This paper presents a Markov-based system model for microfluidic molecular communication (MC) channels. By discretizing the advection-diffusion dynamics, the proposed model establishes a physically consistent state-space formulation. The transition matrix explicitly captures diffusion, advective flow, reversible binding, and flow-out effects. The resulting discrete-time formulation enables analytical characterization of both transient and equilibrium responses through a linear system representation. Numerical results verify that the proposed framework accurately reproduces channel behaviors across a wide range of flow conditions, providing a tractable basis for the design and analysis of MC systems in microfluidic environments.

Addressing the challenge of the security recognition through electroencephalogram (EEG) biometrics, we propose an EEG recognition system named EEG-FE rrRS. Leveraging a fuzzy extractor, this system aims to facilitate personalized recognition in the scenarios such as the unmanned aerial vehicle (UAV) and metaverse that require human-computer interaction. This robust and reusable fuzzy extractor framework capitalizes on EEG characteristics for biometric identification and it can be divided into two parts: an EEG signal processing module and a proprietary fuzzy extractor scheme containing the secure sketch and the strong extractor. By employing EEG-FE rrRS, a unique digital identity can be established for each user. The security of the proposed fuzzy extractor is proved from the perspective of entropy loss. Furthermore, the simulations have been conducted to evaluate the performance of EEGFE rrRS, showcasing its highly promising recognition accuracy. Specifically, the recognition rate of the system on the motor imagery database has reached 0.92% FRR and 0.08% FAR, respectively. While on the SEED database, the recognition rate has achieved 0% FRR and 0% FAR, respectively.

DNA-based Molecular Communication (MC) has received significant attention for its potential in nanoscale bio-communication systems such as drug delivery, biosensing, and the Internet of Bio-Nano Things (IoBNT). This paper presents a theoretical model for DNA-based MC in spatially confined environments. In the proposed framework, the transmitter (TX) releases complementary DNA (cDNA) molecules in the channel, which reversibly bind to immobilized probe DNAs at the receiver (RX). The stochastic dynamics of diffusion, binding, and release are characterized by a Markov process. Closed-form channel characteristics, including the Channel Impulse Response (CIR), equilibrium distribution, and settling time, are derived. System performance is evaluated through both the proposed model and Particle-Based Simulation (PBS). The results demonstrate that the proposed model effectively describes the dynamic behavior of DNA hybridization-based MC, offering accurate predictions with significantly lower computational complexity than conventional simulation methods.

As one of the most prevalent diseases worldwide, plaque formation in human arteries, known as atherosclerosis, is the focus of many research efforts. Previously, molecular communication (MC) models have been proposed to capture and analyze the natural processes inside the human body and to support the development of diagnosis and treatment methods. In the future, synthetic MC networks are envisioned to span the human body as part of the Internet of Bio-Nano Things (IoBNT), turning blood vessels into physical communication channels. By observing and characterizing changes in these channels, MC networks could play an active role in detecting diseases like atherosclerosis. In this paper, building on previous preliminary work for simulating an MC scenario in a plaque-obstructed blood vessel, we evaluate different analytical models for non-Newtonian flow and derive associated channel impulse responses (CIRs). Additionally, we add the crucial factor of flow pulsatility to our simulation model and investigate the effect of the systole-diastole cycle on the received particles across the plaque channel. We observe a significant influence of the plaque on the channel in terms of the flow profile and CIR across different emission times in the cycle. These metrics could act as crucial indicators for early non-invasive plaque detection in advanced future MC methods.

Graph Neural Network(GNN) have gained extensive adoption in diverse fields, including IoT anomaly detection, social network analysis, and drug molecule prediction, due to their exceptional ability to handle graph-structured data. However, as GNN models become more widely used, the issue of dataset leakage has become increasingly prominent, posing significant risks to the rights and interests of data owners. In this paper, we propose a data ownership authentication method based on GNN backdoor watermarking, termed Graph Data Ownership Authentication(GDOA). Specifically, the data owner injects covert backdoor watermark triggers into some of the samples in the dataset. When an unauthorized user uses the data to train their own model, the model will be injected with the backdoor. The data owner can then validate the target model using samples with the same triggers to determine whether the target model was trained using the unauthorized dataset. GDOA utilizes the clean-label backdoor method to achieve ownership authentication of the dataset due to the greater stealthiness of the clean-label backdoor. Specifically, the watermark samples are selected within the target class, and the samples’ own labels are used as target labels, effectively avoiding the issue of label confusion. In addition, to improve the authentication success rate, GDOA optimizes the injection position of feature triggers in the feature vector. Our extensive experiments across multiple models and benchmark datasets demonstrate that GDOA achieves an average authentication success rate of over 90%, validating its effectiveness for graph data ownership verification.

