Recent research has been exploring ways to streamline device pairing by introducingtouch-to-accessthat minimizes user interaction. It generates pairing keys by extracting features from a shared information source to ascertain if two devices are being held by the same person. While these solutions focus on verifying the authenticity of the device, they do not consider the legitimacy and pairing intent of the device holder. Moreover, the pairing keys exchanged over an open wireless link may be susceptible to eavesdropping attacks. In this paper, we propose a secure device pairing mechanism that utilizes the unique electrical responses of the human body to generate and transmit user-specific pairing keys, ensuring both the user's legitimacy and pairing intent while also improving key transmission reliability. We accomplish this by using the device's built-in microphone to capture ambient sound as entropy and converting it into an electrical signal transmitted by the body for device pairing. We have built a prototype and conducted extensive experiments with 31 participants to evaluate its security and usability. The results demonstrate that our proposed mechanism offers a more secure and reliable option for user-specific pairing keys, contributing to the field of device pairing.

Recent efforts in reducing user involvement during device pairing have successfully introduced touch-to-access. To detect whether two devices are being held by the same person, existing touch-to-access solutions extract features from a shared information source to generate pairing keys. They focus on validating the device's authenticity by only requiring the user's simple touching of the device, however, ignore the device holder's legitimacy and pairing intent. Moreover, the pairing keys may be vulnerable to eavesdropping attacks since they are exchanged over an open wireless link (e.g., WiFi or Bluetooth). In this paper, we develop a secure device pairing mechanism that essentially uses the human body to generate and transmit user-specific pairing keys, ensuring the user's legitimacy and pairing intent, as well as improving key transmission reliability. Our work is based on the observation that the human body produces a unique response to the electrical signal flowing through it, and different bodies induce distinct responses to the signal. The built-in microphone on devices captures ambient sound as an entropy source and converts it into an electrical signal, which is subsequently processed and transmitted by the human body for device pairing. We build a prototype using off-the-shelf microphones and conduct extensive experiments with 31 participants to evaluate its security performance and usability. The results show that our system achieves a pairing success rate of 97.74% and an equal error rate of 2.28%.

We present a secure device pairing approach, called BiTouch, using the human body as a conductor to generate and transmit user-specific pairing keys for advancing touch-to-access policy. BiTouch is designed based on the observation that the human body responds uniquely to electrical signals flowing through it. Built-in microphones on devices are essentially used to capture ambient sound as entropy and convert it into an electrical signal, which is subsequently transmitted by the body for device pairing. We implement BiTouch using off-the-shelf microphones and evaluate it with 31 participants. The results demonstrate that BiTouch ensures the user's legitimacy and key transmission reliability, and achieves a pairing success rate of 97.74% and an equal error rate of 2.28%.

Continuous authentication is crucial for protecting user’s privacy throughout their login session. Existing studies employ wireless sensing technologies to provide device-free and unobtrusive authentication; the user’s behavior is continually assessed without their direct involvement until it deviates from their normal pattern. However, these works primarily concentrate on single-user authentication, which poses challenges in multiuser scenarios, such as smart homes and offices, where more than one user usually exists. In this article, we propose HeartPrint, a continuous multiuser authentication system, that employs a single commodity mmWave radar to capture the unique self-driving heartbeat motions from multiple users. Specifically, HeartPrint leverages the effect of skin surface vibrations caused by heartbeat on radio frequency (RF) transmissions. To profile individual heartbeat signals from the entangled components that are induced by multiple users, we first use a clustering method to position each user in the environment, then focus on the signal reflected from each position separately. The irrelevant body movements are eliminated from the RF signal by using a proposed signal energy comparison method for preserving fine-grained heartbeat traits. We then develop a pipeline to extract the most informative features for characterizing each user and feed them to an elaborated classifier for user authentication. We evaluate HeartPrint with 54 participants and demonstrate that it achieves an average authentication accuracy of over 95%. Additionally, we show that it is resilient against spoofing attacks, with an average attack success rate of less than 3%.

The number of single-cell transcriptomic (SCT) studies is rapidly increasing. More than 15000 single cell gene expression data sets are available in public repositories. More than 2400 of these sets involve Peripheral Blood Mononuclear Cells (PBMC) data sets. Main cell types of PBMC are B cells, dendritic cells, monocytes, natural killer cells, and T cells. Labels of individual PBMC are usually provided in metadata accompanying the data sets or are implicit as data set partitions for sorted cells. We analyzed the correctness of labels assigned to individual cells from PBMC in primary reports. The correctness of primary labels was assessed by using Artificial Neural Network (ANN) classifier and Confident Learning (CL) approach. We assessed that the number of mislabels on average in our data sets is about2%. The label accuracy varied broadly between data sets, particularly among those generated by experimental cell sorting followed by SCT.

