge computing

Cooperative Computation for Multiuser Task Offloading in Wireless-Powered MEC Systems

This paper investigates joint computing and relaying for multiuser task offloading in a wireless-powered mobile edge computing (MEC) system comprising an energy node (EN), an edge server (ES), and multiple energy-harvesting users. One user is selected as the helper for the remaining task users. Each task user partitions its workload among local computing, cooperative computing at the helper, and remote execution at the ES. During a parallel cooperation stage, the helper computes one portion of the uploaded tasks locally while forwarding the remaining portion to the ES and also processes its own task through local computing or edge offloading. The weighted sum computation rate (WSCR) is maximized by jointly optimizing helper selection, task partitioning, time allocation, transmission-energy allocation, and CPU-resource allocation under frame-duration, energy-neutrality, communication, and computation constraints. For each candidate helper, transmission-energy variables are introduced to decouple transmission time and power, and the perspective structure of the achievable-rate functions is exploited to reformulate the continuous resource-allocation problem as an equivalent convex problem. By solving the convex problem for all the candidate helpers, the globally optimal helper selection and resource allocation are obtained. The numerical results show that the proposed joint computing-and-relaying scheme consistently outperforms computing-only, relaying-only, and dedicated-helper cooperation. The performance gain stems from adaptively balancing helper computing and ES processing according to the prevailing communication, computation, and energy bottlenecks.

Reconfigurable terahertz spatial differentiation via incident angle tuned beam shifts on a Bi2Te3 platform

Optical spatial differentiation offers an ultrafast, low-power route to edge detection by directly processing the optical field and can address the inherent diffraction-limited blurring of terahertz (THz) imaging. However, existing THz edge detection schemes face persistent bottlenecks. Metasurfaces and diffractive devices are fixed in function once fabricated, whereas tunable approaches demand stringent angular alignment that limits practical deployment. Here, we theoretically demonstrate a reconfigurable THz spatial differentiator built on the topological insulator Bi2Te3. The device achieves single-parameter mode reconfigurability, enabling seamless switching among one-dimensional (1D) x-direction, 1D y-direction, and isotropic two-dimensional edge enhancement solely by varying the incident angle at a fixed Fermi energy. It also offers favorable parameter robustness, including a broad angular tolerance for 1D x-direction mode and insensitivity to external magnetic fields. Transfer function analysis and edge detection simulations validate the device performance, providing a robust platform for dynamic THz optical computing and high-contrast edge imaging.

Quantum phenomenology from a classical integer automaton: an emergent Schrödinger sector, exact exchange statistics, CHSH violation without randomness, foliation-independent Bell statistics, a selection-rule theorem for ensemble entanglement, a stimulated parametric pair channel with a capacity wall, a surviving isotropy no-go, two engine-measured forced rationals, and obstruction theorems for dynamic geometry

