Message composition and customization in a user's handwriting style includes obtaining electronic source text from a user, the electronic source text to be sent to a recipient, ascertaining properties of the electronic source text, the properties including words used in the electronic source text and a context of the electronic source text, and the context including an emotion of the electronic source text, and building an electronic message based on the ascertained properties, the electronic message including the electronic source text presented graphically in a handwriting style of the user.

Certain aspects of the present disclosure provide techniques for recommending trip purposes to users of an application. Embodiments include receiving labeled travel data from the application running on a remote device including a plurality of trip purposes. Embodiments include building a topic model representing words associated with a plurality of topics. Embodiments include training a topic prediction model, using the plurality of topics and one or more features derived from each of the plurality of trip records, to output a topic based on an input trip record. Embodiments include training a purpose prediction model, using the topic model and the plurality of trip purposes, to output a trip purpose based on an input topic. The trip purpose may be recommended to a user via a user interface of the application running on the remote device.

Techniques are described relating to data distribution and security in a multilayer storage infrastructure. An associated computer-implemented method includes receiving file data associated with a user for storage in a managed services domain, applying an ensemble learning model to devise a data distribution technique for the file data based upon contextual information associated with the user, and encrypting the file data. The method further includes, based upon the data distribution technique, dividing the file data to store among a cloud computing layer, a fog computing layer, and a local computing layer by performing a hash transformation and applying at least one cyclic error correcting code. In an embodiment, the method further includes receiving a data access request associated with the file data, authenticating the data access request, and restoring the file data via decryption.

Embodiments of the present disclosure are directed to a method and an apparatus for generating training data for a visual question answering (VQA) system, and a computer readable medium. The method for generating training data for a visual question answering system includes: obtaining a first set of training data of the visual question answering system, the first set of training data comprising a first question for an image in the visual question answering system and a first answer corresponding to the first question; obtaining information related to the image; generating a second question corresponding to the first answer based on the information to obtain a second set of training data for the image in the visual question answering system, the second set of training data comprising the second question and the first answer.

Embodiments include methods, systems and computer program products for performing named entity recognition. Aspects include obtaining a text having a plurality of words and comparing each of the plurality of words to a dictionary. Aspects also include creating, based on the comparison, an annotation for at least one of the plurality of words that the least one of the plurality of words refers to a named entity. Aspects further include parsing the text to identify a part of speech for each of the plurality of words and removing the annotations from each of the at least one of the plurality of words that has a part of speech that is not one or a noun and a noun supporting adjective.

Provided are methods, systems, and devices for generating semantic objects and an output based on the detection or recognition of the state of an environment that includes objects. State data, based in part on sensor output, can be received from one or more sensors that detect a state of an environment including objects. Based in part on the state data, semantic objects are generated. The semantic objects can correspond to the objects and include a set of attributes. Based in part on the set of attributes of the semantic objects, one or more operating modes, associated with the semantic objects can be determined. Based in part on the one or more operating modes, object outputs associated with the semantic objects can be generated. The object outputs can include one or more visual indications or one or more audio indications.

The disclosed embodiments describe methods, systems, and apparatuses for determining user intent. A method is disclosed comprising obtaining a session text of a user; calculating, by the processor, a feature vector based on the session text; determining probabilities that the session text belongs to a plurality of intent labels, the probabilities calculated using a multi-level hierarchal intent classification model, the intent labels assigned to levels in the multi-level hierarchal intent classification model; and assigning a user intent to the session text based on the probabilities.

A method and apparatus for joint image and per-pixel annotation synthesis with a generative adversarial network (GAN) are provided. The method includes: by inputting data to a generative adversarial network (GAN), obtaining a first image from the GAN; inputting, to a decoder, a first feature value that is obtained from at least one intermediate layer of the GAN according to the inputting of the data to the GAN; and obtaining a first semantic segmentation mask from the decoder according to the inputting of the first feature value to the decoder.

A learning apparatus learns a machine learning model for performing semantic segmentation of determining a plurality of classes in an input image in units of pixels by extracting, for each layer, features which are included in the input image and have different frequency bands of spatial frequencies. A learning data analysis unit analyzes the frequency bands included in an annotation image of learning data. A learning method determination unit determines a learning method using the learning data based on an analysis result of the frequency bands by the learning data analysis unit. A learning unit learns the machine learning model via the determined learning method using the learning data.

There are numerous features in video that can be detected using computer-based systems, such as objects and/or motion. The detection of these features, and in particular the detection of motion, has many useful applications, such as action recognition, activity detection, object tracking, etc. The present disclosure provides a neural network that learns motion from unlabeled video frames. In particular, the neural network uses the unlabeled video frames to perform self-supervised hierarchical motion learning. The present disclosure also describes how the learned motion can be used in video action recognition.