Abstract: An information management system is provided herein that uses machine learning (ML) to predict what data to store in a secondary storage device and/or when to perform the storage. For example, a client computing device can be initially configured to store data in a secondary storage device according to one or more storage policies. A media agent in the information management system can monitor data usage on the client computing device, using the data usage data to train a data storage ML model. The data storage ML model may be trained such that the model predicts what data to store in a secondary storage device and/or when to perform the storage. The client computing device can then be configured to use the trained data storage ML model in place of the storage polic(ies) to determine which data to store in a secondary storage device and/or when to perform the storage.

Abstract: A computing system trains a clustering model. A responsibility parameter vector is initialized for each observation vector that includes a probability value of a cluster membership in each cluster. (A) Beta distribution parameter values are computed for each cluster. (B) Parameter values are computed for a normal-Wishart distribution for each cluster. (C) Each responsibility parameter vector defined for each observation vector is updated using the computed beta distribution parameter values, the computed parameter values for the normal-Wishart distribution, and a respective observation vector of the plurality of observation vectors. (D) A convergence parameter value is computed. (E) (A) to (D) are repeated until the computed convergence parameter value indicates the responsibility parameter vector defined for each observation vector is converged. A cluster membership is determined for each observation vector using a respective, updated responsibility parameter vector.

Abstract: A method for expanding an answer key to verify a question and answer system is provided in the illustrative embodiments. A definition is constructed of an extended answer type. The extended answer type represents an answer type of an unrepresented answer. The unrepresented answer is unrepresented in the answer key as a valid response to a question in a set of valid responses to the question in the answer key. The extended answer type is created in the answer key according to the definition. The extended answer type is populated such that the unrepresented answer becomes as additional valid response to the question, the creating and the populating extending the answer key to form an extended answer key. The populated extended answer type in the extended answer key is used to verify that a generated answer from the Q and A system is correct.

Abstract: Systems and methods are provided herein for use in identifying and/or detecting electronic message containing malicious content. One exemplary method includes receiving multiple electronic tags. Each of the multiple electronic tags corresponds to an electronic message and a use, and the user caused the electronic tag to be associated with the electronic message based on a perception that the electronic message included malicious content. The exemplary method further includes assigning, for each electronic tag, point(s) to the corresponding user when the corresponding electronic message includes malicious content, totaling, for each user, the point(s) assigned during a predefined interval, and identifying one of the user(s) with a highest total point(s), for the defined interval, as a winner, thereby incentivizing users to associate electronic tags with electronic message perceived to include malicious content.

Abstract: Systems, methods, and computer-readable media for performing network assurance in a traditional network. In some examples, a system can collect respective sets of configurations programmed at network devices in a network and, based on the respective sets of configurations, determine a network-wide configuration of the network, the network-wide configuration including virtual local area networks (VLANs), access control lists (ACLs) associated with the VLANs, subnets, and/or a topology. Based on the network-wide configuration of the network, the system can compare the ACLs for each of the VLANs to yield a VLAN consistency check, compare respective configurations of the subnets to yield a subnet consistency check, and perform a topology consistency check based on the topology. Based on the VLAN consistency check, the subnet consistency check, and the topology consistency check, the system can determine whether the respective sets of configurations programmed at the network devices contain a configuration error.

Abstract: When an electronic mail is transmitted, a computer outputs to a display device a transmission confirmation screen on which a destination or an attached file name is displayed, the destination or the attached file name being different from a destination or an attached file name that has been set by a user in the electronic mail, and on which whether transmission of the electronic mail will be performed is selectable. Upon receipt of selection of whether the transmission will be performed, the computer outputs to the display device message information that varies according to a result of the selection of whether the transmission will be performed on the transmission confirmation screen.

Abstract: A method for dialog state tracking uses a neural network model, such as an MemN2N model, which has been trained to receive a representation of a question and a representation of a subpart of a dialog and to output an answer to the question. For at least one iteration, a subpart of a dialog is received. A representation of the subpart of the dialog is generated. The representation of the subpart of the input dialog and representation of a question are input to the trained neural network model. An answer is output by the neural network model, based on the representation of the question and the representation of the subpart of the input dialog. A dialog state for the dialog is updated, based on the answer to the question. The dialog state includes a set of variables. The updating includes predicting a value for at least one of the variables.

