Abstract: The embodiments of the disclosure relates to a comment display method, apparatus, equipment and medium of a document, wherein the method comprises: in response to a comment operation of a first user on an online document, displaying a first comment(s) in a first area of the online document; in response to a trigger operation of a second user on a comment mode control in the online document, determining at least one second comment to be displayed from the first comment(s); and displaying the at least one second comment in the first area of the online document.

Abstract: An information processing apparatus input a document represented by hierarchical structure including a plurality of drawing commands; analyze each of the plurality of drawing commands of respective hierarchical levels corresponding to from a rearmost side to a frontmost side in the document, starting from the rearmost side; determine whether or not the drawing command being analyzed is a command for drawing a graphic object; generate a rendered image of the graphic object by rendering the graphic object in a case where the determining determines that the drawing command is a command to draw the graphic object; and perform character recognition processing on the rendered image.

Abstract: Methods and systems for manual and programmatic remediation of websites. JavaScript code is accessed by a user device and optionally calls TTS, ASR, and RADAE modules from a remote server to thereby facilitate website navigation by people with diverse abilities.

Abstract: A variety of data structures, components, runtime viewers and methods are described for defining, delivering and rendering wrapped packages of cards in a manner particularly well suited, but not limited to, display on mobile devices. Each card may selectively include media content, a palette of application functionality and/or supporting e-commerce related services. A wrap descriptor, composed of a plurality of card descriptors defines the structure, layout and content of a set of cards that constitute the wrap. The wrap descriptor may take the form of a JSON data object. A runtime viewer that executes on the rendering computing device may be used to transform the wrap descriptor into a runtime instance of the wrap that includes the content and functionality defined by the descriptor. In some implementations, the runtime viewer creates an object graph, which may then be converted to an object model suitable for rendering on the consuming device.

Abstract: Embodiments disclosed herein provide for dynamically modifying image and/or video content based on user preferences. Embodiments provide a server, a database containing user preference information and one or more content templates, and a content generator comprising a plurality of generative adversarial networks, wherein each generative adversarial network is associated with a corresponding style transfer, and the content generator is configured to apply one or more style transfers to a base image to convert the base image into a desired style. Once created, the customized content is sent to a client device to for display.

Abstract: Systems and methods for generating a report, including acquiring data for topics of interest from a plurality of data sources. Content from the acquired data is extracted, prioritized, and categorized into corresponding categories and subcategories based on a predetermined hierarchical set of priority rules, and a writing style and report threshold levels are selected as constraints for generating a customized report based on the prioritized and categorized content. A final report draft can be iteratively generated by generating report drafts by sequentially utilizing the prioritized and categorized content from a highest priority level to a lowest priority level for the categories and subcategories until one or more report threshold level constraints are reached. An overall quality score for the final report draft is determined based on a category content accuracy score for each of the categories determined by calculating an average of subcategory content accuracy scores for each related subcategory.

Abstract: A computing device receives user interaction with a portion of a data visualization that is displayed on the computing device. In response to receiving the user interaction, the computing device generates a first prompt that includes (i) a first parameter specifying a visualization type of the data visualization, (ii) a second parameter specifying a data array corresponding to data marks of the data visualization, and (iii) a third parameter specifying a title of the data visualization. The computing device inputs the first prompt into a large language model (LLM) and obtains, from the LLM, a text narrative for the portion of the data visualization. The device generates a second prompt in accordance with the text narrative, inputs the second prompt into the LLM, and obtains from the LLM a title for the text narrative. The device generates a self-contained story based on the text narrative and the title.

Abstract: Described herein is a computer implemented method. The method includes accessing, by a computer system including a processing unit, portable document format (PDF) data defining a plurality of glyphs, sorting the plurality of glyphs into one or more glyph sets, and calculating an expanded glyph bounding box for each glyph. Each glyph set is processed to determine one or more text areas, each text area being associated with one or more glyphs from the glyph set which have collectively overlapping expanded bounding boxes.

