Abstract: Systems and methods for color prediction are described. Embodiments of the present disclosure receive an image that includes an object including a color, generate a color vector based on the image using a color classification network, where the color vector includes a color value corresponding to each of a set of colors, generate a bias vector by comparing the color vector to teach of a set of center vectors, where each of the set of center vectors corresponds to a color of the set of colors, and generate an unbiased color vector based on the color vector and the bias vector, where the unbiased color vector indicates the color of the object.

Abstract: Systems and methods for intent discovery and video summarization are described. Embodiments of the present disclosure receive a video and a transcript of the video, encode the video to obtain a sequence of video encodings, encode the transcript to obtain a sequence of text encodings, apply a visual gate to the sequence of text encodings based on the sequence of video encodings to obtain gated text encodings, and generate an intent label for the transcript based on the gated text encodings.

Abstract: Multitask machine-learning model training and training data augmentation techniques are described. In one example, training is performed for multiple tasks simultaneously as part of training a multitask machine-learning model using question pairs. Examples of the multiple tasks include question summarization and recognizing question entailment. Further, a loss function is described that incorporates a parameter sharing loss that is configured to adjust an amount that parameters are shared between corresponding layers trained for the first and second tasks, respectively. In an implementation, training data augmentation techniques are also employed by synthesizing question pairs, automatically and without user intervention, to improve accuracy in model training.

Abstract: Systems and methods are provided for providing a navigation interface to access or otherwise use electronic content items. In one embodiment, an augmentation application identifies at least one entity referenced in a document. The entity can be referenced in at least two portions of the document by at least two different words or phrases. The augmentation application associates the at least one entity with at least one multimedia asset. The augmentation application generates a layout including at least some content of the document referencing the at least one entity and the at least one multimedia asset associated with the at least one entity. The augmentation application renders the layout for display.

Abstract: Systems and methods for key-phrase extraction are described. The systems and methods include receiving a transcript including a text paragraph and generating key-phrase data for the text paragraph using a key-phrase extraction network. The key-phrase extraction network is trained to identify domain-relevant key-phrase data based on domain data obtained using a domain discriminator network. The systems and methods further include generating meta-data for the transcript based on the key-phrase data.

Abstract: Embodiments of the present invention provide systems, methods, and non-transitory computer storage media for parsing a given input referring expression into a parse structure and generating a semantic computation graph to identify semantic relationships among and between objects. At a high level, when embodiments of the preset invention receive a referring expression, a parse tree is created and mapped into a hierarchical subject, predicate, object graph structure that labeled noun objects in the referring expression, the attributes of the labeled noun objects, and predicate relationships (e.g., verb actions or spatial propositions) between the labeled objects. Embodiments of the present invention then transform the subject, predicate, object graph structure into a semantic computation graph that may be recursively traversed and interpreted to determine how noun objects, their attributes and modifiers, and interrelationships are provided to downstream image editing, searching, or caption indexing tasks.

Abstract: Systems and methods for natural language processing are described. One or more aspects of a method, apparatus, and non-transitory computer readable medium include receiving a text phrase; encoding the text phrase using an encoder to obtain a hidden representation of the text phrase, wherein the encoder is trained during a first training phrase using self-supervised learning based on a first contrastive loss and during a second training phrase using supervised learning based on a second contrastive learning loss; identifying an intent of the text phrase from a predetermined set of intent labels using a classification network, wherein the classification network is jointly trained with the encoder in the second training phase; and generating a response to the text phrase based on the intent.

Abstract: Embodiments of the present invention provide systems, methods, and non-transitory computer storage media for parsing a given input referring expression into a parse structure and generating a semantic computation graph to identify semantic relationships among and between objects. At a high level, when embodiments of the preset invention receive a referring expression, a parse tree is created and mapped into a hierarchical subject, predicate, object graph structure that labeled noun objects in the referring expression, the attributes of the labeled noun objects, and predicate relationships (e.g., verb actions or spatial propositions) between the labeled objects. Embodiments of the present invention then transform the subject, predicate, object graph structure into a semantic computation graph that may be recursively traversed and interpreted to determine how noun objects, their attributes and modifiers, and interrelationships are provided to downstream image editing, searching, or caption indexing tasks.

Abstract: Methods and systems are provided for identifying subparts of a text. A neural network system can receive a set of sentences that includes context sentences and target sentences that indicate a decision point in a text. The neural network system can generate context vector sentences and target sentence vectors by encoding context from the set of sentences. These context sentence vectors can be weighted to focus on relevant information. The weighted context sentence vectors and the target sentence vectors can then be used to output a label for the decision point in the text.

Abstract: Techniques are disclosed for abstractive summarization process for summarizing documents, including long documents. A document is encoded using an encoder-decoder architecture with attentive decoding. In particular, an encoder for modeling documents generates both word-level and section-level representations of a document. A discourse-aware decoder then captures the information flow from all discourse sections of a document. In order to extend the robustness of the generated summarization, a neural attention mechanism considers both word-level as well as section-level representations of a document. The neural attention mechanism may utilize a set of weights that are applied to the word-level representations and section-level representations.