Abstract: Methods and apparatus are described herein for identifying tasks in messages. In various implementations, natural language processing may be performed on a received message to generate an annotated message. The annotated message may be analyzed pursuant to a grammar. A portion of the message may be classified as a user task entry based on the analysis of the annotated message.

Abstract: Systems and methods for responding to a user query are provided. More specifically, the systems and methods provide static search results along with a relevant chat bot user interface for combined display on a browser interface in response to a user query. As such, the systems and methods utilize a conversation layer to communicate with a selected chat bot service. Additionally, the systems and methods may update the static search results and other related content displayed in the browser interface in response to user chat bot inputs and/or chat bot responses provided on the chat bot user interface in the combined display. Accordingly, the systems and methods for responding to a user query provide a more efficient and better search engine when compared to previously utilized search engines that did not provide access to a chat bot interface on the same browser interface as the search results.

Abstract: Techniques are described herein that are capable of suggesting intent frame(s) for user request(s). For instance, the intent frame(s) may be suggested to elicit a request from a user. An intent frame is a natural language phrase (e.g., a sentence) that includes at least one carrier phrase and at least one slot. A slot in an intent frame is a placeholder that is identified as being replaceable by one or more words that identify an entity and/or an action to indicate an intent of the user. A carrier phrase in an intent frame includes one or more words that suggest a type of entity and/or action that is to be identified by the one or more words that may replace the corresponding slot. In accordance with these techniques, the intent frame(s) are suggested in response to determining that natural language functionality of a processing system is activated.

Abstract: In an approach for providing a dynamic definition of a problem statement during an IT critical situation with an associated conflict resolution, a processor receives an initial problem definition. A processor identifies related social collaboration channels associated with the initial problem definition. A processor updates users within the related social collaboration channels. A processor determines there is at least one problem update from the related social collaboration channels. A processor identifies a set of subject matter experts who are relevant to the initial problem definition. A processor initiates a different social collaboration channel for the set of subject matter experts. A processor determines a consensus to a solution to the initial problem definition is gained by the set of subject matter experts. A processor updates the initial problem definition to include the solution.

Abstract: Systems and methods for embodiments of a graph based artificial intelligence systems for identity management are disclosed. Embodiments of the identity management systems disclosed herein may utilize a network graph approach to analyzing identities or entitlements of a distributed networked enterprise computing environment. Specifically, in certain embodiments, an artificial intelligence based identity governance systems may include an intelligent decision support agent to provide an approval or denial recommendation for an access request. To provide an approval or denial recommendation, the intelligent agent may utilize a classifier trained on historical certification data. The intelligent agent may utilize features which represent relevant signals to the approval or denial decision including features that may be associated with a network graph of the identities and entitlements of the enterprise computing environment.

Abstract: Techniques and solutions for measuring web page value are described. For example, web page content information for a web page can be obtained. The web page can be decomposed into web page content components based on the web page content information. Content types can be determined for the web page content components. The web page content components can be weighted based on the content types and/or based on other criteria. Web page resources can be obtained in advance for web pages and/or content components based on the weights.

Abstract: A system and method of detecting an aberrant message is provided. An ordered set of words within the message is detected. The set of words found within the message is linked to a corresponding set of expected words, the set of expected words having semantic attributes. A set of grammatical structures represented in the message is detected, based on the ordered set of words and the semantic attributes of the corresponding set of expected words. A cognitive noise vector comprising a quantitative measure of a deviation between grammatical structures represented in the message and an expected measure of grammatical structures for a message of the type is then determined. The cognitive noise vector may be processed by higher levels of the neural network and/or an external processor.

Abstract: A simulation application receives simulation parameters associated with a simulation to be generated. The simulation parameters include geometry associated with the simulation and corresponding boundary conditions. The simulation engine processes the simulation parameters and then, using a neural network, generates a solution estimate. Based on the estimated solution, the simulation engine then executes a finite element analysis solver using the solution estimate as a starting point. The FEA solver iterates until a converged solution is reached. The converged solution is then provided to the end-user.

Abstract: Techniques for customizing knowledge representation systems including identifying, based on a plurality of concepts in a knowledge representation (KR), a group of one or more concepts relevant to user context information, and providing the identified group of one more concepts to a user. The KR may include a combination of modules. The modules may include a kernel and a customized module customized for the user. The kernel may accessible via a second KR.