What this is. A single bit-exactly reversible, integer-valued lattice automaton in which a broad range of quantum phenomenology arises as measurable, checksum-reproducible behaviour. Evolution is a permutation of integer microstates, not approximate numerical integration — which is why the gates below can be bit-exact rather than statistical. This is an existence result strengthened by a survived no-go and one falsifiable prediction; it is not a claim that our universe is this automaton. Quantum sector. An emergent Schrödinger sector (envelope bridge with effective mass fixed by the curvature of the exact discrete dispersion; two microscopically different substrates collapsing onto the single universal spreading law √(1+τ²)); deterministic double-slit single-object interference (p<10⁻⁴) with trajectories matching the weak-measurement reconstruction of Kocsis et al. 2011; CHSH = 2.79 from position-valued outcomes with a no-signalling control; a measured branch-coherence decay law (R²=0.997); exact exchange statistics, with Pauli exclusion holding as a bit-conserved invariant for all time; a quantum eraser with literal bit-reversible un-measurement; and partial coarse-grained Born relaxation with exact Loschmidt reversal. Where the quantum–classical boundary lies. The two-particle interaction is substrate-derived (cross-/self-phase ratio 2.31 against a parameter-free prediction of 2), so both terms of the two-particle Hamiltonian are emergent — but the reconstructed state stays Schmidt rank 1 where the L² reference reaches K=4.9. The substrate derives the Hamiltonian, not the L² state space; entanglement is carried, not derived. Spin is implemented. Lorentz structure. Velocity anisotropy vanishes as k² in the infrared with first-principles coefficients in 2D and 3D (≈1%), so rotational isotropy emerges as an infrared fixed point. Bell statistics are foliation-order-independent for causally separated measurement operations (|ΔS|<0.005, S>2 in every ordering) while 41% of individual trajectory pairs flip — a preferred foliation that trajectories know about and no measurement reveals, with the residual order-dependence scaling away as A⁻⁰⋅⁹³. Full boost covariance of states and fields remains open. New in v4 — a dynamic-geometry sector, reported with its obstructions. Six pre-registered pilots ask whether matter can make the local-time geometry and be acted on by it. Three results are derived and measured: (i) an exactly conserved geometric ledger forces the static response to be contact-only — the lattice counterpart of Gauss’s law on a compact space — so no two-body force is possible in that class (hypotheses stated explicitly; this is a no-go for a class, not for discrete gravity in general); (ii) a source entering the first-order equation yields a flat plateau in 1D (depth 46.3 against the parameter-free 45.25, constant in time, exactly zero outside the causal cone), 1/t dilution in 2D and nothing in 3D, so transport cannot build a static sourced field in any dimension; (iii) four independently probed force channels are blocked, the only directional one carrying the winding (electromagnetic) sign. What the ledger does permit. Relaxing conservation of the free field while keeping the total exact restores long range without losing bit-exactness (far/source 0.98–1.06 against ≈0 for contact), and a carry-based binding rule makes mass = number of bound tokens a 1% measurement rather than a framing (rate drift 1.3% while the object’s volume changed 3.05×). The substance carries a second bitwise-exact Gauss law (div e − ΔQ = 0 over 2×10⁴ ticks, flux = enclosed charge exactly) whose charge is the conserved energy. Letting the clocks read it closes the matter→geometry→matter loop kinematically: a body slows its own clock by 30%, matching the clock-rate law to 1.3%, with a causal onset. In 2D the clock field refracts a mobile probe, and a binary, untunable band-edge prediction tabulated before the run is confirmed at both clock ratios (transmission 0.008–0.046 in the predicted reflecting region against 0.58–0.93 just below it); the quantitative refraction law is closed as unproven after two attempts, per its own pre-registration. None of this is general relativity: the geometry is a scalar lapse, no attraction is measured, and the obstruction theorem forbids one in the conserving class. Falsifiable prediction, factor-corrected in v4. The quadratically Planck-suppressed photon dispersion coefficient is measured on the engine rather than derived from a formula: ξiso(3D) = 0.024299 ± 1.1×10⁻⁴ against the exact rational 187/7680 (0.20%), with the subluminal sign resolved from zero by 221 numerical standard errors of the fit. Three coefficients are now stated separately: phase (ξ), group velocity (3ξ) and the Lang convention, δγ,2 = −2ξ/EPl² = −3.27×10⁻⁴⁰ GeV⁻². Our earlier headline quoted the phase coefficient under the Lang symbol; that factor of 2 is the nineteenth self-correction on record and was found by an external automated review, credited as such. The prediction sits ~30× below the older model-dependent Lang et al. (2017) bound, but is in tension (~3.3×) with the conditional Pierre Auger (2022) limit for a source scenario with a subdominant proton component — so it is now decidable by UHECR composition measurements, with a threshold signature at E*≈2 EeV. Its fate depends on the proton fraction at the highest energies, the source model, and the assumption of standard interaction-vertex kinematics, which this substrate does not derive. New in v5 — a selection-rule theorem, a stimulated pair channel, a capacity wall, and a second forced rational. The two-particle boundary is turned into a theorem, attacked adversarially, and mapped constructively. An exact selection rule for the pre-registered phase-ensemble reconstruction (the e−iφ weight pins harmonic pairs to n+m=1) forces any U(1)-equivariant dynamics to factorize exactly — Schmidt rank 1 regardless of interaction strength — with the Schmidt excess bounded at fourth order in the equivariance defects (a one-sided defect buys nothing); the proof passed an adversarial proof-check. A red-team scan of six rank-maximizing mechanisms returned NULL in the autonomous class, and the sharpest impostor — a Floquet drive that initially passed every discriminator — was convicted by the scaling of its rank signal with the integer quantization step itself (423×): the twentieth self-correction, retracted the day it was born. Constructively, a coherent pump prepared in the initial condition of the time-homogeneous update converts through the independently measured cubic coefficient into stimulated, phase-sensitive correlated envelope pairs in two exactly momentum-matched channels — one an umklapp channel whose recoil G = 2π is supplied by the lattice itself — with a parameter-free coupling κ = γP, gain locked to twice the pump phase and scaling as A², and a Manley–Rowe ledger (pump depletion = 2× pairs) closing at the percent level. A capacity scan across pump power, interaction time, conversion-zone length and channel number finds the Schmidt number saturated at the pair (K ≤ 1.12, never more than 2 significant modes, ~65× below the contact L² reference): the substrate makes pairs, not volume — the tensor-product wall stands. A theorem-backed classification fixes where Bell violation can live: configuration-space dynamics realizes it by construction (the earlier S = 2.79 is reread as a confirmation of that class, not an empirical surprise), while for the physically local pumped class per-run CHSH ≤ 2 is a theorem — the driven correlations are classical entanglement: nonseparability without nonlocality. One order deeper in the dispersion, the lattice forces a second exact rational: <ξ₂>3D = 103351/495452160, pre-registered and engine-confirmed to 0.31% (S/N 312), with the first-allowed cos 8θ harmonic detected at S/N 29 at its forced amplitude −1/36864 — a structural forced constant ~56 orders below current bounds, stated as such, not a new testable prediction. A closing section translates the results into renormalization-group vocabulary (isotropy as an IR fixed point; a two-substrate universality class; quantization-floor artefacts as lattice-scale operators) — a dictionary, explicitly not a computed Wilsonian flow. The archive adds the full pre-registrations, engines and per-run JSONs for all four campaigns (~$10 of cloud compute). Methodology as a co-equal contribution. Criteria are committed to version control before each run; the commit hash is the timestamp. Twenty headline results have been retracted or corrected when the project’s own tests exposed artefacts — five in this version. The archive ships the pre-registration designs themselves (specs/), including the correction blocks for changes that failed: a coupling measured to pump energy and reported as “measured, not validated”; a gate left failed on a statistic carrying 0.4% of the signal power; a predicted field inversion that did not occur; and three measuring instruments retired after being shown ill-posed. v4 also discloses and fixes a reproducibility defect in the v3 archive: argparse defaults differed from the published configuration and one report JSON was omitted, so exact command lines are now given verbatim in the README. Prior art. Individual phenomena have known precedents — Philippidis–Dewdney–Hiley 1979 (double-slit Bohmian trajectories); Valentini–Westman 2005 (Born relaxation); Visscher 1991 (the update scheme), with the integer/floor-reversible lifting already noted by Fredkin 1999 and Martin-Delgado 2004; Dürr et al. 1999 (a preferred foliation hidden from equilibrium statistics) — and are framed accordingly. The contribution is their tested, validated realization on one exactly reproducible integer substrate, several new exact constructions, the obstruction results above, and the pre-registered, falsification-first methodology. Produced in an AI-assisted workflow (Anthropic Claude) under th

