Abstract Smart surfaces that can dynamically respond to environmental stimuli have demonstrated great promise in wearable electronics and optical detectors. Herein, a photopatternable nanolayered polymeric film reversibly display hide structural colors the visible range response relative humidity (RH) changes is reported. This fabricated on silicon substrate using layer‐by‐layer assembly of chitosan photoreactive carboxymethyl cellulose‐azido derivative, selectively crosslinked through UV...

Hydrogels with multifunctional properties such as biocompatibility, adhesion, and self-healing inherent antibacterial are of great significance in biomedical applications, especially for wound dressings drug release. However, it remains a significant challenge to combine all these one-component hydrogel system. Furthermore, most injectable hydrogels mechanically weak not self-healing, which limits their applications dressings. Herein, we designed tunable two-layer structure that endows the...

Abstract With the rapid development of blockchain technology and increasing demand for partial decentralization Internet, application underlying based on has been widely concerned. Along with decentralized objects, programmable financial system represented by Ethereum gotten more attention. However, smart contract sacrifices its security to improve decentralization. So fatal problem a large number users, negligence users in coding threatens entire network. Therefore, this paper aims research...

In federated learning (FL), malicious clients could manipulate the predictions of trained model through backdoor attacks, posing a significant threat to security FL systems. Existing research primarily focuses on attacks and defenses within generic scenario, where all collaborate train single global model. A recent study conducted by Qin et al. [ 24 ] marks initial exploration personalized (pFL) each client constructs based its local data. Notably, demonstrates that pFL methods with...

Face recognition (FR) has been applied to nearly every aspect of daily life, but it is always accompanied by the underlying risk leaking private information. At present, almost all attack models against FR rely heavily on presence a classification layer. However, in practice, model can obtain complex features input via backbone, and then compare with target for inference, which does not explicitly involve outputs layer adopting logit or other losses. In this work, we advocate novel inference...

With the gradual development of blockchain technology and decentralized demand Internet market, programmable financial system with has been proposed. However, uncertainty smart contract in Ethereum application layer leads to fatal problems on network, which affects efficiency usages. Moreover, most schools or enterprises have not mastered technology, so it is difficult carry out experiments Ethereum, such as simple account transfer transaction deployment contract. The paper research private...

Abstract With the gradual development of big data technology and rapid growth e-commerce industry, comprehensive analysis is particularly important. Therefore, user behavior historical shopping have been focused in recent progress industry. Based on sets, this paper analyzes based Pareto principle (80-20 rule) RFM models. By statistically analyzing characteristics users, groups, preferences, purchase characteristics, several models research algorithms proposed to service targeted products...

The COVID-19 epidemic has swept the world for more than a year. Besides efforts of medical staff to fight outbreak, number researchers at home and abroad have conducted visual analysis data epidemic. They made contributions against mainly in two aspects: situation display prediction. With advance research, countries developed effective vaccines. In this paper, we exploratory on existing vaccine. It reveals types quantities vaccines currently use, shows comparison vaccination from different...

The utilization of personal sensitive data in training face recognition (FR) models poses significant privacy concerns, as adversaries can employ model inversion attacks (MIA) to infer the original data. Existing defense methods, such augmentation and differential privacy, have been employed mitigate this issue. However, these methods often fail strike an optimal balance between accuracy. To address limitation, paper introduces adaptive hybrid masking algorithm against MIA. Specifically,...

Due to the rising privacy concerns on sensitive client data and trained models like Transformers, secure multi-party computation (MPC) techniques are employed enable inference despite attendant overhead. Existing works attempt reduce overhead using more MPC-friendly non-linear function approximations. However, integration of quantization widely used in plaintext into MPC domain remains unclear. To bridge this gap, we propose framework named Ditto efficient quantization-aware Transformer...

Proteases play an essential role in the four sequential but overlapping phases of wound healing: hemostasis, inflammation, proliferation, and remodeling. In chronic wounds, excessive protease secretion damages newly formed extracellular matrix, thereby delaying or preventing normal healing process. Peptide-based fluorogenic sensors provide a visual platform to sense analyze activity through changes fluorescence intensity. Here, we have developed integrated microfluidic chip coated with...

Existing research primarily focuses on backdoor attacks and defenses within the generic federated learning scenario, where all clients collaborate to train a single global model. A recent study conducted by Qin et al. (2023) marks initial exploration of personalized (pFL) each client constructs model based its local data. Notably, demonstrates that pFL methods with \textit{parameter decoupling} can significantly enhance robustness against attacks. However, in this paper, we whistleblow...

Personalized federated learning has gained significant attention as a promising approach to address the challenge of data heterogeneity. In this paper, we relatively unexplored problem in learning. When model been trained and deployed, an unlabeled new client joins, providing personalized for becomes highly challenging task. To challenge, extend adaptive risk minimization technique into unsupervised setting propose our method, FedTTA. We further improve FedTTA with two simple yet effective...

The transferability of adversarial examples can be exploited to launch black-box attacks. However, often present poor transferability. To alleviate this issue, by observing that the diversity inputs boost transferability, input regularization based methods are proposed, which craft combining several transformed inputs. We reveal make resultant biased towards flat extreme regions. Inspired this, we propose an attack called flatness-aware (FAA) explicitly adds a term in optimization target...

Abstract Diabetes mellitus is a common chronic disease with long phase of asymptomatic. This paper focuses on five classification algorithms in machine learning, which are MLP, SVM, KNN, DT and the improved algorithm ANN. By adjusting appropriate parameters for mining analysing diabetes data, classifier effect analysed according to performance indicators accuracy, precision, recall, F1-score. The suitable researched prediction, provides ideas other data from current medical industry.