Trustworthy data recovery for incomplete multi-view learning

Most oblique photography 3D model watermarking algorithms only support limited data recovery or fail to restore the original model, falling short of meeting diverse user needs. Consequently, this study introduces a novel reversible watermarking scheme specifically tailored for oblique photographic 3D models, which is designed to adjust the accuracy of model recovery freely. Firstly, considering the global stability of the oblique photography 3D model, the feature points are extracted by utilizing the mean angle between vertex normals. Secondly, a mapping is established based on the ratio of distances between feature points and non-feature points. Then, the vertices are grouped, with each group consisting of one feature point and several non-feature points. Finally, by using the feature point as the origin, a spherical coordinate system is constructed for each group. The watermark information is embedded by modifying the radius in the spherical coordinate system. In the process of extracting watermarks, watermarks can be extracted from different radius ranges, thereby achieving a controllable error in model recovery. Experimental results demonstrate that this approach exhibits significant advantages in reversibility and controllable restoration accuracy, achieving error-free extraction under both translation and rotation attacks. Compared to existing algorithms, it achieves average improvements of 0.121 and 0.298 under cropping and simplification attacks, respectively, showcasing enhanced robustness. This enables it to meet better diverse user demands for watermarking and model restoration in oblique photography 3D models.

In digital forensics, retrieving data from physically damaged or inaccessible devices is critical for evidence preservation. This study presents a detailed application of the chip-off technique for data recovery from NAND flash memories. The process involves the physical removal of a TSOP-48 NAND chip from a defective SanDisk flash drive, followed by cleaning, preparation, and raw data extraction using the PC-3000 Flash platform. Each stage — including soldering temperature adjustment, pin cleaning, ECC correction, and XOR decoding — is described step by step to ensure repeatability and data integrity. Experimental findings show that using optimal rework parameters (380 °C, 30 % airflow) and proper flux (Amtech NC-559-ASM) allows for complete and error-free data recovery. The study emphasizes the importance of standardized chip-off procedures in forensic laboratories and provides practical guidance for minimizing physical and logical errors during the recovery of NAND-based storage devices.

Innovative PEPS Tensor Network Decomposition for Enhanced Higher Order Data Recovery

BackgroundEstablished in 2021, the SuperAging Research Initiative is a multisite, longitudinal study focused on identifying resilience and resistance factors that promote successful cognitive aging. “SuperAgers” are defined as individuals age 80+ with episodic memory performance that is average or better for individuals 20‐30 years younger. The SuperAging Research Initiative aims to advance knowledge of the neurobiology of brain aging, resilience, and resistance against “typical” age‐related cognitive decline and pathologic declines seen in Alzheimer's disease and related disorders. The SuperAging Research Initiative is focused on increasing racial‐ethnic, geographical, and educational diversity by enrolling 500+ participants across the United States and Canada. The mid‐project recruitment and enrollment success, baseline participant characteristics, and initial study findings from the unique cohort are highlighted.MethodParticipant enrollment and harmonized data collection is ongoing at five North American sites. The protocol includes behavioral, biological, environmental, genetic, and psychosocial characteristics that may contribute to successful cognitive aging. Two embedded Research Projects provide focused opportunities to extend the depth and breadth of science. Project 1 utilizes state‐of‐the‐art wearable technology to obtain quantitative measurements of daily activity, and Project 2 uses transcriptomic, genetic, and protein profiling to examine immune and inflammatory system parameters.ResultAcross sites, >280 participants (ages 80‐101 with 6‐20 years education) have enrolled in the harmonized protocol using community engaged research (CER) strategies. More than 12 states/provinces are represented. To date, approximately 20% participants identify with a historically underrepresented racial‐ethnic group. Sites leveraging existing CER methodology have enrolled a higher percentage of racially diverse participants (30+%). The Project 1 protocol has shown strong feasibility (>90%), yielding high‐quality data (>95% data recovery) for a fully remote sensor data collection protocol. Project 2 has begun initial analyses, and estimates to date suggest SuperAgers have similar Alzheimer's disease polygenic risk scores compared to their cognitively‐average peers.ConclusionThe prospective, longitudinal study of SuperAgers is feasible and provides a unique opportunity to identify mechanisms conferring cognitive resilience and resistance against “typical” and pathologic age‐related cognitive decline. Outcomes may identify novel modifiable factors that promote successful cognitive aging.

