Abstract: Methods and systems are provided for generating, for respective mutually exclusive classes of model inputs, separate output thresholds that can be applied to the continuous-valued output of a neural network or other machine learning model in order to classify inputs in a class-sensitive manner. Such classes could be related to operational or other constraints with respect to the classifier outputs that vary across the classes of inputs. Thus, the machine learning model can be improved by using training data from all of the available classes while allowing the end performance of the model plus threshold classifier to be separately set for each input class. These automated methods for class-specific threshold setting also provide improvements with respect to accuracy, time, and cost. Also provided are methods and systems for per-class calibration of model outputs.

Abstract: A method includes accessing a subject entity and a subject relation of a focal platform and accessing a knowledge graph representative of control performance data. Further, the method includes computing a set of ranked target entities that cause the subject entity based on the subject relation or are an effect of the subject entity based on the subject relation. Computing the set of ranked target entities is performed using relational hops from the subject entity within the knowledge graph performed using the subject relation and reward functions. The method also includes transmitting the set of ranked target entities to the focal platform. The set of ranked target entities is usable for modifying a user interface of an interactive computing environment provided by the focal platform.

Abstract: This disclosure covers methods, non-transitory computer readable media, and systems that use an intelligent analytics interface to process natural-language and other inputs to configure an analytics task for the system. The disclosed methods, non-transitory computer readable media, and systems provide the intelligent analytics interface to facilitate an exchange between the systems and a user to determine values for the analytics task. The methods, non-transitory computer readable media, and systems then use these values to execute an analytics task.

Abstract: A content saliency network is a machine-learned neural network that predicts the saliency of elements of a content item. The content saliency network may be used in a method that includes determining a set of elements in draft content and computing a first pixel-level vector for the content. The method may also include, for each element in the set of elements, computing a vector of simple features for the element, the simple features being computed from attributes of the element, computing a second pixel-level vector for the element, computing a third pixel-level vector for an intermediate context of the element, and providing the vectors to the content saliency network. The content saliency network provides a saliency score for the element. The method further includes generating an element-level saliency map of the content using the respective saliency scores for the set of elements and providing the saliency map to a requestor.

Abstract: Techniques of forecasting web metrics involve generating, prior to the end of a period of time, a probability of a metric taking on an anomalous value, e.g., a value indicative of an anomaly with respect to web traffic, at the end of the period based on previous values of the metric. Such a probability is based on a distribution of predicted values of the metric at some previous period of time. For example, a web server may use actual values of the number of bounces collected at hourly intervals in the middle of a day to predict a number of bounces at the end of the current day. Further, the web server may also compute a confidence interval to determine whether a predicted end-of-day number of bounces may be considered anomalous. The width of the confidence interval indicates the probability that a predicted end-of-day number of bounces has an anomalous value.

Abstract: A machine-learning framework uses partial-click feedback to generate an optimal diverse set of items. An example method includes estimating a preference vector for a user based on diverse cascade statistics for the user, the diverse cascade statistics including previously observed responses and previously observed topic gains. The method also includes generating an ordered set of items from the item repository, the items in the ordered set having highest topic gain weighted by similarity with the preference vector, providing the ordered set for presentation to the user, and receiving feedback from the user on the ordered set. The method also includes, responsive to the feedback indicating a selected item, updating the diverse cascade statistics for observed items, wherein the updating results in penalizing the topic gain for items of the observed items that are not the selected item and promoting the topic gain for the selected item.

Abstract: Techniques are provided for detecting new topics and themes and assigning new posts to existing topic and/or theme clusters in online community discussions. A post posted to an online community is received and a post feature vector representative of the post is created. The post is compared to a plurality of centroid feature vectors, each centroid feature vector being representative of a respective post cluster and associated with a theme. Upon determining that similarity between the post feature vector and one of a plurality of centroid feature vectors satisfies a minimum similarity threshold, the post is assigned to the post cluster of which the centroid feature vector is representative. Upon determining that similarity between the post feature vector and any of the plurality of centroid feature vectors is below the minimum similarity threshold, a new theme cluster is created and the post is assigned to the new theme cluster.

Abstract: A method includes accessing a subject entity and a subject relation of a focal platform and accessing a knowledge graph representative of control performance data. Further, the method includes computing a set of ranked target entities that cause the subject entity based on the subject relation or are an effect of the subject entity based on the subject relation. Computing the set of ranked target entities is performed using relational hops from the subject entity within the knowledge graph performed using the subject relation and reward functions. The method also includes transmitting the set of ranked target entities to the focal platform. The set of ranked target entities is usable for modifying a user interface of an interactive computing environment provided by the focal platform.

Abstract: A content saliency network is a machine-learned neural network that predicts the saliency of elements of a content item. The content saliency network may be used in a method that includes determining a set of elements in a UI and computing a first context vector for the content. The method may also include, for each element in the set of elements, computing a vector of simple features for the element, the simple features being computed from attributes of the element, computing a second context vector for the element, computing a third context vector for an intermediate context of the element, and providing the vectors to the content saliency network. The content saliency network provides a saliency score for the element. The method further includes generating an element-level saliency map of the content using the respective saliency scores for the set of elements and providing the saliency map to a requestor.

