Abstract: A system for proctoring online exams includes a client-side computing system and a visual analytics system. The client-side computing system includes a camera configured to obtain video data corresponding to a user while taking an exam and one or more input devices configured to obtain interaction data, wherein the interaction data includes mouse movements of the user while taking the exam. The visual analytics system is configured to obtain the video data and the interaction data from the client-side computing system, analyze the exam data to detect abnormal behavior by the user based at least in part on mouse movement data, and generate one or more visualizations of the analyzed exam data to be used in determining whether or not the user has cheated during the exam.

Abstract: The invention relates to a computer implemented method and system for analyzing a video.

Abstract: A system for proctoring online exams includes a client-side computing system and a visual analytics system. The client-side computing system includes a camera configured to obtain video data corresponding to a user while taking an exam and one or more input devices configured to obtain interaction data, wherein the interaction data includes mouse movements of the user while taking the exam. The visual analytics system is configured to obtain the video data and the interaction data from the client-side computing system, analyze the exam data to detect abnormal behavior by the user based at least in part on mouse movement data, and generate one or more visualizations of the analyzed exam data to be used in determining whether or not the user has cheated during the exam.

Abstract: A deep sequence model with prototypes may be steered. A prototype overview is displayed, the prototype overview including a plurality of prototype sequences learned by a model through backpropagation, each of the prototype sequences including a series of events, where for each of the prototype sequences, statistical information is presented with respect to use of the prototype sequence by the model. Input is received adjusting one or more of the prototype sequences to fine-tune the model. The model is updated using the plurality of prototype sequences, as adjusted, to create an updated model. The model, as updated, is displayed in the prototype overview.

Abstract: A computer implemented method and system for processing an audio signal. The method includes the steps of extracting prosodic features from the audio signal, aligning the extracted prosodic features with a script derived from or associated with the audio signal, and segmenting the script with the aligned extracted prosodic features into structural blocks of a first type. The method may further include determining a distance measure between a structural block of a first type derived from the script with another structural block of the first type using, for example, the Damerau-Levenshtein distance.

Abstract: A computer implemented method and system processing a video signal.

Abstract: A method for predicting student performance in interactive question pools, comprising: a data processing and feature extraction module for: extracting from historical score data records statistical student features reflecting the students' past performances on the questions; and statistical question features indicating a question popularity of each question and the average students' scores on each question; and extracting from the pointing device movement data records a plurality of interactive edge features representing the characteristics of the students' problem-solving behaviors; a network construction module for: performing an Edge2Node transformation to transform the interactive edge features into one or more interaction nodes; and constructing a heterogeneous student-interaction-question (SIQ) network; and a prediction module for processing the SIQ network using a residual relational graph neural network to predict a student's score on unattempt questions.

Abstract: A deep sequence model with prototypes may be steered. A prototype overview is displayed, the prototype overview including a plurality of prototype sequences learned by a model through backpropagation, each of the prototype sequences including a series of events, where for each of the prototype sequences, statistical information is presented with respect to use of the prototype sequence by the model. Input is received adjusting one or more of the prototype sequences to fine-tune the model. The model is updated using the plurality of prototype sequences, as adjusted, to create an updated model. The model, as updated, is displayed in the prototype overview.

Abstract: A computer implemented method and system processing a video signal.

Abstract: A computer implemented method and system for processing an audio signal. The method includes the steps of extracting prosodic features from the audio signal, aligning the extracted prosodic features with a script derived from or associated with the audio signal, and segmenting the script with the aligned extracted prosodic features into structural blocks of a first type. The method may further include determining a distance measure between a structural block of a first type derived from the script with another structural block of the first type using, for example, the Damerau-Levenshtein distance.

Abstract: The present teaching relates to analyzing user activities related to a video. The video is provided to a plurality of users. The plurality of users is monitored to detect one or more types of user activities performed in time with respect to different portions of the video. One or more visual representations of the monitored one or more types of user activities are generated. The one or more visual representations capture a level of attention paid by the plurality of users to the different portions of the video at any time instance. Interests of at least some of the plurality of users are determined with respect to the different portions of the video based on the one or more visual representations.

Abstract: The present disclosure provides a data visualization method and apparatus. The method includes: displaying a distribution curve diagram by using a display device, where the distribution curve diagram is used to represent time-based distribution of multiple pieces of spatial-temporal data, and the distribution curve diagram includes a timeline, a curve changing with the timeline, and an irregular region enclosed by the curve; and displaying a tree map in the irregular region by using the display device, where the tree map divides the irregular region into m sub-regions that are in a one-to-one correspondence with m regions, and an area of each of the m sub-regions is used to indicate an amount of spatial-temporal data distributed in a region corresponding to each sub-region. The present disclosure implements visualization of spatial-temporal data by using the tree map.

Abstract: The present disclosure provides a data visualization method and apparatus. The method includes: displaying a distribution curve diagram by using a display device, where the distribution curve diagram is used to represent time-based distribution of multiple pieces of spatial-temporal data, and the distribution curve diagram includes a timeline, a curve changing with the timeline, and an irregular region enclosed by the curve; and displaying a tree map in the irregular region by using the display device, where the tree map divides the irregular region into m sub-regions that are in a one-to-one correspondence with m regions, and an area of each of the m sub-regions is used to indicate an amount of spatial-temporal data distributed in a region corresponding to each sub-region. The present disclosure implements visualization of spatial-temporal data by using the tree map.

Abstract: A data visualization method includes acquiring a spatiotemporal behavior data set generated by network users, determining, according to a time point at which or a time period in which each piece of spatiotemporal behavior data is generated, an amount of spatiotemporal behavior data that is in the spatiotemporal behavior data set and generated at a time point or in a time period falling within each time period in N time periods, and presenting a column bar corresponding to an ith time period in the N time periods, where the column bar is in a radial direction of a first circle representing a full time domain and intersects the first circle and the length of a column bar represents a quantity of behaviors in a time period corresponding to the column bar.

Abstract: The present teaching relates to analyzing user activities related to a video. The video is provided to a plurality of users. The plurality of users is monitored to detect one or more types of user activities performed in time with respect to different portions of the video. One or more visual representations of the monitored one or more types of user activities are generated. The one or more visual representations capture a level of attention paid by the plurality of users to the different portions of the video at any time instance. Interests of at least some of the plurality of users are determined with respect to the different portions of the video based on the one or more visual representations.

Abstract: A data visualization method includes acquiring a spatiotemporal behavior data set generated by network users, determining, according to a time point at which or a time period in which each piece of spatiotemporal behavior data is generated, an amount of spatiotemporal behavior data that is in the spatiotemporal behavior data set and generated at a time point or in a time period falling within each time period in N time periods, and presenting a column bar corresponding to an ith time period in the N time periods, where the column bar is in a radial direction of a first circle representing a full time domain and intersects the first circle and the length of a column bar represents a quantity of behaviors in a time period corresponding to the column bar.