Search

The paper develops datasets and methods to assess student participation in real-life collaborative learning environments. In collaborative learning environments, students are organized into small groups where they are free to interact within their group. Thus, students can move around freely causing issues with strong pose variation, move out and re-enter the camera scene, or face away from the camera. We formulate the problem of assessing student participation into two subproblems: (i) student group detection against strong background interference from other groups, and (ii) dynamic participant tracking within the group. A massive independent testing dataset of 12,518,250 student label instances, of total duration of 21 hours and 22 minutes of real-life videos, is used for evaluating the performance of our proposed method for student group detection. The proposed method of using multiple image representations is shown to perform equally or better than YOLO on all video instances. Over the entire dataset, the proposed method achieved an F1 score of 0.85 compared to 0.80 for YOLO. Following student group detection, the paper presents the development of a dynamic participant tracking system for assessing student group participation through long video sessions. The proposed dynamic participant tracking system is shown to perform exceptionally well, missing a student in just one out of 35 testing videos. In comparison, a state of the art method fails to track students in 14 out of the 35 testing videos. The proposed method achieves 82.3% accuracy on an independent set of long, real-life collaborative videos.

We use extreme value theory to estimate the probability of successive exceedances of a threshold value of a time-series of an observable on several classes of chaotic dynamical systems. The observables have either a Fr\'echet (fat-tailed) or Weibull (bounded) distribution. The motivation for this work was to give estimates of the probabilities of sustained periods of weather anomalies such as heat-waves, cold spells or prolonged periods of rainfall in climate models. Our predictions are borne out by numerical simulations and also analysis of rainfall and temperature data.

Polarons and spin-orbit (SO) coupling are distinct quantum effects that play a critical role in charge transport and spin-orbitronics. Polarons originate from strong electron-phonon interaction and are ubiquitous in polarizable materials featuring electron localization, in particular $\mathrm{3d}$ transition metal oxides (TMOs). On the other hand, the relativistic coupling between the spin and orbital angular momentum is notable in lattices with heavy atoms and develops in $\mathrm{5d}$ TMOs, where electrons are spatially delocalized. Here we combine ab initio calculations and magnetic measurements to show that these two seemingly mutually exclusive interactions are entangled in the electron-doped SO-coupled Mott insulator $\mathrm{Ba_2Na_{1-x}Ca_xOsO_6}$ ($0< x < 1$), unveiling the formation of spin-orbital bipolarons. Polaron charge trapping, favoured by the Jahn-Teller lattice activity, converts the Os $\mathrm{5d^1}$ spin-orbital $\mathrm{J_{eff}=3/2}$ levels, characteristic of the parent compound $\mathrm{Ba_2NaOsO_6}$ (BNOO), into a bipolaron $\mathrm{5d^2}$ $\mathrm{J_{eff}=2}$ manifold, leading to the coexistence of different J-effective states in a single-phase material. The gradual increase of bipolarons with increasing doping creates robust in-gap states that prevents the transition to a metal phase even at ultrahigh doping, thus preserving the Mott gap across the entire doping range from $\mathrm{d^1}$ BNOO to $\mathrm{d^2}$ $\mathrm{Ba_2CaOsO_6}$ (BCOO).

Early detection and localization of myocardial infarction (MI) can reduce the severity of cardiac damage through timely treatment interventions. In recent years, deep learning techniques have shown promise for detecting MI in echocardiographic images. However, there has been no examination of how segmentation accuracy affects MI classification performance and the potential benefits of using ensemble learning approaches. Our study investigates this relationship and introduces a robust method that combines features from multiple segmentation models to improve MI classification performance by leveraging ensemble learning. Our method combines myocardial segment displacement features from multiple segmentation models, which are then input into a typical classifier to estimate the risk of MI. We validated the proposed approach on two datasets: the public HMC-QU dataset (109 echocardiograms) for training and validation, and an E-Hospital dataset (60 echocardiograms) from a local clinical site in Vietnam for independent testing. Model performance was evaluated based on accuracy, sensitivity, and specificity. The proposed approach demonstrated excellent performance in detecting MI. The results showed that the proposed approach outperformed the state-of-the-art feature-based method. Further research is necessary to determine its potential use in clinical settings as a tool to assist cardiologists and technicians with objective assessments and reduce dependence on operator subjectivity. Our research codes are available on GitHub at https://github.com/vinuni-vishc/mi-detection-echo.