Abstract: A computer-implemented method and an apparatus facilitate user engagement on enterprise interaction channels. Information related to a current journey of a user on one or more enterprise interaction channels is received. The user is categorized as one of a hot lead, a warm lead, and a non-hot lead based, at least in part, based on the received information related to the current journey of the user. If the user is categorized as the non-hot lead, a user interface (UI) displayed to the user is modified. The UI is modified to facilitate user engagement for converting the user from a non-purchasing entity to a purchasing entity.

Abstract: The present invention is directed to intelligent and near real-time monitoring in a streaming environment. In embodiments, a streaming platform streams raw data from raw data capture devices, where the raw data is contextually enriched using multiple information sources. A rule execution operator, integrated within the streaming platform, detects events by executing rules against the contextually enriched data. The rules applied are customizable, flexible, multivariate, and dynamically responsive to changing information. When an event is detected, an action can be instantiated in response to the event detection and/or newly derived information. The action and/or instructions for performing the action are communicated to external sources, where the action can be initiated and/or completed. Accordingly, actions are instantiated in near real time relative to the initial receipt of raw data from which the events were detected. Actions may be delayed using sophisticated temporal reasoning, in some embodiments.

Abstract: Methods and software that allow one or more users to utilize targeted-advertising functionality in the context of a product program. For example, such logistical functionality may enable users to benefit from the use of one or more functionalities enabled by targeted advertising, such as fully or partially ad-sponsored software, relevant training or assistance recommendations, and user-customized advertising, among others. Various corresponding and related methods and software are described.

Abstract: A security control system for a kinematic robot uses a cognitive assessment agent to map proposed instructions to either legitimate or illegitimate actions based on contextual variables. The agent computes a security anomaly index score representing a variance of a likely kinematic action of the robot compared to acceptable actions. If the score exceeds a predetermined threshold, a security alert is generated for the robot's administrator. The contextual variables include a user profile, a user location, and subject matter of the kinematic actions. The analysis compares input text to predefined classification metadata, and can also compare verbal phrases or body gestures to corresponding baselines. Different numeric weights can be applied to the contextual variables. The computing begins with a default value for the score and thereafter increments or decrements the score based on the weights. The weights can be adjusted based on a supervisory appraisal of the computed score.

Abstract: A security control system for a kinematic robot uses a cognitive assessment agent to map proposed instructions to either legitimate or illegitimate actions based on contextual variables. The agent computes a security anomaly index score representing a variance of a likely kinematic action of the robot compared to acceptable actions. If the score exceeds a predetermined threshold, a security alert is generated for the robot's administrator. The contextual variables include a user profile, a user location, and subject matter of the kinematic actions. The analysis compares input text to predefined classification metadata, and can also compare verbal phrases or body gestures to corresponding baselines. Different numeric weights can be applied to the contextual variables. The computing begins with a default value for the score and thereafter increments or decrements the score based on the weights. The weights can be adjusted based on a supervisory appraisal of the computed score.

Abstract: In an approach for managing a set of tests to run on a set of platforms, a processor identifies a set of tests to run on a set of platforms and one or more priorities associated with the set of tests. A processor sends the set of tests and the one or more priorities associated with the set of tests to the set of platforms. A processor receives information about a first test run of a first test on a first platform of the set of platforms. A processor determines that the first test failed on the first platform. A processor updates the one or more priorities associated with the set of tests. A processor sends the updated one or more priorities associated with the set of tests to at least the second platform.

Abstract: A method and system are provided. The method includes providing an actionable business entity operating model. The method further includes extending the actionable business entity operating model to drive user interface behavior on a user interface device having at least a display device, by extending class and property meta classes of the actionable business entity operating model to include user interface behavior semantics. The extending step includes configuring the user interface behavior semantics to be responsive to an operation state of the actionable business entity operating model.

Abstract: The present invention is directed to a method and computer readable medium for automatically generating a Uniform Resource Locator (URL) from schema and semantics of an organizational structure. The semantically intelligent URL conveys knowledge about content it accompanies by defining patterns of possible URL parameters within an organizational structure of related content stored in a database, each pattern being represented by entities and semantic information formed of attributes and relationships. The patterns of the URL organize categories of information and form relationships among the entities. The patterns are used for automatically generating a URL along with semantic information about the content in a resource or web page, wherein said semantic information explains the content, such as with attributes of entities, such as in a name value pair that correlates to the content of a page.

Abstract: A method is provided for determining a website user behavioral portrait based on navigation on the website and dynamically reconfiguring web pages based on those portraits. In accordance with the method, data relating to the progress of a user through a website is recorded, and an ongoing behavioral portrait of the user is built based on the data. The portrait is then used to dynamically reconfigure web content.