Deviation-aware digital twin-enabled joint user association and task partitioning for industrial IoT offloading

Latency-critical industrial Internet of Things (IIoT) applications outstrip on-device computing capability, yet digital twin (DT)-assisted multi-access edge computing (MEC) is typically modeled as a perfectly synchronized mirror and treats edge association and task partitioning separately. This paper proposes a deviation-aware three-layer DT-assisted offloading framework spanning IIoT devices, micro base stations (MBSs), and a macro base station (BS). Twins of devices and MBSs are maintained at the BS and explicitly parameterized by CPU-frequency, transmit-power, and bandwidth deviations, so decisions are made from realistically imperfect twin state and the resulting delay-estimation error is quantified. Under delay and energy constraints, discrete user association and continuous partitioning ratios are jointly optimized to minimize average offloading time plus a service-failure penalty, with distinct delay models derived for parallel independent subtasks and serially dependent subtasks. Because the problem is non-convex, NP-hard, and hybrid-variable, a deep multi-agent parameterized Q-network (DMAPQN) is developed in which each device twin acts as an agent and a global mixing network couples local hybrid-action Q-values, preserving the discrete–continuous action structure. Simulations show a 10 MB task completes in 330 ms versus 415, 596, 617, and 705 ms for dichotomy, random, full-MBS, and local execution—a 20.5% gain over the strongest baseline.