Abstract: The technology disclosed relates to accessing external data, including massive amounts of data stored in the cloud, in spreadsheet cells: accessing external data direct via a formulaic variable in a spreadsheet, specifying an ordered progression for the accessed external data, selectively propagating data accessed using the formulaic variable vertically or horizontally, within a propagation pattern responsive to normal A$1, $A1 and $A$1 spreadsheet conventions. Two or more external data fields, responsive to the formulaic variable, have an ordered sequence relationship that nests ordering of vectors of the propagated data; and the ordering according to the ordered sequence relationship is maintained during replication by copy and paste. In another disclosed method, the external data is generated using an implicit join of data from at least two external data sources to generate multiple adjoining vectors of spreadsheet cells of data responsive to selection parameters in the formulaic variable.

Abstract: Disclosed is a method of as you type error checking of a spreadsheet formula as a user types the formula in a cell, including analyzing, as you type, the cell contents of the spreadsheet formula and throwing an error message when the cell contents are not valid in the formula being typed. The error message identifies/emphasizes a token or group of tokens that cause the error. Also disclosed is selecting and populating into a spreadsheet formula a formulaic data field to be evaluated, starting from a position in the spreadsheet formula where the formulaic data field to be evaluated is to be populated, and responding to typed inputs and/or UI selection by displaying a list of selectable formulaic data fields with additional information that contains (non-data derived) human generated values. Upon selection among the fields, the method includes automatically populating into the starting position in the spreadsheet formula.

Abstract: Systems and techniques for insight generation and presentation are described to generate natural language descriptions that describe relationships among data in a database. A user input is processed to determine a data analysis task applicable to the user input and the database. The data analysis task is performed by a machine learning model to retrieve a dataset from the database that is relevant to the user input. The dataset is processed by a machine learning model to generate natural language descriptions, and a presentation page is generated that incorporates the natural language descriptions in order to present insightful data relationships in an easy to consume, natural language format.

Abstract: Various embodiments provide methods and systems for generating an article. The method is performed by a server system. The method includes receiving a first input corresponding to selection of an article generation method. In addition, the method includes accessing a template. Further, the method includes receiving a second input corresponding to selection of one or more components from a set of components. Furthermore, the method includes embedding the selected one or more components in the template. Moreover, the method includes receiving at least one Artificial Intelligence (AI) based recommendation for generating the article. Also, the method includes generating the article based, at least in part, on the embedded one or more components and the at least one AI based recommendation.

Abstract: An embodiment of the present invention is directed to a natural language processing platform that builds a Data Dictionary and a Knowledge Graph specific to a particular domain, such as a mortgage domain. The Natural Language Processor (NLP) Engine of an embodiment of the present invention is directed to maintaining an innovative hierarchy that specifies conditions and actions. Data extraction may be performed in a manner that keeps the hierarchy intact. By keeping the hierarchy intact, an output from a data-extraction process ensures that the vertical top-down placement of sentences (e.g., plain, bulleted, indented) is consistent with the input documents.

Abstract: Example solutions for performing key phrase extraction from content items using a large language model (LLM) include: determining a token entropy score for a first token of a content item containing text content by generating and submitting a prompt to the LLM, that includes prefix tokens preceding a first token, receiving a probability distribution from the LLM, and generating a token entropy score for the first token; identifying a candidate phrase that includes one or more tokens, each token having an associated token entropy score; computing a phrase entropy score for the candidate phrase based on the token entropy scores of the one or more tokens; storing the candidate phrase as a key phrase of the content item upon the phrase entropy score exceeding a threshold; and searching a database of content items based on the key phrase, the search returning results including the content item.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods for training and utilizing a transcript segmentation neural network to segment audio recording transcripts. In particular, in one or more embodiments, the disclosed systems implement a pretraining technique for a transcript segmentation neural network based on a specialized dataset that includes contextual information about stored sentences or conversations. For example, the discloses system train a transcript segmentation neural network based on contextual data that indicates semantic similarities and/or distances between sentences of a digital document. In some cases, the disclosed systems also (or alternatively) train a transcript segmentation neural network based on curricular data generated by a classification model. Moreover, in some cases, the disclosed system use the trained transcript segmentation neural network to generate a segmented audio transcript.