Abstract: A method, system and computer program product for improving management and performance of an employee. An indication of a goal sponsored by an owner, such as a manager, is received. After receiving an indication of a user subscribing to the goal sponsored by the owner, communications (e.g., posts) on the social network involving the subscribed user that include a tag designating the goal may be monitored. Upon detecting a completion of the goal by the subscribed user in the monitored communications, a pattern of steps (e.g., “prepare,” “review” and “finish”) to accomplish the goal is determined based on the monitored communications. The pattern may then be used to recommend actions to other users subscribed to the goal concerning accomplishing the goal based on the pattern. Furthermore, the owner may receive indications as to the progress of the user in accomplishing the goal based on the pattern.

Abstract: A method, system and computer program product for improving management and performance of an employee. An indication of a goal sponsored by an owner, such as a manager, is received. After receiving an indication of a user subscribing to the goal sponsored by the owner, communications (e.g., posts) on the social network involving the subscribed user that include a tag designating the goal may be monitored. Upon detecting a completion of the goal by the subscribed user in the monitored communications, a pattern of steps (e.g., “prepare,” “review” and “finish”) to accomplish the goal is determined based on the monitored communications. The pattern may then be used to recommend actions to other users subscribed to the goal concerning accomplishing the goal based on the pattern. Furthermore, the owner may receive indications as to the progress of the user in accomplishing the goal based on the pattern.

Abstract: Conversation user interfaces that are configured for virtual assistant interaction may include tasks to be completed that may have repetitious entry of the same or similar information. User preferences may be learned by the system and may be confirmed by the user prior to the learned preference being implemented. Learned preferences may be identified in near real-time on large collections of data for a large population of users. Further, the learned preferences may be based at least in part on previous conversations and actions between the system and the user as well as user-defined occurrence thresholds.

Abstract: A security platform employs a variety techniques and mechanisms to detect security related anomalies and threats in a computer network environment. The security platform is “big data” driven and employs machine learning to perform security analytics. The security platform performs user/entity behavioral analytics (UEBA) to detect the security related anomalies and threats, regardless of whether such anomalies/threats were previously known. The security platform can include both real-time and batch paths/modes for detecting anomalies and threats. By visually presenting analytical results scored with risk ratings and supporting evidence, the security platform enables network security administrators to respond to a detected anomaly or threat, and to take action promptly.

Abstract: An apparatus is provided to interface with users to perform group assessment processing and display of results in conjunction with selection of items.

Abstract: A method and system is provided for computer-aided simulation of multi-layer selective laser sintering and melting in an additive manufacturing processes. The method may include receiving a solid model. The method may also include slicing the solid model geometry along a build direction and creating 3D meshes that represent manufacturing layers. In addition, the method may include simulating manufacture of each of the 3D meshes to produce corresponding deformed 3D meshes. Further, the method may include building a 3D mesh model from the deformed 3D meshes and displaying the 3D mesh model.

Abstract: Techniques for creating, configuring, and employing diffusion light devices are presented. Such light device(s) can comprise or be associated with a light management component (LMC) that can employ sensors to monitor environmental conditions in a defined area of people or vehicles, and a diffusion component that can diffuse or otherwise process light. At a least a portion of the diffusion component and/or a light component can be formed of a fabric that can emit light and/or diffuse light. LMC can enhance function of the light device to manage diffusion of light or perform other tasks to enhance user experience and safety and security of people or vehicles. Based on results of analyzing sensor data relating to the conditions, LMC can determine and facilitate implementing an adjustment(s) to a parameter(s) of the diffusion component or light component to achieve desired emission or diffusion of light.

Abstract: Methods for computer-aided simulation of additive manufacturing processes and corresponding systems and computer-readable mediums. A method includes receiving a finite element (FE) model of a part to be manufactured. The method includes intersecting a depositing layer line with the FE model to define an FE layer mesh that represents a manufacturing layer. The method includes simulating manufacture of the FE layer mesh and correspondingly modifying the FE model. The method includes displaying the modified FE model.

Abstract: In one embodiment, a method includes receiving, an input from a client device corresponding to n-grams; accessing a media-item index that indexes a plurality of media items with their respective keywords, each keyword being extracted from communications made by users of an online social network; identifying media items from the media-item index, each media item being identified based on its associated keywords matching the n-grams corresponding to the input; calculating a media-item score for each of the identified media items based on a number of occurrences in which the media item has been used in communications on the online social network in connection with its matched keywords; and sending, to the client device, data corresponding to the identified media items having a media-item score greater than a threshold media-item score.