We performed classification of healthy Peripheral Blood Mononuclear Cells cell types using four methods Artificial Neural Network (ANN), Profiles, Protein Markers (PMs), and RNA markers (RNAMs). Profiles represent patterns of gene expressions characteristic of the subtypes of cells. PMs are protein found exclusively in certain types or subtypes of cells, or represent particular cell states, RNAMs are genes which demonstrate significant differential expressions between cell types. A total of 109 datasets from four different sources containing $\sim$ 120,000 single cells gene expression were used to train and test prediction models. We combined the methods which perform prediction using the whole set of gene features (ANN and Profiles), and those that used specific gene features (PMs and RNAMs) to predict the cell type. The overall classification accuracy was 94.8% for ANN, 94.5% for Profiles, 90.7% for PMs, 67.9% for RNAMs. The combination of four methods showed accuracy of 90.9% with high confidence of positive predictions. The combination of four methods allowed identification of mislabeled cell types in test data sets.

Single cell transcriptomics measures gene expression data of large number of genes, concurrently, from tens of thousands of cells present in a studied biological sample. It is difficult to obtain good classification results due to high data dimensionality and variability of biological states. We performed a preliminary study to assess the feasibility of using supervised machine learning methods to classify peripheral blood mononuclear cell (PBMC) types from single cell gene expression data. We analyzed a large PBMC data set $(\sim 120,000$ PBMC cells), selected 47 genes (from 30698 features) suitable as SML classification features, and performed classification using 20 machine learning algorithms. Data sets represented three sample processing strategies: PBMC separation (two data sets), and experimental cell sorting by (two data sets). The accuracy in 5-class classification among 20 methods was 91-97% (PBMC separation), 97-100% (magnetic-activated cell sorting), and 82-99% (fluorescence-activated cell sorting). Our results indicate the feasibility of supervised machine learning for classification of cells into major PBMC cell types using a small number of classification features from single cell gene expression data.

We implemented an automated system for single-cell classification using artificial neural networks (ANN). Our system takes single-cell gene expression sparse matrices and trains ANN to classify cell types and subtypes. The assemblies of ANNs predict cell classes by voting. We tested the system in a case study where we trained ANNs with a dataset containing approximately 120,000 single cells and tested the resulting model using an independent data set of 13,000 single cells. The overall accuracy of the 5-class classification was 95%. We trained and tested a total of 100 ANNs in 10 cycles. The prediction system demonstrated excellent reproducibility. The analysis of misclassifications indicated that 2% were likely classification errors, while the remaining 3% were likely due to mislabeled types and subtypes in the test set.

We trained artificial neural network (ANN) models to classify peripheral blood mononuclear cells (PBMC) in chronic lymphoid leukemia (CLL) patients. The classification task was to determine differences in gene expression profiles in PBMC pre-treatment (with ibrutinib) and on days 30, 120, 150, and 280 after the start of treatment. Twelve datasets represented clinical samples containing a total 48,016 single cell profiles were used to train and test ANN models to classify the progress of therapy by gene expression changes. The accuracy of ANN classification was $ \gt 92$% in internal cross-validation. External cross-validation, using independent data sets for training and testing, showed the accuracy of classification of post-treatment PBMCs to more than 80%. To the best of our knowledge, this is the first study that has demonstrated the potential of ANNs with 10x single cell gene expression data for detecting the changes during treatment of CLL.

Single cell transcriptomics (SCT) enables high-throughput measurement of mRNA expression concurrently from tens of thousands of single cells. Gene expression profiles in single cells cover only a small fraction of expressed genes and these data are inherently noisy. We developed a method that utilizes artificial neural networks (ANN) for classification of single cells by their tissue of origin. Data sets representing 10 different organs and tissues from C57BL/6 laboratory mice were standardized and used for training and testing ANN models. Each organ was represented by at least two datasets derived from different mice. We achieved 80% accuracy in 10-class classification. After combining data sets from spleen, bone marrow, and lung into one super-class and mammary tissue and muscle into another, we achieved overall cell classification accuracy of 98% across two tissue super-classes and five organs.

We used 27 human single cell transcriptomics (SCT) data sets to develop an artificial neural network (ANN) model for classification of Peripheral Blood Mononuclear Cells (PBMC). We demonstrated that highly accurate models for the classification of PBMC subtypes can be developed by combining multiple independent data sets to form training data sets. A significant data preparation effort was needed for building predictive models. Using a data set of ~120,000 single cell instances we showed the accuracy of classification of PBMC call of ~ 90%. Optimization techniques and the addition of new high-quality data sets for model training are expected to improve PBMC subtype classification accuracy.