Abstract: A syntax-directed mathematical expression recognition system combines online handwritten mathematical expression, offline handwritten and printed mathematical expression and structured mathematical expressions to generate a huge amount of mathematical expressions which are usable in training of neural networks. An end to end learning way is used to train the neural network for increment of learning effect and recognition ability. The problem of overfitting in multi-layer neural networks is reduced effectively so as to increase accuracy of testing set. The mode of neural network is usable for online handwritten mathematical expression, offline handwritten and printed mathematical expression. Therefore only one recognition system is used, no other system is needed.

Abstract: This disclosure relates to a device, a method, and a program capable of removing erroneous data from learning data used for machine learning used in natural language processing, for example. The method includes storing a forward direction learned model of a discrete series converter. The model is trained based on a plurality of pairs of discrete series of texts. Each pair comprises a first discrete series indicates an input of discrete series. A second discrete series indicates an output of discrete series. The first discrete series and the second discrete series are correctly associated. The method further includes converting the first discrete series to the second discrete series, and generating a quality score using the forward direction learned model, using a second learning pair of discrete series texts including an error in relationship.

Abstract: Systems and methods provides for dialog management in real-time rather than turn taking. An example method included generating first candidate responses to triggering event. The triggering event may be receipt of a live stream chunk for the dialog or receipt of a backend response to a previous backend request for a dialog shema. The method also includes updating a list of candidate responses that are accepted or pending with at least on of the first candidate responses, and determining, for the triggering event, whether the list of candidate responses includes a candidate response that has a confidence score that meets a triggering threshold. The method also includes waiting for a next triggering event without providing a candidate response when the list does not include a candidate response that has a confidence score that meets the triggering threshold.

Abstract: A method including detecting, in a written electronic communication, an input sentence satisfying a readability metric threshold. The method also includes transforming, by a sentence transformer model, the input sentence to output suggested sentences. The method also includes evaluating the suggested sentences along a set of acceptability criteria. The method also includes determining, based on the evaluating, that the set of acceptability criteria is satisfied. The method also includes modifying, based on determining that the set of acceptability criteria is satisfied, the written electronic communication with the suggested sentences to obtain a modified written electronic communication. The method also includes returning the modified written electronic communication.

Abstract: A communication broker service of a microservices platform may direct a distributed queue management service to monitor multiple data distribution outboxes of a plurality of microservices of the microservices platform for message data. The communication broker service may receive an indication from the distributed queue management service that the distributed queue management service has retrieved message data from a data distribution outbox of a first microservice. The communication broker service may apply a business logic to identify one or more second microservices that are configured to receive the message data from the first microservice. The communication broker server may then direct the distributed queue management service to distribute the message data of the first microservice from a data store of the distributed queue management service to one or more corresponding data distribution inboxes of the one or more second microservices.

Abstract: A method and system is disclosed for speech synthesis of streaming text. At a text-to-speech (“ITS) system, a real-time streaming text string having a starting point and an ending point may be received, and a first sub-string comprising a first portion of the text string received from an initial point to a first trigger point may be accumulated. The initial point is no earlier than the starting point and is prior to the first trigger point, and the first trigger point is no further than the ending point. A punctuation model of the ITS system may be applied to the first sub-string to generate a pre-processed first sub-string comprising the first sub-string with added grammatical punctuation as determined by the punctuation model. TTS synthesis processing may be applied to at least the pre-processed first sub-string to generate first synthesized speech, and audio play out of the first synthesized speech produced.

Abstract: An embodiment for a method of supporting neurodivergent comprehension of content by proactively identifying and transforming predicted difficult comprehension areas for a given user. The embodiment may receive an activation command from a registered user. The embodiment may automatically identify, within displayed content, predicted difficult comprehension areas based on a series of preferences and behaviors associated with the registered user. The embodiment may mark a textual element associated with the predicted difficult comprehension areas that may be transformed to improve comprehensibility of the predicted difficult comprehension areas for the registered user. The embodiment may detect user interaction with the marked textual element. The embodiment may, in response to detecting user interaction with the marked textual element, transform the marked textual element and provide transformed content to the registered user.