Abstract: A method for electronic decomposition of a data string into structurally meaningful parts. The method includes receiving the data string as an input data and splitting the data string into a plurality of tokens. The method further includes performing an inter token modelling on the plurality of tokens to obtain a first probability of arrangement of tokens with respect to each other and performing an intra token modelling on the plurality of tokens to obtain a second probability indicative of a number of occurrences of each individual token. The method further includes combining the first probability of arrangement of tokens, with the second probability to obtain a final probability of an arrangement for the data string and displaying one or more arrangements for the data string based on the final probability. The method is advantageous to remove ambiguity while determining structurally meaningful parts of the data string.

Abstract: A computerized method categorizes questions from a transcript and provides those categorized questions to targets. A transcript associated with a meeting is obtained and a question in the obtained transcript is detected by a question detection model. A question vector embedding of the detected question is generated using a question clustering model and a category of the detected question is determined by the question clustering model using the generated question vector embedding and a plurality of category vector embeddings. A target of the detected question is identified using the determined category and the detected question is provided to the identified target via a question provision interface. Further, the question clustering model is used to generate the category vector embeddings from example questions of category question catalogs, such that the set of categories for which the question clustering model is configured can be efficiently modified.

Abstract: The present disclosure relates to methods and systems for self-teaching a large language model (LLM). The methods and systems use a self-learning framework with a plurality of phases. In each phase of the self-learning framework, the LLM generates a diverse set of outputs for a question and an aggregation is performed on the diverse set of outputs generate a phase output. The phase output from a previous phase is used as an input to the LLM in a next phase.

Abstract: The present disclosure relates to methods and systems that generate a confidence score for the generated large language model (LLM) output. The methods and systems use the text of the input provided to the LLM and the text from the generated LLM output to produce a feature vector that encodes a readability of the text from the input and the text of the LLM output. The feature vector is used to determine a corresponding confidence score for the generated LLM output. The confidence score is used to evaluate a quality of the generated LLM output.

Abstract: A technique supplements a language model with knowledge information retrieved from external sources. The technique operates by: receiving a query; receiving knowledge information based on the query; generating original model-input information that includes the query and the knowledge information; and presenting the original model-input information to the language model. The technique further includes: receiving an original response from the language model; generating a usefulness measure that identifies usefulness of the original response; and determining whether the usefulness measure satisfies a prescribed test. Upon determining that the usefulness measure does not satisfy the test, the technique includes: generating revised model-input information that includes feedback information; presenting the revised model-input information to the language model; and receiving a revised response from the language model.

Abstract: A method, computer system, and a computer program product for classification training are provided. A deep learning model is trained with a first dataset that includes annotated samples that include a first and second class. Last hidden state features corresponding to respective tokens from prototypes of the first and second class are saved. The trained deep learning model is further trained with a second dataset that includes additional annotated samples that include the first and second class and a third class. The further training includes performing a cosine similarity loss optimization and a cross entropy loss optimization. The cosine similarity loss optimization is of last hidden state features resulting from the further trained deep learning model and the first class and the second class compared to the saved last hidden state features. The cross entropy loss optimization is for classification of the first, second, and the third class.

Abstract: The disclosure relates to systems and methods of improved attention for neural networks. A system may access a first array of entries and a second array of entries, wherein the first array represents a first word embedding generated from one or more first characters of an input and the second array represents a second word embedding from the one or more second characters of the input. A system may generate at least a first difference matrix based on the first array and at least a second difference matrix based on the second array. A system may determine a difference value based on vector based at least in part on the difference value. A system may provide the context vector to subsequent layers of a neural network.

Abstract: Disclosed herein are methods, systems, and computer-readable media for automatically classifying and moderating content. In an embodiment, a method may include receiving input data and one or more content policies, and generating a content taxonomy. The method may also include receiving multi-domain cold start data and generating training data. The method may also include accessing a language model based on the input data and the training data, and iteratively classifying the content of the input data using the language model and the content taxonomy, refining the training data based on the classified content of the input data, refining the language model based on the refined training data, probing the refined language model, and updating the threshold value based on the probing of the refined language model. The method may also include moderating the content of the input data based on the optimized language model and the content taxonomy.