Purpose Neural networks are used in diverse applications, making them vulnerable to tampering and reinforcing the need for ownership authentication. The proposed method is based on a steganographic technique that embeds binary information into the weights using the IEEE 754 representation to enhance the security of the neural network and ownership authentication. Design/methodology/approach The proposed method is assessed using a variational autoencoder. Moreover, this technique can be extended to other neural networks. Ownership information is embedded within the most stable layers of the neural network, determined via gradient-based analysis, to enhance robustness against common model alterations, including fine-tuning, compression, pruning, overwriting, noise injection and weight quantization. Findings The experimental results confirm minimal impact on model performance and ensure reliable data recovery. The bit error rate evaluates the robustness of the proposed method, which obtained values ranging from 0.0131 to 0.129 for different weight pruning (10–50%). These results were further corroborated by extensive experimental validation. Originality/value The proposed method introduces a steganographic technique that embeds ownership information using the IEEE 754 representation. Unlike existing techniques, this approach embeds information into the weights without modifying the model structure and maintains the model’s performance without structural changes.

Older adults undergoing abdominal surgery may experience delayed recovery due to changes in physical and psychological aspects. Additionally, the process for effectively involving family members in enhancing recovery for older adults undergoing abdominal surgery in Thailand remains unclear. A quasi-experimental, two-group post-test-only design was employed to investigate the effects of the Recovery Promotion Program with Family Support for older adults undergoing abdominal surgery. The sample consisted of 66 older individuals who had abdominal surgery in the surgical department of a supertertiary hospital, in the lower North of Thailand. The participants in the control group (n = 33) were purposively selected and completed the study first. Then, the participants in the experiment group (n = 33) were recruited by pair matching to ensure similarity in gender, age, and the type of surgery with the participants in the control group. The instruments used to collect data were: a Demographic Data Form, a Convalescence and Recovery Evaluation Form, a Fall Risk Assessment Tool, and a Pressure Ulcer Risk Assessment Tool. Data were summarized with descriptive statistics and analyzed using the Chi-square and independent t-test. The results showed that on day 5 post-surgery, older adults who received the Recovery Promotion Program with Family Support had a significantly better overall recovery, with an extremely large effect size, and also experienced substantially better recovery in terms of pain, gastrointestinal function, emotional status, and activity compared to participants who received routine care alone. In terms of safety, post-intervention analysis revealed no difference between the groups. When comparing the incidence of falls and pressure ulcers, both the experimental and control groups demonstrated 100% safety, with no reported adverse events, such as severe pain, persistent nausea or vomiting, or difficulty breathing. The study findings indicate that the Recovery Promotion Program with Family Support can enhance postoperative recovery in older adults undergoing abdominal surgery. Nurses can apply this intervention in promoting recovery among this population. However, further testing with a multisite study and randomized control is needed before it can be widely used.

