Abstract: Techniques for real-time generation and customization of text classification models. An initial dataset of input text samples are manually assigned labels, and the labeled input text samples are tokenized and provided as training data to train machine learning classifiers for various classes or categories of the input text samples. As the machine learning classifiers train with the training data, feedback in the form of suggestions (or predictions) are provided in real time by the text classification models regarding which label(s) to assign to any input text sample(s) currently in the training data or any new input text sample(s) further provided as training data for the respective machine learning classifiers. The suggested (or predicted) label(s) can be manually assigned to the input text sample(s), if deemed appropriate, and the newly labeled input text sample(s) can be provided to supplement the existing training data for the respective machine learning classifiers.

Abstract: Technology for generating a consistently labeled training dataset. For each one of multiple previously labeled texts, a distance between the previously labeled text and a current text to be labeled is generated by comparing a list of tokens for the previously labeled text to a list of tokens for the current text to determine an overlap value equal to a number of tokens that match between the list of tokens for the previously labeled text and the list of tokens for the current text, and using the overlap value to calculate a distance between the previously labeled text and the current text that is inversely correlated to the overlap value. Previously labeled texts that are most similar to the current text are identified as those previously labeled texts having the shortest distances to the current text, and are displayed with their previously assigned labels in a label selection user interface.

Abstract: A computer-implemented technique provides electronic text recommendations to a user. The technique involves capturing content from a conference among multiple participants. The technique further involves receiving electronic typing input from the user while the user is typing to document the conference. The technique further involves outputting, in response to the electronic typing input from the user, electronic text recommendations to the user based on the captured content. Such recommendations can take the form of predictive typing for taking notes based on content shared during an online meeting. The words and/or phrases in the predictive typing may be derived from the words spoken and/or from visual data displayed during the conference. Accordingly, a user is able to document what is discussed with minimal typing and thus reduced distraction. Furthermore, the information that is documented can be made more on-point and less verbose than a lengthy transcription of an audio feed.

Abstract: Technology for generating a consistently labeled training dataset. For each one of multiple previously labeled texts, a distance between the previously labeled text and a current text to be labeled is generated by comparing a list of tokens for the previously labeled text to a list of tokens for the current text to determine an overlap value equal to a number of tokens that match between the list of tokens for the previously labeled text and the list of tokens for the current text, and using the overlap value to calculate a distance between the previously labeled text and the current text that is inversely correlated to the overlap value. Previously labeled texts that are most similar to the current text are identified as those previously labeled texts having the shortest distances to the current text, and are displayed with their previously assigned labels in a label selection user interface.

Abstract: Techniques for real-time generation and customization of text classification models. An initial dataset of input text samples are manually assigned labels, and the labeled input text samples are tokenized and provided as training data to train machine learning classifiers for various classes or categories of the input text samples. As the machine learning classifiers train with the training data, feedback in the form of suggestions (or predictions) are provided in real time by the text classification models regarding which label(s) to assign to any input text sample(s) currently in the training data or any new input text sample(s) further provided as training data for the respective machine learning classifiers. The suggested (or predicted) label(s) can be manually assigned to the input text sample(s), if deemed appropriate, and the newly labeled input text sample(s) can be provided to supplement the existing training data for the respective machine learning classifiers.

Abstract: A technique provides a sync notification to indicate when a shared screen view of a presenter device has been delivered to an audience. The technique involves generating a sync time estimation based on communications with a set of audience devices, and detecting occurrence of an update to the shared screen view of the presenter device. The technique further involves outputting, after the sync time estimation has elapsed since the detected occurrence of the update, the sync notification on the presenter device to indicate when the shared screen view of the presenter device has been delivered to the audience. Such a technique is well suited for a variety of online collaboration systems such as online conference systems, online webinar systems, online meeting systems, and so on.

Abstract: A computer-implemented technique provides electronic text recommendations to a user. The technique involves capturing content from a conference among multiple participants. The technique further involves receiving electronic typing input from the user while the user is typing to document the conference. The technique further involves outputting, in response to the electronic typing input from the user, electronic text recommendations to the user based on the captured content. Such recommendations can take the form of predictive typing for taking notes based on content shared during an online meeting. The words and/or phrases in the predictive typing may be derived from the words spoken and/or from visual data displayed during the conference. Accordingly, a user is able to document what is discussed with minimal typing and thus reduced distraction. Furthermore, the information that is documented can be made more on-point and less verbose than a lengthy transcription of an audio feed.

Abstract: A technique provides a sync notification to indicate when a shared screen view of a presenter device has been delivered to an audience. The technique involves generating a sync time estimation based on communications with a set of audience devices, and detecting occurrence of an update to the shared screen view of the presenter device. The technique further involves outputting, after the sync time estimation has elapsed since the detected occurrence of the update, the sync notification on the presenter device to indicate when the shared screen view of the presenter device has been delivered to the audience. Such a technique is well suited for a variety of online collaboration systems such as online conference systems, online webinar systems, online meeting systems, and so on.

Abstract: Techniques of finding a meeting time involve recommending a meeting time based on attendee meeting history. Along these lines, a meeting scheduler accesses meeting logs that contain trends involving previous meetings to which attendees were invited. For example, such meeting logs may indicate whether an attendee accepted meeting invitations to meetings occurring at particular times during a workweek. Based on the accessed meeting logs, the meeting scheduler recommends a time at which to hold the meeting. For example, the meeting logs may indicate that an attendee declines invitations at certain times of the day, or that another attendee has flexibility during a time labeled as unavailable in the calendar. The meeting schedule then sends a message to the meeting organizer that contains the recommended time.

Abstract: A technique smartly displays video feeds during an online meeting. The technique involves receiving, by processing circuitry, a plurality of video feeds during the online meeting. The technique further involves selecting, by the processing circuitry, a group of video feeds from the plurality of video feeds during the online meeting. The number of video feeds in the group of video feeds is less than the number of video feeds in the plurality of video feeds. Additionally, the number of video feeds in the group is at least two. The technique further involves outputting, by the processing circuitry, the group of video feeds to an online meeting participant during the online meeting. Such a technique enables the number of participants of the online meeting to be essentially unlimited, where only the video feeds deemed most significant (e.g., based on scoring) are shared during the online meeting.

Abstract: A technique smartly displays video feeds during an online meeting. The technique involves receiving, by processing circuitry, a plurality of video feeds during the online meeting. The technique further involves selecting, by the processing circuitry, a group of video feeds from the plurality of video feeds during the online meeting. The number of video feeds in the group of video feeds is less than the number of video feeds in the plurality of video feeds. Additionally, the number of video feeds in the group is at least two. The technique further involves outputting, by the processing circuitry, the group of video feeds to an online meeting participant during the online meeting. Such a technique enables the number of participants of the online meeting to be essentially unlimited, where only the video feeds deemed most significant (e.g., based on scoring) are shared during the online meeting.

Abstract: A technique for recording dictation, meetings, lectures, and other events includes automatically segmenting an audio recording into portions by detecting speech transitions within the recording and selectively identifying certain portions of the recording as noteworthy. Noteworthy audio portions are displayed to a user for selective playback. The user can navigate to different noteworthy audio portions while ignoring other portions. Each noteworthy audio portion starts and ends with a speech transition. Thus, the improved technique typically captures noteworthy topics from beginning to end, thereby reducing or avoiding the need for users to have to search for the beginnings and ends of relevant topics manually.