Advanced Digital Sciences Center

This paper develops an automated approach for conjunctival hyperemia grading from slit-lamp images using semisupervised learning. We conducted a retrospective study including slit-lamp images from two study sites. Two independent graders assessed the severity of hyperemia according to the Efron Grading Scales. Segmentation of the conjunctiva and its vessels was performed using semisupervised segmentation with limited labeled data. Conjunctival vessel densities were estimated from the model outputs and compared against the manual clinical Efron gradings. Three hundred and seventeen slit-lamp images from the primary site and 164 from an external site were included. The semisupervised models with unlabeled data demonstrated significantly improved segmentation compared to a baseline fully supervised model using only the labeled data (p < 0.001). Calculated conjunctival vessel densities showed correlations of 0.86 [0.76, 0.93] with ground truth vessel densities. Comparisons of vessel densities against mean manual clinical Efron gradings showed correlations of 0.83 and 0.80 for the test and external datasets, which were comparable to the inter-rater agreements of 0.82 [0.68, 0.90] and 0.75 [0.67, 0.81] in the datasets, respectively. Conjunctival vessel densities obtained with semisupervised learning showed good agreement with clinical grading of conjunctival hyperemia. This approach may be applied toward an automatic, objective assessment of the conjunctiva.

Hybrid energy storage system (HESS) is effective to compensate for fluctuation power in renewables and fast fluctuation loads in DC microgrids. To regulate DC bus voltage, a power management strategy is an essential issue. In the meantime, the increasing integration of constant power loads (CPLs) in DC microgrids brings great challenges to stable operation due to their negative incremental impedance. In this paper, a fast composite backstepping control (FBC) method is proposed for the HESS to achieve faster dynamics, smaller voltage variations, and large-signal stabilization. In the FBC method, a higher order sliding mode observer (HOSMO) is adopted to estimate the coupled disturbances. Furthermore, the FBC method is integrated with the droop control; so that the FBC-based decentralized power allocation (FBC-DPA) strategy for HESS in DC microgrids is developed. The proposed FBC method is designed based on the Lyapunov function to ensure its stability. Moreover, the design guidelines are provided to facilitate the application of the proposed method. Both simulation and experimental studies under different operating scenarios show that the proposed method achieves faster voltage recovery and smaller voltage variations than the conventional backstepping control method.

With the widespread use of Internet of Things (IoT) devices, malware detection has become a hot spot for both academic and industrial communities. Existing approaches can be roughly categorized into network-side and host-side. However, existing network-side methods are difficult to capture contextual semantics from cross-source traffic, and previous host-side methods could be adversary-perceived and expose risks for tampering. More importantly, a single perspective cannot comprehensively track the multi-stage lifecycle of IoT malware. In this paper, we present <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math>${\sf CMD}$</tex-math></inline-formula> , a co-analyzed IoT malware detection and forensics system by combining hardware and network domains. For the network part, <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math>${\sf CMD}$</tex-math></inline-formula> proposes a tailored capsule neural network to capture the contextual semantics from cross-source traffic. For the hardware part, <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math>${\sf CMD}$</tex-math></inline-formula> designs an entire file operation recovery process in a side-channel manner by leveraging the Serial Peripheral Interface (SPI) signals from on-chip traces. These traffic provenance and operating logs information could benefit the anti-virus countermeasures for security practitioners. By practical evaluation, we demonstrate that <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math>${\sf CMD}$</tex-math></inline-formula> realizes outstanding detection effects ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e.g.,</i> <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\sim$</tex-math></inline-formula> 99.88% F1-score) compared with seven state-of-the-art methods, and recovers 96.88% <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\sim$</tex-math></inline-formula> 99.75% operation commands even if against adaptive adversaries (that could kill processes or tamper with operation log files). A by-product benefit of such an external monitor is <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math>${\sf CMD}$</tex-math></inline-formula> introduces zero latency on the IoT device, and incurs negligible IoT CPU utilization. Also, since SPI focuses on file operations, the proposed hardware trace forensics does not have the data explosion problem like previous work, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e.g.,</i> recovered logs of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math>${\sf CMD}$</tex-math></inline-formula> only take up limited extra space overhead ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e.g.,</i> <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\sim$</tex-math></inline-formula> 0.2 MB per malware). Furthermore, we provide the model interpretability for the capsule network and develop a case study (Hajime) of the operation logs recovery.

Ceramic membranes are taking center stage for separation technologies in water treatment. Among them, ceramic nanofiltration membranes are at the forefront of membrane technologies. The desalination of seawater using ceramic nanofiltration membranes is a potential application toward increasing the global water supply and tackling water scarcity. However, while the high fabrication cost poses a challenge to their large-scale applications, high-value separation applications can help to offset the overall cost. In this regard, ceramic nanofiltration membranes can also be explored as a viable option for high-value lithium extraction from the waste seawater brine. In order to determine the potential of nanofiltration ceramic membranes for desalination and lithium recovery from seawater, the current efficiency of salt rejection across various operation parameters must be thoroughly evaluated. Specifically, the interactions between the Donnan exclusion, steric exclusion, zeta potential, and salt concentration play an important role in determining the salt rejection efficiency. Several strategies are then proposed to guide ceramic nanofiltration membranes toward potentially practical applications regarding desalination and lithium recovery.

Traditional testing of Anti-Virus (AV) products is usually performed on a curated set of malware samples. While this approach can evaluate an AV's overall performance on known threats, it fails to provide details on the coverage of exact attack techniques used by adversaries and malware. Such coverage information is crucial in helping users understand potential attack paths formed using new code and combinations of known attack techniques.