Understanding Photon-Counting CT: Physics, Detector Technology, and Image Reconstructions

Photon-counting computed tomography (PCCT) represents a detector-level transformation in CT imaging. Unlike conventional energy-integrating detectors, photon-counting detectors directly convert individual X-ray interactions into electrical pulses and classify them according to energy. This architecture enables electronic-noise rejection, smaller detector pixels, improved geometric dose efficiency, and intrinsic spectral acquisition. However, the images available to radiologists are not produced directly by the detector; energy-resolved photon counts must first undergo calibration, correction, projection formation, reconstruction, and material decomposition. This narrative review provides an educational framework linking X-ray attenuation physics, detector materials and architectures, energy thresholds, and detector nonidealities to the resulting PCCT images. It describes conventional polyenergetic and ultra-high-resolution images, virtual monoenergetic imaging, iodine maps, virtual non-contrast imaging, calcium and bone subtraction, virtual non-calcium imaging, effective atomic number maps, electron-density maps, and emerging K-edge techniques. Particular emphasis is placed on the clinical purpose and limitations of each reconstruction, including noise, artifacts, partial-volume effects, misregistration, incomplete subtraction, calibration dependence, and limited cross-platform comparability. Practical considerations for protocol design, image selection, interpretation workflow, and spectral-data archiving are also discussed. Understanding the pathway from photon detection to image formation is essential for selecting the appropriate reconstruction, avoiding misinterpretation, and integrating PCCT effectively into clinical radiology.

Quantum phenomenology from a classical integer automaton: an emergent Schrödinger sector, exact exchange statistics, CHSH violation without randomness, foliation-independent Bell statistics, a selection-rule theorem for ensemble entanglement, a stimulated parametric pair channel with a capacity wall, a surviving isotropy no-go, two engine-measured forced rationals, and obstruction theorems for dynamic geometry