Abstract: A computing system for revising large language model (LLM) input prompts is provided herein. In one example, the computing system includes at least one processor configured to receive, via a prompt interface, a prompt from a user including an instruction for a trained LLM to generate an output, and generate a first response to the prompt. The at least one processor is configured to assess the first response according to assessment criteria to generate an assessment report for the first response, and generate a revised prompt in response to second input including the first prompt, the first response, the assessment report, and a prompt revision instruction for the LLM to revise the prompt in view of the assessment report. The at least one processor is configured to, in response to final input including the revised prompt, generate a final response to the revised prompt, and output the final response.

Abstract: Disclosed is an improved approach to review and analysis a history of recorded facts for automation runs. This can be used to improve automation accuracy and to create dynamic dashboards.

Abstract: Disclosed is an improved approach to implement software upgrades, exports, and/or imports. Natural language processing and code generation are employed to implement the improved software upgrades, exports, and/or imports.

Abstract: Disclosed is an improved approach to implement an improved approach to retrospectively examiner and edit facts for an automation run. The user can then continue processing from the point of the fact change, rather than being required to restart the entire process from the very beginning.

Abstract: Provided are computer implemented systems and methods for generating a user interface for language translation, including: providing domain specific machine translation models; generating a machine translation user interface comprising an input text element; outputting the machine translation user interface; receiving user input text at the input text element in a first language from a first user; selecting a selected domain specific machine translation model in the domain specific machine translation models by applying a machine selector model of the machine selector module, the machine selector model for selecting the selected domain specific machine translation model by classifying the input text element as in-domain for the selected domain specific translation model; translating the input text element from a first language into a second language as an output text element based on the selected domain specific machine translation model; and updating the machine translation user interface to comprise the out

Abstract: In implementations of systems for generating digital content, a computing device implements a generation system to receive a user input specifying a characteristic for digital content. The generation system generates input text based on the characteristic for processing by a first machine learning model. Output text generated by the first machine learning model based on processing the input text is received. The output text describes a digital content component. The generation system generates the digital content component by processing the output text using a second machine learning model. The generation system generates the digital content including the digital content component for display in a user interface based on the characteristic.

Abstract: Methods and systems are disclosed herein for a context assistance system converting a reference in a content item based on context associated with the content item. For example, the context assistance system identifies a first entity referenced in the content item associated with a first context. The context assistance system then determines at least one attribute of the first entity. Next, the context assistance system identifies a second context in which the content item is to be presented. The context assistance system determines the second entity based on the second context and on the at least one attribute of the first entity. The context assistance system replaces reference to the first entity in the content item with reference to the second entity.

Abstract: The output of a machine learning model may be customized for a particular application using context satisfying the requirements of that application. This may be done using a method comprising receiving an artificial intelligence request, retrieving a set of exemplars corresponding to the artificial intelligence request, generating an artificial intelligence prompt comprising the exemplars and a payload of the artificial intelligence request, and obtaining an artificial intelligence response by providing the artificial intelligence prompt to a trained machine learning model.

Abstract: An improved translation experience is provided by a system that listens for speech input spoken in a first language or a second language; receives, in response to the listening for the speech input, first speech input spoken in the first language; determines that the first speech input is spoken in the first language; and generates and presents, in association with the receiving of the first speech input, a translation of the first speech input in the second language. The system also detects an endpoint in the first speech input, and, in response to detecting the endpoint, listens for additional speech input spoken in the first language or in the second language. Corresponding systems, methods, and media are also disclosed.

Abstract: Barcode reader and cradle assemblies for inductive charging include a reader for capturing at least one image of an object and a cradle that retains the reader. The reader includes a head portion including an opening at least partially defining the FOV, a handle portion, a foot portion forming a tapered foot region, a rechargeable power source coupled with the reader to be electrically charged and to supply an electric charge to power the reader, and a first coil positioned within or around the tapered foot region coupled with the rechargeable power source. The cradle includes a body portion including a tapered body region and a second coil positioned therewithin or therearound that generates an oscillating electromagnetic field in response to an alternating current. The tapered foot region removably couples with the tapered body region to retain the reader. When the reader is retained within the cradle, at least one of the first or second coils is nested within the other of the coils.