Face recognition is a classical problem in Computer Vision that has experienced significant progress. Yet, in digital videos, face recognition is complicated by occlusion, pose and lighting variations, and persons entering/leaving the scene. The thesis's goal is to develop a fast method for face recognition in digital videos that is applicable to large datasets. The thesis introduces several methods to address the problems associated with video face recognition. First, to address issues associated with pose and lighting variations, a collection of face prototypes is associated with each student. Second, to speed up the process, sampling, K-means Clustering, and a combination of both are used to reduce the number of face prototypes per student. Third, the videos are processed at different frame rates. Fourth, the thesis proposes the use of active sets to address occlusion and to eliminate face recognition application on video frames with slow face motions. Fifth, the thesis develops a group face detector that recognizes students within a collaborative learning group, while rejecting out-of-group face detections. Sixth, the thesis introduces a face DeID for protecting the students' identities. Seventh, the thesis uses data augmentation to increase the training set's size. The different methods are combined using multi-objective optimization to guarantee that the full method remains fast without sacrificing accuracy. To test the approach, the thesis develops the AOLME dataset of 138 student faces (81 boys and 57 girls) of ages 10 to 14, who are predominantly Latina/o students. Compared to the baseline method, the final optimized method resulted in fast recognition times with significant improvements in face recognition accuracy. Using face prototype sampling only, the proposed method achieved an accuracy of 71.8% compared to 62.3% for the baseline system, while running 11.6 times faster.

Face recognition in collaborative learning videos presents many challenges. In collaborative learning videos, students sit around a typical table at different positions to the recording camera, come and go, move around, get partially or fully occluded. Furthermore, the videos tend to be very long, requiring the development of fast and accurate methods. We develop a dynamic system of recognizing participants in collaborative learning systems. We address occlusion and recognition failures by using past information about the face detection history. We address the need for detecting faces from different poses and the need for speed by associating each participant with a collection of prototype faces computed through sampling or K-means clustering. Our results show that the proposed system is proven to be very fast and accurate. We also compare our system against a baseline system that uses InsightFace [2] and the original training video segments. We achieved an average accuracy of 86.2% compared to 70.8% for the baseline system. On average, our recognition rate was 28.1 times faster than the baseline system.

Analysis of human mobility from GPS trajectories becomes crucial in many aspects such as policy planning for urban citizens, location-based service recommendation/prediction, and especially mitigating the spread of biological and mobile viruses. In this paper, we propose a method to find temporal factors affecting the human mobility lifestyle. We collected GPS data from 100 smartphone users in Japan. We designed a model that consists of 13 temporal patterns. We then applied a multiple linear regression and found that people tend to keep their mobility habits on Thursday and the days in the second week of a month but tend to lose their habits on Friday. We also explained some reasons behind these findings.

Most of the current user authentication systems are based on PIN code, password, or biometrics traits which can have some limitations in usage and security. Lifestyle authentication has become a new research approach. A promising idea for it is to use the location history since it is relatively unique. Even when people are living in the same area or have occasional travel, it does not vary from day to day. For Global Positioning System (GPS) data, the previous work used the longitude, the latitude, and the timestamp as the features for the classification. In this paper, we investigate a new approach utilizing the distance coherence which can be extracted from the GPS itself without the need to require other information. We applied three ensemble classification RandomForest, ExtraTrees, and Bagging algorithms; and the experimental result showed that the approach can achieve 99.42%, 99.12%, and 99.25% of accuracy, respectively.

A visual homograph attack is a way that the attacker deceives the web users about which domain they are visiting by exploiting forged domains that look similar to the genuine domains. T. Thao et al. (IFIP SEC'19) proposed a homograph classification by applying conventional supervised learning algorithms on the features extracted from a single-character-based Structural Similarity Index (SSIM). This paper aims to improve the classification accuracy by combining their SSIM features with 199 features extracted from a N-gram model and applying advanced ensemble learning algorithms. The experimental result showed that our proposed method could enhance even 1.81% of accuracy and reduce 2.15% of false-positive rate. Furthermore, existing work applied machine learning on some features without being able to explain why applying it can improve the accuracy. Even though the accuracy could be improved, understanding the ground-truth is also crucial. Therefore, in this paper, we conducted an error empirical analysis and could obtain several findings behind our proposed approach.