Abstract: Implementations disclose predicting video start times for maximizing user engagement. A method includes receiving a first content item comprising content item segments, processing the first content item using a trained machine learning model that is trained based on interaction signals and audio-visual content features of a training set of training segments of training content items, and obtaining, based on the processing of the first content item using the trained machine learning model, one or more outputs comprising salience scores for the content item segments, the salience scores indicating which content item segment of the content item segments is to be selected as a starting point for playback of the first content item.

Abstract: A system, method, and computer-readable medium are disclosed for performing a cloud environment governance policy operation which provides a governance policy framework for use within cloud environments. In various embodiments, the cloud environment governance policy operation expands identify access management (IAM) systems to understand not only users but also resources. Additionally, in various embodiments, the cloud environment governance policy operation provides governance packs to abstract cloud resources and compliance/governance/regulatory regimes via a common open language.

Abstract: Implementations of the present disclosure include methods, systems, and computer-readable storage mediums for utilizing multiple AI service providers by a dialog management system. The dialog management system can include a dispatcher bot, multiple worker bots, and multiple AI adapters that are each associated with a different cloud-based AI service provider. In response to receiving a query, the dispatcher bot selects a particular worker bot to handle the query. The particular worker bot is assigned to a particular AI service provider. An AI adapter associated with the particular AI service provider, generates a query message based on the query. The AI adapter sends the query message to the particular AI service provider and receives a response message. The dialog management system sends a representation of the response message to the particular worker bot, receives an answer for the query from the particular worker bot, and provides the answer for output.

Abstract: Systems and methods for multiple topic automated chatting are provided. The systems and method provide multiple topic automated (or artificial intelligence) chatting by analyzing user inputs in a conversation to determine a plurality topics, to determine and score features related to the determined topics and different users, and to create a knowledge graph of the determined topics. Based on these determinations, the systems and methods may determine if a reply should be provided and then predict a reply.

Abstract: An approach is provided for securing data. Concept expansion on a data structure and names of attributes of the data is performed. Values of fields of the data are matched to a pattern in a repository that includes patterns that specify concepts. Based on the concept expansions and the field values matched to the pattern, a concept of the data is identified and an inference score is determined. The inference score indicates whether the identified score requires a validation. If the inference score exceeds a threshold score, a policy associated with the concept is identified and an action to secure the data based on the policy is performed. If the inference score does not exceed the threshold score, (i) a validation of the concept is received or (ii) an indication that the identified concept is not valid and a new concept of the data are received.

Abstract: A method of performing aspects of an energy industry operation includes simulating an energy industry operation parameter by an analysis module, wherein simulating includes predicting values of the operational parameter via one or more mathematical models of the assembly and a formation region, and estimating an uncertainty range of at least one of the input data and the predicted values of the operational parameter; generating a performance envelope based on the predicted values and the uncertainty range; generating an operating envelope by the analysis module, the operating envelope prescribing minimum constraints on the operational parameter; during the energy industry operation, receiving real time measurements of the operational parameter by a production performance module, and comparing the measured operational parameter to the operating envelope and the performance envelope; and based on the measured operational parameter being outside the performance envelope, controlling the energy industry operation

Abstract: Automatically learning and providing software development team structure and methodologies. A software development knowledgebase repository is generated by mining for software processes data over a network of computer systems. A team structure specification and project requirement associated with a target project is received. A software development methodology is selected from the software development knowledgebase repository based on the team structure specification and project requirement associated with a target project, a team members knowledgebase, a teams knowledgebase, and a past projects knowledgebase. A machine learning module automatically learns a software development methodology to select. Based on the software development methodology, a software development environment infrastructure for the target project is built.

Abstract: A cognitive system and method for optimized video playback. A semantic and sentiment analysis of the video presentation is performed to identify one or more topics, speakers, and sentiments therein. An analysis of at least one viewer is performed to identify the topics, speakers, and sentiments of interest. The topics, speakers, and sentiments are associated with a time-based mapping of the video presentation based on the analysis of the viewer. A navigation aid for the video presentation is generated using the time-based mapping, wherein the navigation aid includes entry and exit pointers to portions of the video presentation that match the viewer's areas of interest. Portions of the video presentation may be recommended and displayed to the viewer using the navigation aid.

