Abstract: A determination device includes a memory, and processing circuitry coupled to the memory and configured to accept input of a plurality of sequences provided as candidates for a solution to one given input, and determine, for two sequences of the plurality of sequences, a sequence that has a higher accuracy than the other sequence of the two sequences, using a model expressed as a neural network.

Abstract: A voice control method and display apparatus are provided. The voice control method includes converting a voice of a user into text in response to the voice being input during a voice input mode; performing a control operation corresponding to the text; determining whether speech of the user has finished based on a result of the performing the control operation; awaiting input of a subsequent voice of the user during a predetermined standby time in response to determining that the speech of the user has not finished; and releasing the voice input mode in response to determining that the speech of the user has finished.

Abstract: Methods for replacing or obscuring objects detected in an image or video on the basis of image context are disclosed. Context of the image or video may be obtained via pattern recognition on audio associated with the image or video, by user-supplied context, and/or by context derived from image capture, such as the nature of an application used to capture the image. The image or video may be analyzed for object detection and recognition, and depending upon policy, the image or video context used to select objects related or unrelated to the context for replacement or obfuscation. The selected objects may then be replaced with generic objects rendered from 3D models, or blurred or otherwise obscured.

Abstract: A processor may generate a plurality of intermediate feature layers of an image using convolutional neural network (CNN) processing. For each intermediate feature layer, the processor may generate a plurality of text proposals using a region proposal network (RPN). Each text proposal may comprise a portion of the intermediate feature layer that is predicted to contain text. The processor may perform OCR processing on image data within a plurality of regions of the image to generate a text result for each region. Each region may comprise at least one of the text proposals. The processor may assemble the text results into a text string comprising the text results ordered according to a spatial order in which the plurality of regions appear within the image.

Abstract: We propose a neural network-implemented method for base calling analytes. The method includes accessing a sequence of per-cycle image patches for a series of sequencing cycles, where pixels in the image patches contain intensity data for associated analytes, and applying three-dimensional (3D) convolutions on the image patches on a sliding convolution window basis such that, in a convolution window, a 3D convolution filter convolves over a plurality of the image patches and produces at least one output feature. The method further includes beginning with output features produced by the 3D convolutions as starting input, applying further convolutions and producing final output features and processing the final output features through an output layer and producing base calls for one or more of the associated analytes to be base called at each of the sequencing cycles.

Abstract: An ensemble of machine learning models used for real-time prediction of text for an electronic chat with an expert user. A global machine learning model, e.g., a transformer model, trained with domain specific knowledge makes a domain specific generalized prediction. Another machine learning model, e.g., an n-gram model, learns the specific style of the expert user as the expert user types to generate more natural, more expert user specific text. If specific words cannot be predicted with a desired probability level, another word level machine learning model, e.g., a word completion model, completes the words as the characters are being typed. The ensemble therefore produces real-time, natural, and accurate text that is provided to the expert user. Continuous feedback of the acceptance/rejection of predictions by the expert is used to fine tune one or more machine learning models of the ensemble in real time.

Abstract: An image forming apparatus that utilizes a communication device is disclosed. A communication unit of a housing which corresponds to an antenna portion is provided at a position at which the communication unit can be seen without obstruction from the upper side. A communication device such as a smartphone can be brought close to or into contact with the communication unit readily. Accordingly, wireless communication between the antenna portion and the communication device can be established.

Abstract: Methods, apparatus, and articles manufacture to decode documents based on images using artificial intelligence are disclosed. An example apparatus includes a model executor to input an image into a first artificial intelligence (AI)-based model to generate detected columns of text in the image; and input the image into a second AI-based model to classify the detected columns into categories; a cell identifier to identify rows or cells in the detected columns; and a report generator to: link information corresponding to the rows or cells in the detected columns with corresponding categories; and generating a report based on the linked information.

Abstract: Embodiments of the present disclosure disclose a method and apparatus for constructing a quality evaluation model, an electronic device and a computer-readable storage medium. A specific implementation mode of the method comprises: acquiring samples of knowledge contents; extracting statistical features, semantic features, and image features respectively from the samples of knowledge contents; and constructing a quality evaluation model for knowledge according to the statistical features, the semantic features, and the image features. On the basis of the prior art, this implementation mode additionally uses semantic features and image features of knowledge contents to construct a more accurate quality evaluation model based on multi-dimensional features that characterize the actual quality of a knowledge, which may well discover some brief but very useful summary knowledge in an enterprise and may recommend high-quality knowledge more accurately for employees in the enterprise.

Abstract: Systems and methods are presented for inferring an embedding vector of an item of a first type into the embedding space. Upon receiving a first time for which there is no embedding vector, documents of a document corpus that include (co-occurrence) both the received item and other items of the same type are identified. Of those other items that have embedding vectors, those embedding vectors are retrieved and averaged. The resulting averaged embedding vector is established as an inferred embedding vector for the received item.

Abstract: Systems and methods enabling rendering an avatar attuned to a user. The systems and methods include receiving audio-visual data of user communications of a user. Using the audio-visual data, the systems and methods may determine vocal characteristics of the user, facial action units representative of facial features of the user, and speech of the user based on a speech recognition model and/or natural language understanding model. Based on the vocal characteristics, an acoustic emotion metric can be determined. Based on the speech recognition data, a speech emotion metric may be determined. Based on the facial action units, a facial emotion metric may be determined. An emotional complex signature may be determined to represent an emotional state of the user for rendering the avatar attuned to the emotional state based on a combination of the acoustic emotion metric, the speech emotion metric and the facial emotion metric.