What this is. A single bit-exactly reversible, integer-valued lattice automaton in which a broad range of quantum phenomenology arises as measurable, checksum-reproducible behaviour. Evolution is a permutation of integer microstates, not approximate numerical integration — which is why the gates below can be bit-exact rather than statistical. This is an existence result strengthened by a survived no-go and one falsifiable prediction; it is not a claim that our universe is this automaton. Quantum sector. An emergent Schrödinger sector (envelope bridge with effective mass fixed by the curvature of the exact discrete dispersion; two microscopically different substrates collapsing onto the single universal spreading law √(1+τ²)); deterministic double-slit single-object interference (p<10⁻⁴) with trajectories matching the weak-measurement reconstruction of Kocsis et al. 2011; CHSH = 2.79 from position-valued outcomes with a no-signalling control; a measured branch-coherence decay law (R²=0.997); exact exchange statistics, with Pauli exclusion holding as a bit-conserved invariant for all time; a quantum eraser with literal bit-reversible un-measurement; and partial coarse-grained Born relaxation with exact Loschmidt reversal. Where the quantum–classical boundary lies. The two-particle interaction is substrate-derived (cross-/self-phase ratio 2.31 against a parameter-free prediction of 2), so both terms of the two-particle Hamiltonian are emergent — but the reconstructed state stays Schmidt rank 1 where the L² reference reaches K=4.9. The substrate derives the Hamiltonian, not the L² state space; entanglement is carried, not derived. Spin is implemented. Lorentz structure. Velocity anisotropy vanishes as k² in the infrared with first-principles coefficients in 2D and 3D (≈1%), so rotational isotropy emerges as an infrared fixed point. Bell statistics are foliation-order-independent for causally separated measurement operations (|ΔS|<0.005, S>2 in every ordering) while 41% of individual trajectory pairs flip — a preferred foliation that trajectories know about and no measurement reveals, with the residual order-dependence scaling away as A⁻⁰⋅⁹³. Full boost covariance of states and fields remains open. New in v4 — a dynamic-geometry sector, reported with its obstructions. Six pre-registered pilots ask whether matter can make the local-time geometry and be acted on by it. Three results are derived and measured: (i) an exactly conserved geometric ledger forces the static response to be contact-only — the lattice counterpart of Gauss’s law on a compact space — so no two-body force is possible in that class (hypotheses stated explicitly; this is a no-go for a class, not for discrete gravity in general); (ii) a source entering the first-order equation yields a flat plateau in 1D (depth 46.3 against the parameter-free 45.25, constant in time, exactly zero outside the causal cone), 1/t dilution in 2D and nothing in 3D, so transport cannot build a static sourced field in any dimension; (iii) four independently probed force channels are blocked, the only directional one carrying the winding (electromagnetic) sign. What the ledger does permit. Relaxing conservation of the free field while keeping the total exact restores long range without losing bit-exactness (far/source 0.98–1.06 against ≈0 for contact), and a carry-based binding rule makes mass = number of bound tokens a 1% measurement rather than a framing (rate drift 1.3% while the object’s volume changed 3.05×). The substance carries a second bitwise-exact Gauss law (div e − ΔQ = 0 over 2×10⁴ ticks, flux = enclosed charge exactly) whose charge is the conserved energy. Letting the clocks read it closes the matter→geometry→matter loop kinematically: a body slows its own clock by 30%, matching the clock-rate law to 1.3%, with a causal onset. In 2D the clock field refracts a mobile probe, and a binary, untunable band-edge prediction tabulated before the run is confirmed at both clock ratios (transmission 0.008–0.046 in the predicted reflecting region against 0.58–0.93 just below it); the quantitative refraction law is closed as unproven after two attempts, per its own pre-registration. None of this is general relativity: the geometry is a scalar lapse, no attraction is measured, and the obstruction theorem forbids one in the conserving class. Falsifiable prediction, factor-corrected in v4. The quadratically Planck-suppressed photon dispersion coefficient is measured on the engine rather than derived from a formula: ξiso(3D) = 0.024299 ± 1.1×10⁻⁴ against the exact rational 187/7680 (0.20%), with the subluminal sign resolved from zero by 221 numerical standard errors of the fit. Three coefficients are now stated separately: phase (ξ), group velocity (3ξ) and the Lang convention, δγ,2 = −2ξ/EPl² = −3.27×10⁻⁴⁰ GeV⁻². Our earlier headline quoted the phase coefficient under the Lang symbol; that factor of 2 is the nineteenth self-correction on record and was found by an external