Abstract: A computer-implemented method for performing few-shot imitation is disclosed. The method comprises obtaining at least one set of training data, wherein each set of training data is associated with a task and comprises (i) one of samples of rewards and a reward function, (ii) one of samples of state transitions and a transition distribution, and (iii) a set of first demonstrations, training a policy network embodied in an agent using reinforcement learning by inputting at least one set of first demonstrations of the at least one set of training data into the policy network, and by maximizing a risk measure or an average return over the at least one set of first demonstrations of the at least one set of training data based on respective one or more reward functions or respective samples of rewards, obtaining a set of second demonstrations associated with a new task, and inputting the set of second demonstrations and an observation of a state into the trained policy network for performing the new task.

Abstract: A method includes: initializing a list of token embeddings, each of the token embeddings corresponding to a tokenized word from text in a corpus of text; generating a graph for a group of consecutive words s from the text, said graph including binary relations between pairs of tokenized words of the group of consecutive words; selecting the token embeddings representing the words of the group of consecutive words from the list of token embeddings; computing a tensor of binary relations as the product between a matrix of the selected token embeddings and a tensor representing discourse relations, the computed tensor representing the binary relations between the pairs of tokenized words; computing a loss using the computed tensor; optimizing the list of token embeddings using the computed loss. The above may be repeated until the computed loss is within a predetermined range.

Abstract: A training system includes: a first training dataset including first entries, wherein each of the first entries includes: a first sentence; a second sentence; and an indicator of a relationship between the first and second sentences; a training module configured to: generate a second dataset including second entries based on the first entries, respectively, wherein each of the second entries includes: the first sentence of one of the first entries; the second sentence of the one of the first entries; a first surface realization corresponding to first facts regarding the first sentence; the indicator of the one of the first entries; and train a model using the second dataset and store the model in memory.

Abstract: A distributed transcoding method includes splitting image content into segments of a fixed interval; allocating the split segments to a plurality of workers and performing parallel transcoding; and concatenating the transcoded segments and merging same into one encoded file.

Abstract: Provided are a computer system for transmitting audio content to realize a user-customized being-there and a method thereof. The computer system may be configured to detect audio files that are generated for a plurality of objects at a venue, respectively, and metadata including spatial features that are set for the objects at the venue, respectively, and to transmit the audio files and the metadata for a user. An electronic device of the user may realize a being-there at the venue by rendering the audio files based on the spatial features in the metadata. That is, the user may feel a user-customized being-there as if the user directly listens to audio signals generated from corresponding objects at a venue in which the objects are provided.

Abstract: A method of performing image retrieval includes: obtaining a query image; generating a global feature descriptor of the query image by inputting the query image into a convolutional neural network (CNN) and obtaining the global feature descriptor as an output of the CNN, where parameters of the CNN are learned during training of the CNN on a batch of training images using a listwise ranking loss function and optimizing a quantized mean average precision ranking evaluation metric; determining similarities between the query image and other images based on distances between the global feature descriptor of the query image and global feature descriptors of the other images, respectively; ranking the other images based on the similarities, respectively; and selecting a set of the other images based on the similarities between the query image and the other images.

Abstract: Electronic devices and/or operating methods of the electronic device to provide visual localization based on outdoor three-dimensional (3D) map information may be provided. Such electronic devices may be configured to acquire two-dimensional (2D) image information about an outdoor environment, generate 3D map information about the outdoor environment based on the 2D image information, and determine a position of a point in the 3D map information corresponding to a query image.

Abstract: A method preprocesses a set of non-scheduled lines within a multimodal transportation network of predetermined stations, by (a) for each non-scheduled line (l) of the set of non-scheduled lines, associating, to each station (plj) of a sequence of stations ({right arrow over (p)}(l)) defining the non-scheduled line (l), at least one time interval (I(l,j)) during which a trip on the non-scheduled line (l) can depart from the station (plj); (b) for each first station (plj) of a non-scheduled line (l) which is reachable from a second station (pti) of a scheduled line, adding to a set of feasible transfers between a scheduled line and a non-scheduled line, if there exists a trip (t) on the scheduled line such that a departure time at the first station (plj) after transferring is compatible with the at least one time interval (I(l,j)) associated to the first station (plj), the earliest transfer from the second station (pti) on the trip (t) to the first station (plj); and (c) outputting the set of feasible transfers

Abstract: A system for generating whole body poses includes: a body regression module configured to generate a first pose of a body of an animal in an input image by regressing from a stored body anchor pose; a face regression module configured to generate a second pose of a face of the animal in the input image by regressing from a stored face anchor pose; an extremity regression module configured to generate a third pose of an extremity of the animal in the input image by regressing from a stored extremity anchor pose; and a pose module configured to generate a whole body pose of the animal in the input image based on the first pose, the second pose, and the third pose.