Abstract: An information code reading device includes a processor and a memory, wherein the processor executes a program stored in the memory to: perform reading processing by using electric power supplied from an external power supply through a cable; wirelessly transmit a reading result to a predetermined external device; determine whether power supply through the cable is shut off; and perform a predetermined notification using electric power supplied from a power storage circuit element in response to determination that the power supply is shut off, the reading processing is processing for optically reading an information code, and the information code reading device comprises the power storage circuit element.

Abstract: An example of an apparatus to track assets is provided. The apparatus includes a reader interface to connect to a radio-frequency identification reader. The apparatus further includes a first antenna interface to connect to a first antenna. In addition, the apparatus includes a second antenna interface to connect to a second antenna. Furthermore, the apparatus includes a switch to connect the first antenna interface to the reader interface in a first state and to connect the second antenna interface to the reader interface in a second state. The radio-frequency identification reader is to transmit a signal via the first antenna and to receive a first response from a tag via the first antenna in the first state. The radio-frequency identification reader is also to transmit and receive the signal via the second antenna in the second state. The apparatus also includes a switch controller to operate the switch automatically.

Abstract: A display control system includes a first wireless device attached to a management target and storing first identification information for identifying the management target, a second wireless device installed at a predetermined position and storing second identification information, a reading apparatus, at least one processor, and at least one memory. The processor and memory are configured to estimate a located position of the management target based on reading results of the first identification information at a first reading time and the second identification information at a second reading time and cause a display apparatus to display an estimated area within which the management target is estimated to be located based on the located position. A magnitude of the estimated area is based on an amount of movement of the reading apparatus between the first reading time and the second reading time.

Abstract: A gaming table contains a number of antennas that are de-tuned from the resonant frequency of a single RFID tag. This increases the power required to read the single tag but matches the resonant frequency for reading a stack of RFID tags. The gaming table may further include a network analyzer and a set of capacitors that are dynamically switched among the antennas according to measuring the reflection coefficient of the antennas.

Abstract: This disclosure presents an indicia reader equipped with two distinct imaging assemblies and a pair of mirrors, one of which is dichroic. This reader can generate both two-dimensional and three-dimensional image data. The dichroic mirror is designed to pass light in a specific wavelength range while reflecting light in another range, effectively directing different light towards the appropriate imaging assembly. This novel design enables efficient packaging of essential hardware, enhancing the reader's functionality for industries that require item tracking.

Abstract: Indicia readers with multiple imaging assemblies and dichroic mirrors are disclosed herein. In an example, an indicia reader like a barcode reader includes a first (2D) imaging assembly operable to sense light in the 700 nm or below wavelength range, a second (3D) imaging assembly operable to sense light in the above 700 nm wavelength range, and a dichroic mirror. The mirror is situated such that (i) light that is received by the first imaging assembly is received from a product scanning region and is reflected by the mirror such that its path is redirected towards and the first imaging assembly and (ii) light that is received by the second imaging assembly is received from the product scanning region by passing through the mirror toward the second imaging assembly.

Abstract: The present disclosure provides new and innovative systems, methods, and apparatuses for detecting non-decode events in indicia scanning settings.

Abstract: Systems and methods for adjusting scan volume are provided. An example includes an indicia reading system comprising a two-dimensional (2D) imaging assembly having at least one 2D imaging sensor and operable to capture 2D image data over at least one 2D field of view (FOV); a three-dimensional (3D) imaging assembly having at least one depth sensor and operable to capture depth data over at least one 3D FOV; and a processor configured to: perform a decoding operation on an indicia appearing in the 2D image data and transmit a payload of the indicia to a host when (i) the indicia is captured in the 2D image data and (ii) at least one of the indicia or an object bearing the indicia is positioned within a 3D scan volume.

Abstract: A method for identifying consumption of a food product held in a food package by using a mobile device can include capturing image data depicting a section of the food package by using a camera of the mobile device. The section includes a printed code and an opening device. The method can further include identifying a first sub-set of the image data depicting the printed code, determining identification data based on the first sub-set of the image data, identifying a second sub-set of the image data depicting the opening device, and determining a state of the opening device based on the second sub-set of the image data. The state is selected from a group consisting of an opened state or a closed state.