automated review, credited as such. The prediction sits ~30× below the older model-dependent Lang et al. (2017) bound, but is in tension (~3.3×) with the conditional Pierre Auger (2022) limit for a source scenario with a subdominant proton component — so it is now decidable by UHECR composition measurements, with a threshold signature at E*≈2 EeV. Its fate depends on the proton fraction at the highest energies, the source model, and the assumption of standard interaction-vertex kinematics, which this substrate does not derive. New in v5 — a selection-rule theorem, a stimulated pair channel, a capacity wall, and a second forced rational. The two-particle boundary is turned into a theorem, attacked adversarially, and mapped constructively. An exact selection rule for the pre-registered phase-ensemble reconstruction (the e−iφ weight pins harmonic pairs to n+m=1) forces any U(1)-equivariant dynamics to factorize exactly — Schmidt rank 1 regardless of interaction strength — with the Schmidt excess bounded at fourth order in the equivariance defects (a one-sided defect buys nothing); the proof passed an adversarial proof-check. A red-team scan of six rank-maximizing mechanisms returned NULL in the autonomous class, and the sharpest impostor — a Floquet drive that initially passed every discriminator — was convicted by the scaling of its rank signal with the integer quantization step itself (423×): the twentieth self-correction, retracted the day it was born. Constructively, a coherent pump prepared in the initial condition of the time-homogeneous update converts through the independently measured cubic coefficient into stimulated, phase-sensitive correlated envelope pairs in two exactly momentum-matched channels — one an umklapp channel whose recoil G = 2π is supplied by the lattice itself — with a parameter-free coupling κ = γP, gain locked to twice the pump phase and scaling as A², and a Manley–Rowe ledger (pump depletion = 2× pairs) closing at the percent level. A capacity scan across pump power, interaction time, conversion-zone length and channel number finds the Schmidt number saturated at the pair (K ≤ 1.12, never more than 2 significant modes, ~65× below the contact L² reference): the substrate makes pairs, not volume — the tensor-product wall stands. A theorem-backed classification fixes where Bell violation can live: configuration-space dynamics realizes it by construction (the earlier S = 2.79 is reread as a confirmation of that class, not an empirical surprise), while for the physically local pumped class per-run CHSH ≤ 2 is a theorem — the driven correlations are classical entanglement: nonseparability without nonlocality. One order deeper in the dispersion, the lattice forces a second exact rational: <ξ₂>3D = 103351/495452160, pre-registered and engine-confirmed to 0.31% (S/N 312), with the first-allowed cos 8θ harmonic detected at S/N 29 at its forced amplitude −1/36864 — a structural forced constant ~56 orders below current bounds, stated as such, not a new testable prediction. A closing section translates the results into renormalization-group vocabulary (isotropy as an IR fixed point; a two-substrate universality class; quantization-floor artefacts as lattice-scale operators) — a dictionary, explicitly not a computed Wilsonian flow. The archive adds the full pre-registrations, engines and per-run JSONs for all four campaigns (~$10 of cloud compute). Methodology as a co-equal contribution. Criteria are committed to version control before each run; the commit hash is the timestamp. Twenty headline results have been retracted or corrected when the project’s own tests exposed artefacts — five in this version. The archive ships the pre-registration designs themselves (specs/), including the correction blocks for changes that failed: a coupling measured to pump energy and reported as “measured, not validated”; a gate left failed on a statistic carrying 0.4% of the signal power; a predicted field inversion that did not occur; and three measuring instruments retired after being shown ill-posed. v4 also discloses and fixes a reproducibility defect in the v3 archive: argparse defaults differed from the published configuration and one report JSON was omitted, so exact command lines are now given verbatim in the README. Prior art. Individual phenomena have known precedents — Philippidis–Dewdney–Hiley 1979 (double-slit Bohmian trajectories); Valentini–Westman 2005 (Born relaxation); Visscher 1991 (the update scheme), with the integer/floor-reversible lifting already noted by Fredkin 1999 and Martin-Delgado 2004; Dürr et al. 1999 (a preferred foliation hidden from equilibrium statistics) — and are framed accordingly. The contribution is their tested, validated realization on one exactly reproducible integer substrate, several new exact constructions, the obstruction results above, and the pre-registered, falsification-first methodology. Produced in an AI-assisted workflow (Anthropic Claude) under th