Abstract: Recurrent neural networks (RNNs) can be visualized. For example, a processor can receive vectors indicating values of nodes in a gate of a RNN. The values can result from processing data at the gate during a sequence of time steps. The processor can group the nodes into clusters by applying a clustering method to the values of the nodes. The processor can generate a first graphical element visually indicating how the respective values of the nodes in a cluster changed during the sequence of time steps. The processor can also determine a reference value based on multiple values for multiple nodes in the cluster, and generate a second graphical element visually representing how the respective values of the nodes in the cluster each relate to the reference value. The processor can cause a display to output a graphical user interface having the first graphical element and the second graphical element.

Abstract: A computer-implemented computer documentation validation method, system, and computer program product for discovering problems and errors in an entire set of program/software documentation written by humans and flagging the documentation errors to suggest corrections to humans.

Abstract: Mechanisms are provided for implementing a virtual corpus engine that receives an inquiry to be processed and analyzes the inquiry to extract one or more features of the inquiry. The virtual corpus engine selects a weight matrix associated with a virtual corpus based on the extracted one or more features of the inquiry. The virtual corpus comprises a plurality of actual corpora of information. The weight matrix comprises a separate weight value for each actual corpus in the plurality of actual corpora. The virtual corpus engine processes the inquiry using a set of selected actual corpora selected from the plurality of actual corpora based on the weight values in the weight matrix and receives results of the processing of the inquiry using the set of selected actual corpora. The virtual corpus engine outputs the results of the processing of the inquiry.

Abstract: The disclosure describes systems and methods of ranking user interest in physical entities based on the attention given to those entities as determined by an analysis of communications from devices over multiple communication channels. The attention ranking systems allow any “Who, What, When, Where” entity to be defined and ranked based, at least in part, on information obtained from communications between users and user proxy devices. An entity rank is generated for entity known to the system in which the entity rank is derived from the information in communications that are indicative of user actions related to the entity. The entity ranks are then used to modify the display of information or data associated with the entities. The system may also generate a personal rank for each entity based on the relation of the entity to a specified user.

Abstract: A computer program product, a method, and an information handling system is provided to associate keywords used for unsuccessfully searching for a shared URL contents with a confidence level of predicting an adjustment of search engine optimization (SEO) for finding the shared URL due to changing the shared URL contents and performing an action based on the confidence level.

Abstract: Historical power output measurements of a wind turbine for a time period immediately preceding a specified time are received. Historical wind speed micro-forecasts for the wind turbine for a time periods immediately preceding the specified past time and immediately following the specified past time are received. The historical wind speed micro-forecasts are converted to wind power values. Based on the historical power output measurements and the wind power output values, a machine learning model for predicting wind power output is trained. Real-time power output measurements of the wind turbine and real-time wind speed micro-forecasts for the wind turbine are received. The real-time wind speed micro-forecasts are converted to real-time wind power values. Using the machine learning model with the real-time power output measurements and the real-time wind power values, a wind power output forecast for the wind turbine at a future time is outputted.

Abstract: Systems and methods for planning location-sensitive probabilistic behavior based evacuation paths, in a building, from source nodes to sink nodes in a network of routes including a plurality of vertices and edges are disclosed. For example, input parameters including layouts, number of evacuees at each source node, transit time, predetermined time period and maximum capacity associated with each edge and vertex are received. Further, weak evacuation schedules at a state of the evacuees are defined based on the layouts of the building, number of evacuees, transit time and predetermined time period. Furthermore, a strong evacuation schedule is defined based on the weak evacuation schedules and maximum capacity associated with each edge and vertex. Moreover, a mapping is defined from the strong evacuation schedule to the weak evacuation schedules to obtain a probabilistic behavior model. Also, an evacuation path is planned, in real-time, based on the probabilistic behavior model.

Abstract: A method described herein includes receiving a digital image, wherein the digital image includes a first element that corresponds to a first domain and a second element that corresponds to a second domain. The method also includes automatically assigning a label to the first element in the digital image based at least in part upon a computed probability that the label corresponds to the first element, wherein the probability is computed through utilization of a first model that is configured to infer labels for elements in the first domain and a second model that is configured to infer labels for elements in the second domain. The first model receives data that identifies learned relationships between elements in the first domain and elements in the second domain, and the probability is computed by the first model based at least in part upon the learned relationships.