Abstract: This disclosure covers methods, non-transitory computer readable media, and systems that use an intelligent analytics interface to process natural-language and other inputs to configure an analytics task for the system. The disclosed methods, non-transitory computer readable media, and systems provide the intelligent analytics interface to facilitate an exchange between the systems and a user to determine values for the analytics task. The methods, non-transitory computer readable media, and systems then use these values to execute an analytics task.

Abstract: This disclosure covers methods, non-transitory computer readable media, and systems that use an intelligent analytics interface to process natural-language and other inputs to configure an analytics task for the system. The disclosed methods, non-transitory computer readable media, and systems provide the intelligent analytics interface to facilitate an exchange between the systems and a user to determine values for the analytics task. The methods, non-transitory computer readable media, and systems then use these values to execute an analytics task.

Abstract: A computer system stores digital media content such as images and video along with associated tags and timestamps. The system detects trends in the media content by semantic analysis which includes generation of a temporal tag graph that includes data indicative of a semantic representation of the tags over a plurality of time periods. The data in the tag graph is clustered to generate a set of identified trends reflected by the tags over the plurality of time periods. The set of identified trends is stored in data storage and is available for characterization which includes labeling of the trends, scoring the trends, evaluating changes in the trends over time, and identifying images representative of the detected trends. The temporal tag graph may take the form of a weighted undirected graph where each node in the graph is associated with one of the tags and the edges connecting the nodes represents a temporal correlation between the nodes associated with each edge.

Abstract: A content saliency network is a machine-learned neural network that predicts the saliency of elements of a content item. The content saliency network may be used in a method that includes determining a set of elements in a UI and computing a first context vector for the content. The method may also include, for each element in the set of elements, computing a vector of simple features for the element, the simple features being computed from attributes of the element, computing a second context vector for the element, computing a third context vector for an intermediate context of the element, and providing the vectors to the content saliency network. The content saliency network provides a saliency score for the element. The method further includes generating an element-level saliency map of the content using the respective saliency scores for the set of elements and providing the saliency map to a requestor.

Abstract: Natural language notification generation techniques and system are described. In an implementation, natural language notifications are generated to provide insight into alerts related to a metric, underlying causes of the alert from other metrics, and relationships of the metric to other metrics. In this way, a user may gain this insight in an efficient, intuitive, and time effective manner.

Abstract: A machine-learning framework uses partial-click feedback to generate an optimal diverse set of items. An example method includes estimating a preference vector for a user based on diverse cascade statistics for the user, the diverse cascade statistics including previously observed responses and previously observed topic gains. The method also includes generating an ordered set of items from the item repository, the items in the ordered set having highest topic gain weighted by similarity with the preference vector, providing the ordered set for presentation to the user, and receiving feedback from the user on the ordered set. The method also includes, responsive to the feedback indicating a selected item, updating the diverse cascade statistics for observed items, wherein the updating results in penalizing the topic gain for items of the observed items that are not the selected item and promoting the topic gain for the selected item.

Abstract: A content saliency network is a machine-learned neural network that predicts the saliency of elements of a content item. The content saliency network may be used in a method that includes determining a set of elements in draft content and computing a first pixel-level vector for the content. The method may also include, for each element in the set of elements, computing a vector of simple features for the element, the simple features being computed from attributes of the element, computing a second pixel-level vector for the element, computing a third pixel-level vector for an intermediate context of the element, and providing the vectors to the content saliency network. The content saliency network provides a saliency score for the element. The method further includes generating an element-level saliency map of the content using the respective saliency scores for the set of elements and providing the saliency map to a requestor.

Abstract: This disclosure covers methods, non-transitory computer readable media, and systems that use an intelligent analytics interface to process natural-language and other inputs to configure an analytics task for the system. The disclosed methods, non-transitory computer readable media, and systems provide the intelligent analytics interface to facilitate an exchange between the systems and a user to determine values for the analytics task. The methods, non-transitory computer readable media, and systems then use these values to execute an analytics task.

Abstract: Techniques of forecasting web metrics involve generating, prior to the end of a period of time, a probability of a metric taking on an anomalous value, e.g., a value indicative of an anomaly with respect to web traffic, at the end of the period based on previous values of the metric. Such a probability is based on a distribution of predicted values of the metric at some previous period of time. For example, a web server may use actual values of the number of bounces collected at hourly intervals in the middle of a day to predict a number of bounces at the end of the current day. Further, the web server may also compute a confidence interval to determine whether a predicted end-of-day number of bounces may be considered anomalous. The width of the confidence interval indicates the probability that a predicted end-of-day number of bounces has an anomalous value.

Abstract: A computer system stores digital media content such as images and video along with associated tags and timestamps. The system detects trends in the media content by semantic analysis which includes generation of a temporal tag graph that includes data indicative of a semantic representation of the tags over a plurality of time periods. The data in the tag graph is clustered to generate a set of identified trends reflected by the tags over the plurality of time periods. The set of identified trends is stored in data storage and is available for characterization which includes labeling of the trends, scoring the trends, evaluating changes in the trends over time, and identifying images representative of the detected trends. The temporal tag graph may take the form of a weighted undirected graph where each node in the graph is associated with one of the tags and the edges connecting the nodes represents a temporal correlation between the nodes associated with each edge.

Abstract: Natural language notification generation techniques and system are described. In an implementation, natural language notifications are generated to provide insight into alerts related to a metric, underlying causes of the alert from other metrics, and relationships of the metric to other metrics. In this way, a user may gain this insight in an efficient, intuitive, and time effective manner.