CReIS: Computation Reuse through image similarity in ICN-based edge computing

Telemetry Dataset for Federated Learning on Heterogeneous Edge Devices Using Raspberry Pi 5 and NVIDIA Jetson Nano

This dataset accompanies the manuscript "Telemetry-aware federated learning on heterogeneous edge devices: an experimental study". The dataset contains telemetry measurements collected from a federated learning testbed consisting of a Raspberry Pi 5 and an NVIDIA Jetson Nano running the Flower federated learning framework. Telemetry was sampled at one-second intervals and includes CPU utilization, CPU frequency, CPU temperature, memory usage, disk activity, network activity, system load, process counts, and Jetson-specific GPU metrics where available. The experiments comprise four stages: baseline system stress profiling, local model training on MNIST, Fashion-MNIST, and CIFAR-10 datasets, federated learning experiments under varying local-epoch settings, and controlled hardware throttling experiments to investigate device heterogeneity. The dataset supports reproducibility of the statistical analyses presented in the associated manuscript and may be useful for research on federated learning, edge computing, telemetry analytics, resource-aware scheduling, heterogeneous computing, and performance characterization of embedded AI platforms. The data are provided in Microsoft Excel format with accompanying metadata and are intended for academic and research use.

Telemetry Dataset for Federated Learning on Heterogeneous Edge Devices Using Raspberry Pi 5 and NVIDIA Jetson Nano

This dataset accompanies the manuscript "Telemetry-aware federated learning on heterogeneous edge devices: an experimental study". The dataset contains telemetry measurements collected from a federated learning testbed consisting of a Raspberry Pi 5 and an NVIDIA Jetson Nano running the Flower federated learning framework. Telemetry was sampled at one-second intervals and includes CPU utilization, CPU frequency, CPU temperature, memory usage, disk activity, network activity, system load, process counts, and Jetson-specific GPU metrics where available. The experiments comprise four stages: baseline system stress profiling, local model training on MNIST, Fashion-MNIST, and CIFAR-10 datasets, federated learning experiments under varying local-epoch settings, and controlled hardware throttling experiments to investigate device heterogeneity. The dataset supports reproducibility of the statistical analyses presented in the associated manuscript and may be useful for research on federated learning, edge computing, telemetry analytics, resource-aware scheduling, heterogeneous computing, and performance characterization of embedded AI platforms. The data are provided in Microsoft Excel format with accompanying metadata and are intended for academic and research use.

Smart aquaponics: trends, challenges, and future directions

Abstract Smart aquaponics couples recirculating aquaculture with hydroponic plant cultivation, and a substantial body of recent research applies IoT sensing, machine learning, and edge computing to this domain. The literature, however, lacks a synthesis that maps how predictive models, system architectures, and biological context combine in deployed systems, leaving researchers and practitioners without a clear technical roadmap. This systematic literature review (SLR) addresses this gap through 10 research questions following the Kitchenham and Charters protocol and PRISMA-style screening. The search initially identified 3,123 records from six bibliographic databases and snowballing; after duplicate removal, screening, and quality assessment, 49 primary studies were retained for analysis. First, the literature has a prediction-to-control gap: 24% of the studies report forecasters or classifiers without specifying how the resulting prediction is consumed by an actuator, leaving inference layers technically ahead of control layers. Second, system architectures are transitioning from centralised cloud designs toward federated and edge configurations to address privacy, latency, bandwidth, and sensor-drift issues. Third, hybrid models incorporating physical, or domain knowledge show promising performance under non-stationary aquaponics conditions. Fourth, long-term and multi-site field deployments remain very limited, making it difficult to interpret reported machine-learning accuracy as evidence of commercial readiness. Overall, future progress requires stronger integration of prediction, control, calibration, biological benchmarking, and operator usability.

Blockchain fog attention framework for collusion detection and automated accountability in internet of things networks

Fog–edge Internet of Things (IoT) systems support low-latency distributed services but remain vulnerable to coordinated attacks that evade detectors designed for independent events. Existing solutions also tend to separate attack detection, provenance verification, and accountability enforcement, leaving no unified path from relational evidence to auditable response. This study aims to develop an integrated framework that detects coordinated malicious behavior, ranks provenance relevance, verifies evidence, and activates rule-based accountability in resource-constrained edge environments. The proposed Blockchain-Fog Computing Collaborative Framework with Deep Attention-based Collusion Detection and Automated Accountability (BF3-ACDA) framework combines a hierarchical blockchain–fog architecture with an attention-based collusion graph neural network (AttnCol-GNN), whose reputation-aware attention coefficient incorporates behavioral correlation and blockchain-derived trust information. A shared attention representation supports both collusion classification and provenance ranking, while Fog-BFT consensus, Merkle verification, and smart contracts provide tamper-evident recording and severity-based enforcement. Across 30 matched independent runs on the collusion-augmented CICIoT2023 benchmark, BF3-ACDA achieved 96.80 ± 0.23% accuracy, 96.75 ± 0.21% F1-score, and 0.975 ± 0.005 AUC-ROC. Direct-transfer accuracy without target-domain fine-tuning was 94.20 ± 0.34% on NSL-KDD and 92.50 ± 0.25% on CICIDS2017. Removing detector-side reputation and verification features reduced accuracy by 3.40 percentage points, whereas replacing learned attention with mean aggregation reduced it by 1.80 points. The INT8 edge model required 12.4 MB and 58.5 ± 3.9 ms per inference. The results demonstrate a method-level coupling of coordinated-pattern detection, provenance relevance, and auditable enforcement, while supporting prototype deployment on evaluated edge, fog, and cloud platforms. As the collusion metadata were constructed for this study, the findings characterize robustness under controlled conditions rather than field performance on naturally occurring collusion.

Interpretable Digital Twin-Safe Scheduling for Deadline-Aware Task Allocation in Dynamic Fog-Edge Computing: A Comparative and Prediction-Validity Evaluation