Attackers create adversarial text to deceive both human perception and the current AI systems to perform malicious purposes such as spam product reviews and fake political posts. We investigate the difference between the adversarial and the original text to prevent the risk. We prove that the text written by a human is more coherent and fluent. Moreover, the human can express the idea through the flexible text with modern words while a machine focuses on optimizing the generated text by the simple and common words. We also suggest a method to identify the adversarial text by extracting the features related to our findings. The proposed method achieves high performance with 82.0% of accuracy and 18.4% of equal error rate, which is better than the existing methods whose the best accuracy is 77.0% corresponding to the error rate 22.8%., Comment: PACLIC 2019, 9 pages

Machine-translated text plays a crucial role in the communication of people using different languages. However, adversaries can use such text for malicious purposes such as plagiarism and fake review. The existing methods detected a machine-translated text only using the text's intrinsic content, but they are unsuitable for classifying the machine-translated and human-written texts with the same meanings. We have proposed a method to extract features used to distinguish machine/human text based on the similarity between the intrinsic text and its back-translation. The evaluation of detecting translated sentences with French shows that our method achieves 75.0% of both accuracy and F-score. It outperforms the existing methods whose the best accuracy is 62.8% and the F-score is 62.7%. The proposed method even detects more efficiently the back-translated text with 83.4% of accuracy, which is higher than 66.7% of the best previous accuracy. We also achieve similar results not only with F-score but also with similar experiments related to Japanese. Moreover, we prove that our detector can recognize both machine-translated and machine-back-translated texts without the language information which is used to generate these machine texts. It demonstrates the persistence of our method in various applications in both low- and rich-resource languages., Comment: INLG 2019, 9 pages

Machine-translated text plays an important role in modern life by smoothing communication from various communities using different languages. However, unnatural translation may lead to misunderstanding, a detector is thus needed to avoid the unfortunate mistakes. While a previous method measured the naturalness of continuous words using a N-gram language model, another method matched noncontinuous words across sentences but this method ignores such words in an individual sentence. We have developed a method matching similar words throughout the paragraph and estimating the paragraph-level coherence, that can identify machine-translated text. Experiment evaluates on 2000 English human-generated and 2000 English machine-translated paragraphs from German showing that the coherence-based method achieves high performance (accuracy = 87.0%; equal error rate = 13.0%). It is efficiently better than previous methods (best accuracy = 72.4%; equal error rate = 29.7%). Similar experiments on Dutch and Japanese obtain 89.2% and 97.9% accuracy, respectively. The results demonstrate the persistence of the proposed method in various languages with different resource levels., Comment: 12 pages, CICLING 2019 (LNCS)

Homograph attack is a way that attackers deceive victims about which website domain name they are communicating with by exploiting the fact that many characters look alike. The attack becomes serious and is raising broad attention when recently many brand domains have been attacked such as Apple Inc., Adobe Inc., Lloyds Bank, etc. We first design a survey of human demographics, brand familiarity, and security backgrounds and apply it to 2,067 participants. We build a regression model to study which factors affect participants' ability in recognizing homograph domains. We find that for different levels of visual similarity, the participants exhibit different abilities. 13.95% of participants can recognize non-homographs while 16.60% of participants can recognize homographs whose the visual similarity with the target brand domains is under 99.9%; but when the similarity increases to 99.9%, the number of participants who can recognize homographs significantly drops down to only 0.19%; and for the homographs with 100% of visual similarity, there is no way for the participants to recognize. We also find that female participants tend to recognize homographs better the male but male participants tend to able to recognize non-homographs better than females. Security knowledge is a significant factor affecting both homographs and non-homographs; surprisingly, people who have strong security knowledge tend to be able to recognize homographs but not non-homographs. Furthermore, people who work or are educated in computer science or computer engineering do not appear as a factor affecting the ability in recognizing homographs; however, interestingly, right after they are explained about the homograph attack, people who work or are educated in computer science or computer engineering are the ones who can capture the situation the most quickly.

The modeling of large deformation fracture mechanics has been a challenging problem regarding the accuracy of numerical methods and their ability to deal with considerable changes in deformations of meshes where having the presence of cracks. This paper further investigates the extended finite element method (XFEM) for the simulation of large strain fracture for hyper-elastic materials, in particular rubber ones. A crucial idea is to use a polygonal mesh to represent space of the present numerical technique in advance, and then a local refinement of structured meshes at the vicinity of the discontinuities is additionally established. Due to differences in the size and type of elements at the boundaries of those two regions, hanging nodes produced in the modified mesh are considered as normal nodes in an arbitrarily polygonal element. Conforming these special elements becomes straightforward by the flexible use of basis functions over polygonal elements. Results of this study are shown through several numerical examples to prove its efficiency and accuracy through comparison with former achievements.

We deal with a class of semilinear nonlocal differential equations in Hilbert spaces which is a general model for some anomalous diffusion equations. By using the theory of integral equations with completely positive kernel together with local estimates, some existence, regularity and stability results are established. An application to nonlocal partial differential equations is shown to demonstrate our abstract results.