Abstract: A pre-trained source encoder generates a source encoder representation for each image of a labeled set of source images from a source domain. A target encoder generates a target encoder representation for each image of an unlabeled set of target images from a target domain. A generative adversarial network outputs a first prediction indicating whether each of the source encoder representations and each of the target encoder representations originate from the source domain or the target domain. The generative adversarial network outputs a second prediction of the latent code for each of the source encoder representations and each of the target encoder representations. The target encoder and the generative adversarial network are trained by repeatedly updating parameters of the target encoder and the generative adversarial network until one or more predetermined stopping conditions occur.

Abstract: A multi-document summarization system includes: an encoding module configured to receive multiple documents associated with a subject and to, using a first model, generate vector representations for sentences, respectively, of the documents; a grouping module configured to group first and second ones of the sentences associated with first and second aspects into first and second groups, respectively; a group representation module configured to generate a first vector representation based on the first ones of the sentences and a second vector representation based on the second ones of the sentences; a summary module configured to: using a second model: generate a first sentence regarding the first aspect based on the first vector representation; and generate a second sentence regarding the second aspect based on the second vector representation; and store a summary including the first and second sentences in memory in association with the subject.

Abstract: A method for operating a user terminal providing a video slide service includes uploading a UGC video to a server; receiving, from the server, scene metadata information for each reproduction section corresponding to the UGC video; generating a video slide file based on the scene metadata information for each reproduction section; displaying an item corresponding to the video slide file; and displaying a page screen composed of representative image information and subtitle information, for each reproduction section of the UGC video in response to a user selection on the item.

Abstract: A method and a system for conference assistance provide, as a conference assistance tool, a screen including an original text that is a text input in one of a plurality of languages set to be used in a conference and a translation text that is translated from the original text in another language.

Abstract: A comment processing method may include specifying a desired section in content that includes at least one of a text and an image in response to a selection of a user, and registering a comment input from the user to the specified section; and highlighting the section to which the comment is registered in the content, and displaying at least one comment among comments registered to the highlighted section at a point in time at which the highlighted section is displayed on a screen area.

Abstract: A method includes: selecting a set of access points in a focus zone that identifies an area around the travel destination; determining access paths through one or more access points in the set of selected access points; listing crossings that the access paths pass through; for each of the listed crossings: determining a probability that the listed crossing is passed through by an access path; determining at least one complexity score of the listed crossing; determining a set of points of interest (POIs) along access paths passing through the listed crossing; and determining quality scores for the POIs, respectively, the quality score of one of the POIs reflecting a utility of that one of the POIs to facilitate orientation; selecting POIs from the set of POIs based on the quality scores; generating the access map including the access paths and the selected POIs; and outputting the access map.

Abstract: A multiple sound source mixing method includes dividing a plurality of sound source data into segments each with a desired length; sequentially inputting sound source data of a corresponding segment for each segment through a desired number of nodes with respect to the plurality of sound source data and mixing the input sound source data into a single piece of sound source data; and concatenating the sound source data mixed for the respective segments.

Abstract: Disclosed are an electronic apparatus for recognizing a multimedia signal and an operating method of the electronic apparatus, including segmenting a detection signal into a plurality of frames; segmenting each of the frames into a plurality of blocks; and representing each of the blocks as a hash word based on a time feature and a frequency feature for each of the blocks.

Abstract: There is provided an approach for using street view images, captured from a selected geographical area, to obtain one or more of a landmark saliency score and a street crossing simplicity score, with each score reflecting a degree to which a computer-implemented circuit (including image recognition engines and a visual element matching module) can recognize and identify a landmark in at least one of the street view images. In turn, a navigational plan for a selected geographical area, including travel directions, is generated with the one or more of the landmark saliency score and the street crossing simplicity score.

Abstract: A computer implemented method for multi-document question answering is performed on a server communicating with a client device over a network. The method includes receiving a runtime question from a client device and retrieving runtime documents concerning the runtime question using a search engine. Runtime answer samples are identified in the retrieved runtime documents. A neural network model, trained using distant supervision and distance based ranking loss, is used to compute runtime scores from runtime question data representing the runtime question and from a runtime answer sample representing a first portion of text from the corpus of documents, where each runtime score represents a probability an answer to the runtime question is present in the runtime answer samples. A runtime answer is selected from the runtime answer samples corresponding to the highest runtime score sent to the client device.

Abstract: A method of determining a highlight section of a sound source includes obtaining a sound source and classification information of the sound source, and learning a neural network by using the sound source and the classification information. The neural network includes an input layer including a node corresponding to a feature value of each of a plurality of sections obtained by splitting the sound source according to a time axis, an output layer including a node corresponding to the classification information, a hidden layer defined between the input layer and the output layer, a first function between the input layer and the hidden layer, and a second function between the hidden layer and the output layer, wherein the first function includes an attention model for calculating a weighted sum of the feature value of each section. The highlight section of the sound source is determined based on weight information of a feature value node of each section included in the first function.