BackgroundGuillain-Barré syndrome (GBS) is a life-threatening condition that has been associated with exposure to immune checkpoint inhibitors (ICIs); however, available data remain limited.MethodsWe conducted a retrospective, worldwide, observational analysis of individual case safety reports in VigiBase, the World Health Organization’s pharmacovigilance database. To minimize competition bias, we excluded reports of vaccines and infectious associated with a known or potential risk of GBS. Subsequently, we searched for reports of GBS linked to ICI regimens, whether as monotherapy or dual immunotherapy, from the Food and Drug Administration (FDA) approval date of each agent until 12 February 2024. The primary endpoint of this study was to assess the association between GBS reporting and exposure to ICI regimens (either monotherapy or dual immunotherapy) using disproportionality analysis. This analysis was performed utilizing the Information Component (IC) and its 95% credibility interval lower boundary (IC025).ResultsA total of 412 cases of GBS associated with ICIs were reported in VigiBase. The disproportionality analysis revealed a significant reporting signal between GBS and the anti-CTLA-4 and anti-PD-1 combination therapy (n = 102, IC025 = 4.6), specifically with nivolumab and ipilimumab (n = 100, IC025 = 4.6); anti-CTLA-4 monotherapy (n = 39, IC025 = 3.6) with ipilimumab monotherapy (n = 39, IC025 = 3.6); anti-PD-1 monotherapy (n = 217, IC025 = 3.4), including pembrolizumab (n = 124, IC025 = 3.5), nivolumab (n = 88, IC025 = 3), and cemiplimab (n = 5, IC025 = 1.2); and anti-PD-L1 monotherapy (n = 53, IC025 = 3) with atezolizumab (n = 36, IC025 = 3), durvalumab (n = 11, IC025 = 1.8), and avelumab (n = 6, IC025 = 1.1) in monotherapy. Among cases with available data (n = 123), the median time to onset was 68 days (interquartile range [IQR]: 24.5–119.5), with a shorter delay observed in patients receiving dual immunotherapy compared to those treated with monotherapy. Among the cases for which data was available (n = 242), data recovery or partial recovery was reported in 58.3% (n = 141/242), while a fatal outcome was reported in 9% (n = 21).ConclusionA significant reporting signal of GBS exists with the majority of ICI regimens employed in both monotherapy and dual immunotherapy.Supplementary InformationThe online version contains supplementary material available at 10.1007/s00415-025-13507-0.

This paper describes a compact low-cost telemetry system featuring ready-made sensors and an acquisition unit based on the ESP32, which makes use of the LoRa/Wi-Fi wireless standard for communication, and autonomous fallback logging to guarantee data recovery during communication loss. Ensuring safe atmospheric re-entry requires reliable onboard monitoring of capsule conditions during descent. The system is intended for sub-orbital, low-cost educational capsules and experimental atmospheric descent missions rather than full orbital re-entry at hypersonic speeds, where the environmental loads and communication constraints differ significantly. The novelty of this work is the development of a fully self-contained telemetry system that ensures continuous monitoring and fallback logging without external infrastructure, bridging the gap in compact solutions for CubeSat-scale capsules. In contrast to existing approaches built around UAVs or radar, the proposed design is entirely self-contained, lightweight, and tailored to CubeSat-class and academic missions, where costs and infrastructure are limited. Ground test validation consisted of vertical drop tests, wind tunnel runs, and hardware-in-the-loop simulations. In addition, high-temperature thermal cycling tests were performed to assess system reliability under rapid temperature transitions between −20 °C and +110 °C, confirming stable operation and data integrity under thermal stress. Results showed over 95% real-time packet success with full data recovery in blackout events, while acceleration profiling confirmed resilience to peak decelerations of ~9 g. To complement telemetry, the TeleCapsNet dataset was introduced, facilitating a CNN recognition of descent states via 87% mean Average Precision, and an F1-score of 0.82, which attests to feasibility under constrained computational power. The novelty of this work is twofold: having reliable dual-path telemetry in real-time with full post-mission recovery and producing a scalable platform that explicitly addresses the lack of compact, infrastructure-independent proposals found in the existing literature. Results show an independent and cost-effective system for small re-entry capsule experimenters with reliable data integrity (without external infrastructure). Future work will explore AI systems deployment as a means to prolong the onboard autonomy, as well as to broaden the applicability of the presented approach into academic and low-resource re- entry investigations.

Single-cell RNA sequencing (scRNA-seq) has revolutionized the study of cellular heterogeneity. A major challenge, however, lies in the prevalence of non-biological zeros—false measurements caused by technical limitations that mask a cell’s true transcriptome. This fundamental issue of distinguishing these artifacts from true biological zeros, where a gene is genuinely absent, remains a key hurdle for computational methods, as misclassification can distort biological signals during data recovery. To overcome this, we introduce D3Impute, a discriminative imputation framework built on three key innovations: (1) a distribution-aware normalization step that adapts to dataset-specific characteristics while preserving meaningful biological variation; (2) a dual-network discriminator that uses bulk RNA-seq data as a biological reference to accurately identify non-biological zeros while retaining the true biological zeros; and (3) a density-guided imputation engine that recovers expression values while maintaining local cellular neighborhood structures. Through comprehensive benchmarking against 12 state-of-the-art methods across six diverse datasets, D3Impute demonstrates consistent and significant improvements in essential downstream analyses, including cell clustering, trajectory inference, and differential expression detection. Furthermore, we provide an extensive practical evaluation of D3Impute, demonstrating its robustness across varying data qualities and providing clear guidelines for optimal application. By offering a robust, biologically informed, and user-oriented solution, D3Impute not only enhances scRNA-seq data analysis but also offers a generalizable framework for handling zero-inflated data in computational biology.

