The "replication crisis" is a methodological problem in which many scientific research findings have been difficult or impossible to replicate. Because the reproducibility of empirical results is an essential aspect of the scientific method, such failures endanger the credibility of theories based on them and possibly significant portions of scientific knowledge. An instance of the replication crisis, analytic replication, pertains to reproducing published results through computational reanalysis of the authors' original data. However, direct replications are costly, time-consuming, and unrewarded in today's publishing standards. We propose that bioinformatics and computational biology students replicate recent discoveries as part of their curriculum. Considering the above, we performed a pilot study in one of the graduate-level courses we developed and taught at our University. The course is entitled Intro to R Programming and is meant for students in our Master's and PhD programs who have little to no programming skills. As the course emphasized real-world data analysis, we thought it would be an appropriate setting to carry out this study. The primary objective was to expose the students to real biological data analysis problems. These include locating and downloading the needed datasets, understanding any underlying conventions and annotations, understanding the analytical methods, and regenerating multiple graphs from their assigned article. The secondary goal was to determine whether the assigned articles contained sufficient information for a graduate-level student to replicate its figures. Overall, the students successfully reproduced 39% of the figures. The main obstacles were the need for more advanced programming skills and the incomplete documentation of the applied methods. Students were engaged, enthusiastic, and focused throughout the semester. We believe that this teaching approach will allow students to make fundamental scientific contributions under appropriate supervision. It will teach them about the scientific process, the importance of reporting standards, and the importance of openness.

Ciphertext-policy attribute-based encryption (CP- ABE) with verifiable outsourced decryption is a mechanism for secure fine-grained access control over encrypted data, and it is suitable for cloud computing applications. However, there exists a risk in CP-ABE with verifiable outsourced decryption that can lead to serious consequences and may limit its wide applications: the key generation center may have misbehavior. In this paper, we present BSA, the blockchain-based specification for ABE to mitigate this risk. We introduce the specification to regulate the data access control and a proof mechanism to supervise whether the key generation center has misbehavior. Also, we can provide decentralized and automated incentives with BSA by smart contracts and blockchain-based consensus.

Trust evaluation of people and information on Twitter is critical for maintaining a healthy online social environment. How to evaluate the trustworthiness of users and tweets becomes a challenging question. In this demo, we show how our proposed CoTrRank approach deal with this problem. This approach models users and tweets in two coupled networks and calculate their trust values in different trust spaces. In particular, our solution provides a configurable way when mapping the calculated raw evidences to the trust values. The CoTrRank demo system has an interactive interface to show how our proposed approach produces more effective and adaptive trust evaluation results comparing with baseline methods.

Atrial Fibrillation (AF) is considered to strongly correlate with stroke. Deep Neural Networks (DNNs) improve the accuracy in real-time atrial fibrillation detection. However, the deployment of DNNs on embedded systems is challenging due to hardware resources. To reduce computation loads, we study the feasibility to eliminate the redundant information in the AF detection task by downsampling. It is compatible with kernel-level optimization, quantization optimization, and model compression methods. A state-of-the-art deep learning model is used to estimate the amount of AF detection information among different sampling rates. This work considers both fixed-length and variable-length time intervals of an Electrocardiograph (ECG) segment. Experiment results demonstrate that model performance can be retained perfectly in AF detection. Ablation study experiments demonstrate the robustness with downsampled signals. Using a large time interval, the AF detection accuracy with 60 Hz signals can be compared to that with 300 Hz signals. Our on-going work includes designing lightweight DNN models with downsampled signals, further exploring the robustness of downsampled signals and model compression.

Sleep period determination is essential to accurate sleep quality analysis. In this paper, we propose an automated algorithm for wearable multi-sensor systems to precisely estimate the sleep period. It leverages the information of accelerometer, and vital signs such as heart rate and breathing rate. Compared to the sleep periods determined by the clinical diagnosing-grade equipment, our algorithm achieves average time differences of 9.0 and 10.4 minutes for healthy subjects and clinical patients, respectively.

Conventional diagnosis using polysomnography (PSG) on breathing disorder is expensive and uncomfortable to patients. In this paper, we present a low-cost portable and wearable multi-sensor system to non-invasively acquire a subject's vital signs, and leverage various machine learning methods on features extracted from Electrocardiogram (ECG) and Blood oxygen saturation (SpO2) signals to detect breathing disorder events. Our preliminary predication accuracies on 110 clinical patients is 90.0%.

Human tracking is an important vision task in video surveillance and perceptual human-computer interfaces. This paper presents a novel algorithm for region-based human tracking using color and depth features. We propose an adaptive autoregressive logarithmic search (ARLS) to estimate the target position, and use depth information to further reduce the false alarm rate. The new ARLS algorithm is evaluated on a color and depth (RGBD) video dataset acquired with the Kinect sensor. The dataset contains various real-world scenarios with illumination and speed variations, and partial occlusion. The experimental results show that the ARLS algorithm is able to handle difficult tracking scenarios, it achieves a tracking accuracy of 91.26% on the test dataset. The proposed algorithm is compared with two tracking algorithms, namely the particle filtering and a modified logarithmic search algorithm.

Facial expression is one way humans convey their emotional states. Accurate recognition of facial expressions is essential in perceptual human-computer interface, robotics and mimetic games. This paper presents a novel approach to facial expression recognition from static images that combines fixed and adaptive 2-D filters in a hierarchical structure. The fixed filters are used to extract primitive features. They are followed by the adaptive filters that are trained to extract more complex facial features. Both types of filters are non-linear and are based on the biological mechanism of shunting inhibition. The features are finally classified by a support vector machine. The proposed approach is evaluated on the JAFFE database with seven types of facial expressions: anger, disgust, fear, happiness, neutral, sadness and surprise. It achieves a classification rate of 96.7%, which compares favorably with several existing techniques for facial expression recognition tested on the same database.