Abstract: Systems and methods are provided for analyzing messages generated by a plurality of computing devices associated with a plurality of users in a messaging system to generate training data to train a machine learning model to determine a probability that a media content item was generated inside an enclosed location or outside, receiving a media content item from a computing device, analyzing the media content item using the trained machine learning model to determine a probability that the media content item was generated inside an enclosed location or outside, determining, based on the probability generated by the trained machine learning model, that the media content item was generated inside an enclosed location, and determining an inside temperature associated with the venue based on messages generated by a plurality of computing devices in a messaging system comprising media content items and temperature information for the venue or a similar venue type.

Abstract: Techniques are described for collecting descriptions of an entity from different data sources and using a numeric comparison and textual centrality measure to automatically select a best description. In one implementation, a method includes: retrieving a real property description dataset, the real property description dataset including descriptions from multiple data sources that describe the real property; extracting, from each of the descriptions, numbers that identify the property; performing a numerical comparison of the numbers extracted from each of the descriptions to determine if any descriptions needs to be discarded from further consideration; applying a text cleaning process to normalize the descriptions; and performing a textual centrality measure of remaining descriptions to determine a level agreement of each of the remaining descriptions with each of the other remaining descriptions; and using at least the textual centrality measure to select a description.

Abstract: Certain embodiments involve using a machine-learning tool to generate metadata identifying segments and topics for text within a document. For instance, in some embodiments, a text processing system obtains input text and applies a segmentation-and-labeling model to the input text. The segmentation-and-labeling model is trained to generate a predicted segment for the input text using a segmentation network. The segmentation-and-labeling model is also trained to generate a topic for the predicted segment using a pooling network of the model to the predicted segment. The output of the model is usable for generating metadata identifying the predicted segment and the associated topic.

Abstract: A method, system, and device for audio-based identification interfaces for selecting objects from video generates and stores frequency-based audio identifiers associated with segments of an audio stream that is integrated with a video stream. The generation of the frequency-based audio identifiers may be performed by a hashing function applied to audio frequencies within audio segments. The video stream comprises identified objects that may be identified by application of a trained neural network. An audio segment is received from a user and a corresponding frequency-based audio identifier is generated and matched against stored frequency-based audio identifiers. The matching determines an audio segment and a temporally corresponding identified object, which is then embodied within an interactive user interface.

Abstract: A sentence type recognition method and apparatus, an electronic device, and a storage medium are provided. The method includes: determining a target sentence from a to-be-recognized classroom teaching speech; segmenting the target sentence into a plurality of words, and processing a word vector of each of the plurality of words, to obtain semantic representation information of the target sentence; calculating, when the plurality of candidate sentence types comprise only a question and a non-question, an inner product of the plurality of pieces of dimension information and normalizing the inner product, to obtain a probability that the sentence type corresponding to the semantic representation information is the question; and determining, based on the probability, that the sentence type corresponding to the semantic representation information is the question or the non-question; determining the sentence type corresponding to the semantic representation information as a sentence type of the target sentence.

Abstract: Embodiments of a method and an apparatus for improving a model based on a pre-trained semantic model are provided. The method may include: based on the pre-trained semantic model, obtaining an initial improved model, where semantic result information of an input vector is determined in the initial improved model based on a hash search method; and based on a model distillation method, training the initial improved model to obtain an improved model. Some embodiments can obtain the semantic result information of the input vector by performing the hash search method on the input vector, replace the original complex iterative calculation process of a semantic model, and obtain the improved model with few model parameters and high compression ratio.

Abstract: A method includes obtaining user input interaction data. The user input interaction data includes one or more user interaction input values respectively obtained from the corresponding one or more input devices. The user input interaction data includes a word combination. The method includes generating a user interaction-style indicator value corresponding to the word combination in the user input interaction data. The user interaction-style indicator value is a function of the word combination and a portion of the one or more user interaction input values. The method includes determining, using a semantic text analyzer, a semantic assessment of the word combination in the user input interaction data based on the user interaction-style indicator value and a natural language assessment of the word combination. The method includes generating a response to the user input interaction data according to the user interaction-style indicator value and the semantic assessment of the word combination.

Abstract: Dynamic content tags are generated as content is received by a dynamic content tagging system. A natural language processor (NLP) tokenizes the content and extracts contextual N-grams based on local or global context for the tokens in each document in the content. The contextual N-grams are used as input to a generative model that computes a weighted vector of likelihood values that each contextual N-gram corresponds to one of a set of unlabeled topics. A tag is generated for each unlabeled topic comprising the contextual N-gram having a highest likelihood to correspond to that unlabeled topic. Topic-based deep learning models having tag predictions below a threshold confidence level are retrained using the generated tags, and the retrained topic-based deep learning models dynamically tag the content.

Abstract: A system for measuring total operating deflection shapes of a structure includes one or more imagers, each including two cameras spaced apart from one another and each oriented and positioned to have corresponding fields of view of a different corresponding section of the structure, with the corresponding sections that may include overlap area of the structure within each of the different sections of the structure. Each of the cameras generates a corresponding data stream, which is communicated to a controller, which is configured to measure the response of the structure to an excitation, such as a vibration or an impulse. The system is configured to convert time-domain data from each of the data streams to the frequency-domain data using a Fourier Transform algorithm and stitching the shapes to obtain the total operating deflection shapes of the structure by scaling and stitching together the frequency-domain data.