Continuous strain missing data recovery with incomplete dataset using geostationary meteorological satellite observations for long-span bridges

A 50 Gb/s clock and data recovery circuit in 45 nm CMOS SOI, SPCLO GF process, for high-speed communication on fiber optics in data centers

The World Wide Web has become one of the most valuable resources for data recovery and information releases since it has the largest collection of data and numerous pages or reports. Advances in web mining are the key to unlocking information on the Internet. There are three types of web mining: web content, web structure, and web usage. One of these classes, Web structure mining is the focus of this research. A significant role in the web mining process is played by web structure mining. This paper discusses the experimental results for Link Based Ranking Algorithms and clarifies Web Mining techniques and certain well-known tactics used in Web structure mining.

Real-time prediction of short-term mortality risk in the intensive care unit (ICU) is often hampered by missing medical data. To address this, we developed RealMIP, an end-to-end framework leveraging generative model for the dynamic imputation of missing values and continuous mortality risk assessment. The model was trained on data from 188 centers in the eICU Collaborative Research Database (eICU-CRD), and internally validated on 20 held-out centers. External validation was performed using the Medical Information Mart for Intensive Care IV (MIMIC-IV) and Salzburg Intensive Care Database (SICdb). RealMIP’s predictive performance was compared with nine established approaches. RealMIP achieved robust predictive performance, with AUCs of 0.957 (95% CI, 0.956–0.957) internally, 0.968 (95% CI, 0.968–0.968) in MIMIC-IV, and 0.932 (95% CI, 0.932–0.933) in SICdb, outperforming comparator models (p < 0.05). RealMIP unlocks the potential of real-time ICU mortality prediction by effectively handling missing data and delivering continuous, interpretable risk assessments.

The widespread adoption of Medical Internet of Things (MIoT) devices, particularly portable electrocardiogram (ECG) monitors, has accelerated since the COVID-19 pandemic, revolutionizing remote patient monitoring and healthcare delivery. However, this rapid integration has introduced significant cybersecurity challenges, especially in securing communication within the MIoT ecosystem. To address these concerns, this study presents a systematic security analysis of three popular portable ECG devices: the Beurer BM 95, KardiaMobile 6L, and OMRON Complete. The investigation begins with a structured literature review to develop a catalog of threats and a threat model specific to the devices’ ecosystem. Guided by this threat model, controlled experiments were conducted to perform penetration testing and security assessments. Our findings reveal multiple security weaknesses and vulnerabilities in the Bluetooth Low Energy (BLE) implementations on these devices, exposing them to potential exploitation and attacks. Additionally, simulated attacks on paired smartphones enabled the recovery of sensitive user and patient data, highlighting further risks within the ecosystem. By uncovering these vulnerabilities, this research highlights the urgent need for stronger security measures in MIoT devices. Addressing these issues proactively is essential to enhance device resilience and protect against emerging threats in connected healthcare environments.

The cessation of regular monitoring of groundwater levels in Ukraine prompts the search for methods for reproducing and predicting the level, which will allow estimating groundwater flow rates, creating models of groundwater resource formation and moisture balance in watersheds. Artificial neural networks (ANN) of various architectures are considered as a data recovery tool for further modeling of water resources. In order to determine the optimal ANN architecture that can simulate the groundwater level (GWL) trend and provide forecasts, the effectiveness of different neural networks (RBF and MLP) in predicting the monthly average GWL was investigated. To select the optimal ANN configuration and assess the effectiveness of each network and its ability to make accurate predictions, the following methods and criteria were used: multiple correlation analysis, spectral analysis of Fourier transforms, wavelet analysis, and component separation by the duration of oscillation cycles. The forecast was made for the average monthly groundwater level from one of the few wells in the Western Bug River basin, for which observations were stopped back in June 2011. The most realistic results using ANN were obtained after isolating short-, medium-, and long-term components in the GWL fluctuations and performing forecasts for the last two components, which is a pioneering step for hydrogeological observations in Ukraine. If for the full (undivided) series of input data it is possible to obtain a forecast/recovery of data with low accuracy up to 4-5 years, then for the medium and long-term components - a more accurate forecast with a sufficiently probable trend up to 11-12 years. Wavelet analysis was used to determine the type of aquifer.