Abstract: Techniques of directing a user to content based on a semantic interpretation of a query input by the user involves generating links to specific content in a collection of documents in response to user string query, the links being generated based on an answer suggestion lookahead index. The answer suggestion lookahead index references a mapping between a plurality of groups of semantically equivalent terms and a respective link to specific content of the collection of documents. These techniques are useful for the generalized task of natural language question answering.

Abstract: Embodiments of the present invention provide systems, methods, and non-transitory computer storage media for parsing a given input referring expression into a parse structure and generating a semantic computation graph to identify semantic relationships among and between objects. At a high level, when embodiments of the preset invention receive a referring expression, a parse tree is created and mapped into a hierarchical subject, predicate, object graph structure that labeled noun objects in the referring expression, the attributes of the labeled noun objects, and predicate relationships (e.g., verb actions or spatial propositions) between the labeled objects. Embodiments of the present invention then transform the subject, predicate, object graph structure into a semantic computation graph that may be recursively traversed and interpreted to determine how noun objects, their attributes and modifiers, and interrelationships are provided to downstream image editing, searching, or caption indexing tasks.

Abstract: Computer-readable media, computer systems, and computing devices of a method for facilitating providing offers utilizing population densities are provided. In embodiments, a population density for a geographical space is determined based on locations of a plurality of user devices. The population density associated with the geographical space is used to determine to provide an electronic offer to a user. The electronic offer may be associated with an item in the geographical space to entice the user to move to the geographical space. In accordance with determining to provide an electronic offer to the user, the electronic offer is provided for viewing by the user via a user device.

Abstract: In techniques for causal modeling and attribution, a causal modeling application implements a dynamical causal modeling framework. Input data is received as a representation of communications between users, such as social media interactions between social media users, and causal relationships between the users can be determined based in part on the input data that represents the communications. Influence variables, such as exogenous variables and/or endogenous variables, can also be determined that influence the causal relationships between the users. A causal relationships model is generated based on the influence variables and the causal relationships between the users, where the causal relationships model is representative of causality, influence, and attribution between the users.

Abstract: Systems and methods are disclosed for augmenting or summarizing an information processing task in a bilateral Q&A format. An initial query is received from an analyst user to conduct an analysis on user-provided data attributes. A primary result is generated from processing the initial query. In addition, defined data attributes associated with the primary result are ranked to create relevant follow-up queries. A summary of the primary result, using both graphical representations and natural language summaries, are provided to the analyst user. The relevant follow-up queries can also be provided to the analyst user, thereby progressing a contextually-based conversation regarding the data. The analytics session can progress as the user traverses the results and follow-up queries, until the user terminates the session or all relevant follow-up queries are exhausted.

Abstract: Methods and systems are provided for identifying subparts of a text. A neural network system can receive a set of sentences that includes context sentences and target sentences that indicate a decision point in a text. The neural network system can generate context vector sentences and target sentence vectors by encoding context from the set of sentences. These context sentence vectors can be weighted to focus on relevant information. The weighted context sentence vectors and the target sentence vectors can then be used to output a label for the decision point in the text.

Abstract: A tutorial for a given application may be leveraged to generate executable code that can then be executed within a native instruction service of the application. In this way, a software application may thus provide an integrated, interactive learning experience for a user, in a manner that extends beyond the instructional content included in the native instruction service, i.e., that includes at least a portion of the instructional content of the tutorial.

Abstract: Natural language and user interface control techniques are described. In one or more implementations, a natural language input is received that is indicative of an operation to be performed by one or more modules of a computing device. Responsive to determining that the operation is associated with a degree to which the operation is performable, a user interface control is output that is manipulable by a user to control the degree to which the operation is to be performed.

Abstract: A framework for annotating image edit requests includes a structure for identifying natural language request as either comments or image edit requests and for identifying the text of a request that maps to an executable action in an image editing program, as well as to identify other entities from the text related to the action. The annotation framework can be used to aid in the creation of artificial intelligence networks that carry out the requested action. An example method includes displaying a test image, displaying a natural language input with selectable text, and providing a plurality of selectable action tag controls and entity tag controls. The method may also include receiving selection of the text, receiving selection of an action tag control for the selected text, generating a labeled pair, and storing the labeled pair with the natural language input as an annotated natural language image edit request.

Abstract: Techniques of directing a user to content based on a semantic interpretation of a query input by the user involves generating links to specific content in a collection of documents in response to user string query, the links being generated based on an answer suggestion lookahead index. The answer suggestion lookahead index references a mapping between a plurality of groups of semantically equivalent terms and a respective link to specific content of the collection of documents. These techniques are useful for the generalized task of natural language question answering.