Abstract: The present disclosure provides new and innovative devices, methods, and apparatuses for discerning the presence and characteristics of objects that are passing in front of an indicia decoding device. In an example, a device comprises a 2D imaging assembly, a depth imaging assembly, and a processing device configured to analyze data from the depth imaging assembly to determine a number of items that are in a field of view of the depth imaging assembly during a period of time, analyze data from the 2D imaging assembly to determine a number of indicia that are present in a field of view of the 2D imaging assembly during at least a portion of the period of time, compare the number of indicia with the number of items, output a determination as to whether the number of indicia matches the number of items, and responsive to determining that the number of indicia does not match the number of items, transmit a message to an alert module.

Abstract: Systems and methods for decoding multiple indicia are provided. An example includes an indicia reading device comprising an imaging assembly configured to capture 2D image data of an environment appearing in a first FOV, the first FOV directed at a product scanning region; a depth sensor assembly including at least one depth sensor, wherein the at least one depth senor is configured to capture 3D image data of an environment appearing in a second FOV, the second FOV at least partially overlapping the first FOV; a controller, having at least a processor and a memory, communicably coupled to the imaging assembly and depth sensor assembly.

Abstract: Image data may indicate an average number of drops of printer fluid per pixel. Print data may be generated such that a total number of drops of printing fluid are deposited to a print medium during a pass in a forward direction and a pass in the reverse direction according to a distribution.

Abstract: A system and method for capturing media and for printing the captured media onto an article of merchandise. The system includes a mobile application that overlays one or more print dielines corresponding to the printable areas of one or more types of merchandise onto an associated camera display. In this way, the system provides, prior to the capturing of the photograph, a visual indication of the subject matter within the photograph (e.g., within the overlaid print dielines) that will be cropped and printed onto the merchandise. As such, a user may properly arrange the subject matter of the photograph (e.g., people) within the print dieline prior to capturing the photograph. The system subsequently communicates the portion of the photograph within the print dieline to a printing system to print the portion onto the merchandise. The system derives the print dielines from graphic files that correspond to the merchandises' printable areas.

Abstract: A printing apparatus includes a main controlling circuit, a sub-controlling circuit group including first-n-th sub-controlling circuits connected in series, a head group configured to eject liquid, a first communication path configured to connect the main controlling circuit and the first sub-controlling circuit, and a second communication path configured to connect the main controlling circuit and the n-th sub-controlling circuit. When the main controlling circuit transmit first print data to the first sub-controlling circuit, the sub-controlling circuit group transfers the first print data in a forward direction. When the main controlling circuit transmit second print data to the n-th sub-controlling circuit, the sub-controlling circuit group transfers the second print data in a reverse direction.

Abstract: An image recording apparatus includes a driving control circuit configured to record an image on a recording medium, and a timer configured to generate a first period to control the driving control circuit, wherein the timer is configured to generate a second period that is different from the first period, and wherein, in a power-saving mode that uses less power than a normal mode in which an image is recorded, the timer is configured to measure time based on the second period.

Abstract: An image-forming system includes: an image-forming apparatus configured to form an image on a recording medium using a consumable material in a container device; an inhibiting mechanism capable of inhibiting supply of the consumable material from the container device to the image-forming apparatus or insertion of the container device into the image-forming apparatus; an obtainment unit configured to obtain a code unique to the container device in a state where the inhibiting mechanism inhibits the supply of the consumable material or the insertion of the container device; a server apparatus configured to perform authentication for the container device based on the code; and a control unit configured to set, when the authentication is successful, setting data related to the container device received from the server apparatus in the image-forming apparatus.

Abstract: System to generate a secure code for authenticating a product is provided. The system includes a processing subsystem including a code generation module to generate a code based on product details. The processing subsystem includes a watermarking module to watermark points in the code. The processing subsystem includes a visible pattern generation module to generate a visible pattern. The processing subsystem includes a concealed pattern generation module to store a concealed pattern including second control points. The processing subsystem includes an overlay module to provide the secure code. The processing subsystem includes a scanning module to analyze the secure code to obtain the first points and discrete points. The scanning module is to interconnect the first points and the discrete points to obtain a reconstructed pattern including third control points. The scanning module is to compare the third control points with the second control points to authenticate the secure code.