This study aimed to design, develop, and evaluate an Automated Visitors Monitoring System (AVMS) using fingerprint biometric technology to enhance visitor management efficiency and security in the Bureau of Jail Management and Penology (BJMP). A mixed methods approach was employed, integrating quantitative data from user surveys with qualitative insights from interviews and focus group discussions. Pilot testing at the Quezon District Jail in Pagbilao, Quezon involved 20 personnel and 200 visitors, with biometric accuracy evaluated through 918 fingerprint entries. The AVMS achieved 100% accuracy, with high user satisfaction (WAM = 4.88), confirming its reliability and institutional relevance. The AVMS achieved 100% accuracy with zero false acceptance or rejection rates. Users rated the system as “very satisfactory” (WAM = 4.88) and reliability “very reliable” (WAM = 4.88). Respondents noted faster processing times, improved record accuracy, and enhanced security. Suggestions for system refinement included visitor restriction tagging, service provider monitoring, alert notifications, and data recovery functions. The study was limited to a single facility and focused exclusively on fingerprint biometrics, which may affect its applicability in other settings. Future research involving multiple facilities and incorporating other biometric modalities is recommended to strengthen the findings. Nonetheless, the results provide strong evidence of the AVMS’s feasibility and highlight its potential for wider institutional adoption.

Low-rank tensor recovery methods within the tensor singular value decomposition (t-SVD) framework have demonstrated considerable success by leveraging the inherent low-dimensional structures of multi-dimensional data. However, previous approaches in this framework often rely on linear transforms or, in some cases, nonlinear transforms constructed with fully connected networks (FCNs). These methods typically promote a global low-rank structure, which may not fully exploit the nature of multiple subspaces in real-world data. In this work, we propose a nonlinear transform to capture long-range dependencies and diverse patterns across multiple subspaces of the data within the t-SVD framework. This approach provides a richer and more nuanced representation compared to the localized processing typically seen in FCN-based transforms. In the transform domain, we construct a low-rank self-representation layer that fully exploits the multi-subspace structure inherent in tensor data. Instead of merely enforcing overall low-rankness, our method minimizes the nuclear norm of a self-representation tensor, allowing for a more precise and joint characterization of multiple subspaces. This results in a more accurate representation of the data's intrinsic low-dimensional structures, leading to superior recovery performance. This new framework, termed the DEep Low-rank Tensor representAtion (DELTA), is evaluated across several typical multi-dimensional data recovery applications, including tensor completion, robust tensor completion, and spectral snapshot imaging. Experiments on various real-world multi-dimensional data illustrate the superior performance of our DELTA.

ABSTRACT This paper reports a full‐rate reference‐less bang‐bang clock and data recovery (BBCDR) circuit with current mismatch elimination functionality. Specifically, a simplified frequency acquisition loop (FAL) based on lock detection (LD) is proposed to achieve efficient and robust frequency acquisition without the need to determine the polarity of frequency errors. This technique eliminates the need for multiphase clocks and additional high‐speed samplers, significantly saving power and area. In addition, a compact current mismatch elimination circuit is introduced to mitigate the impact of the bang‐bang phase detector (BBPD) metastability characteristic. Prototyped in 28‐nm CMOS, the BBCDR circuit automatically tracks a PRBS‐11 none‐return‐to‐zero (NRZ) input between 28.4 and 30.5 Gb/s, with the total chip area being 0.12 mm 2 . At a rate of 30.5 Gb/s, the peak‐peak jitter of the recovered clock and data are 2.25 and 7.31 , respectively, with a core power efficiency of 1.80 pJ/bit.

An incomplete multiview clustering approach considering missing data recovery based on consistency.