To obtain accurate brain source activities, the highly ill-posed source imaging of electro- and magneto-encephalography (E/MEG) requires proficiency in incorporation of biophysiological constraints and signal-processing techniques. Here, we propose a spatio-temporal-constrainted E/MEG source imaging framework-STARTS that can reconstruct the source in a fully automatic way. Specifically, a block-diagonal covariance was adopted to reconstruct the source extents while maintain spatial homogeneity. Temporal basis functions (TBFs) of both sources and noise were estimated and updated in a data-driven fashion to alleviate the influence of noises and further improve source localization accuracy. The performance of the proposed STARTS was quantitatively assessed through a series of simulation experiments, wherein superior results were obtained in comparison with the benchmark ESI algorithms (including LORETA, EBI-Convex, BESTIES & SI-STBF). Additional validations on epileptic and resting-state EEG data further indicate that the STARTS can produce neurophysiologically plausible results. Moreover, a computationally efficient version of STARTS: smooth STARTS was also introduced with an elementary spatial constraint, which exhibited comparable performance and reduced execution cost. In sum, the proposed STARTS, with its advanced spatio-temporal constraints and self-adapted update operation, provides an effective and efficient approach for E/MEG source imaging.

Automated medical image segmentation for organs or lesions plays an essential role in clinical diagnoses and treatment plannings. However, training an accurate and robust segmentation model is still a long-standing challenge due to the time-consuming and expertise-intensive annotations for training data, especially 3-D medical images. Recently, self-supervised learning emerges as a promising approach for unsupervised visual representation learning, showing great potential to alleviate the expertise annotations for medical images. Although global representation learning has attained remarkable results on iconic datasets, such as ImageNet, it can not be applied directly to medical image segmentation, because the segmentation task is non-iconic, and the targets always vary in physical scales. To address these problems, we propose a Multi-scale Visual Representation self-supervised Learning (MsVRL) model, to perform finer-grained representation and deal with different target scales. Specifically, a multi-scale representation conception, a canvas matching method, an embedding pre-sampling module, a center-ness branch, and a cross-level consistent loss are introduced to improve the performance. After pre-trained on unlabeled datasets (RibFrac and part of MSD), MsVRL performs downstream segmentation tasks on labeled datasets (BCV, spleen of MSD, and KiTS). Results of the experiments show that MsVRL outperforms other state-of-the-art works on these medical image segmentation tasks.

Synthetic Molecular communications (SMC), as one of the most promising communication paradigms for internet of nano-things (IoNT), is expected to advance many revolutionary areas such as precision drug delivery and biological engineering. Many of the envisioned applications of SMC are in microscale. However, the state-of-the-art SMC testbeds reported in the literature are mostly in macroscale. The lack of microscale communication sub-systems to enable connectivity between individual nanomachines for basic coordination in IoNT is its key technology hindrance. To solve this issue, we propose a microscale SMC modulator which is a key component of microscale SMC system. The proposed microscale SMC modulator is a signal conversion interface to link the macroworld to microworld. It translates an electric signal into biological DNA signal by electrochemical and electrodissolution technique. The modulator is realized by a layer-by-layer assembly technique to immobilize DNA on the surface of gold thin film. And it triggers the selective release of DNA upon the exclusive control of external electric potential signal. The amount of released DNA depends on the amplitude of applied electric stimuli. The DNA release process can be switched off by removing the external electric stimuli and reactivated by reapplying the stimuli. To examine the effectiveness of the proposed modulator, the released DNA concentration is measured by using Nanodrop which is capable to quantify DNA, RNA, and protein samples. And a detection scheme is proposed for signal detection based on the measured data. Experiments show that the proposed setup is able to successfully convert an electric signal representing a sequence of binary symbols into a DNA biological signal with a bit rate of 1 bit/min. This work may help SMC and IoNT to advance from theoretical research towards practical applications.

The use of large-scale machine learning~(ML) is becoming ubiquitous in various domains ranging from business intelligence to self-driving cars. Many companies are building ML pipelines in a unified data processing environment, and leveraging well-tuned numerical optimization packages for obtaining model parameters. However, most existing optimization tools are specifically designed for a single machine setup, and cannot handle vast volume of data. In this work, we build a distributed computing framework towards optimizing generalized linear models with billions of variables. We at first design a new distributed vector to represent data points from extremely large feature space. Then, we introduce an efficient and scalable approach to compute the second order derivatives of loss function, and optimizes model parameters with limited memory requirement. Experiments on real-world datasets demonstrate that our proposed techniques can scale up for ML models with billions of variables, and achieves better performance than state-of-the-art systems on a wide range of applications, e.g., ad CTR prediction and rideshare price bidding.