Abstract: A method and system for identifying enhancement suggestions for a document and displaying the identified suggestions in a user-friendly manner may include examining the document to identify an enhancement fora portion of content in the document, enabling display of a first user interface element for previewing the identified enhancement as applied to the content, receiving a request to select the identified enhancement, and upon receiving the request enabling display of a second user interface element for accepting the identified enhancement. The first user interface element may provide a preview of the document post enhancement.

Abstract: A translation method, a model training method, apparatuses, electronic devices and storage mediums, which relate to the field of artificial intelligence technologies, such as machine learning technologies, information processing technologies, are disclosed. In an implementation, a weight for each translation model in at least two pre-trained translation models translating a to-be-translated specified sentence is acquired based on the specified sentence and a pre-trained weighting model; and the specified sentence is translating using the at least two translation models based on the weight for each translation model translating the specified sentence.

Abstract: The technology provides a model-based approach for multilingual text rewriting that is applicable across many languages and across different styles including formality levels or other textual attributes. The model is configured to manipulate both language and textual attributes jointly. This approach supports zero-shot formality-sensitive translation, with no labeled data in the target language. An encoder-decoder architectural approach with attribute extraction is used to train rewriter models that can thus be used in “universal” textual rewriting across many different languages. A cross-lingual learning signal can be incorporated into the training approach. Certain training processes do not employ any exemplars. This approach enables not just straight translation, but also the ability to create new sentences with different attributes.

Abstract: A method includes detecting an incomplete entigen group within a knowledge database. The incomplete entigen group includes entigens and one or more entigen relationships between at least some of the entigens. The incomplete entigen group represents knowledge of a topic. The method further includes obtaining additive content for the topic based on the incomplete entigen group and generating an additive entigen group based on the additive content. The method further includes updating the incomplete entigen group utilizing the additive entigen group to produce an updated entigen group. The method further includes indicating that the updated entigen group has an un-curated status when the additive entigen group conflicts with the incomplete entigen group.

Abstract: Learning frameworks for processing text transcripts may include receiving, by an application, a query comprising a topic. A topic model may determine a plurality of subtopics based on the topic. The application may receive, from a database based on the topic and plurality of subtopics, a plurality of text transcripts. A sentiment model may compute, for each text transcript, a respective sentiment score based on a text of the respective text transcript. The application may determine, for each text transcript, a duration of a communication session associated with the respective text transcript. The application may compute, for each text transcript, a total score based on the sentiment score and the duration of the respective text transcript. The application may return, as responsive to the query, a subset of the plurality of text transcripts having a total score that exceeds a threshold.

Abstract: Aspects of the present disclosure involve a system comprising a computer-readable storage medium storing a program and method for replacing text in a message thread. The program and method provide for receiving an indication of user input provided by a first user, the user input comprising a text-based term for updating a message thread between the first user and a second user; determining that the text-based term is included within a set of predefined terms, each predefined term within the set of predefined terms having corresponding replacement content for including in the message thread, the replacement content having been specified by the first or second user with respect to messages exchanged between the first and second user; and providing, in response to the determining and based on the set of predefined terms, the replacement content corresponding to the text-based term within the message thread.

Abstract: Embodiments provide for improvements in generating focused data from an unstructured data set. The focused data generated from the unstructured data set may provide data insight(s) into analysis of the unstructured data set, and/or provide for improved capabilities for a user to efficiently navigate through relevant data portions of such data utilizing a user interface, even when such relevant data portions are not immediately distinguishable without further processing of the unstructured data set.

Abstract: Methods, apparatus and systems for recognizing sign language movements using multiple input and output modalities. One example method includes capturing a movement associated with the sign language using a set of visual sensing devices, the set of visual sensing devices comprising multiple apertures oriented with respect to the subject to receive optical signals corresponding to the movement from multiple angles, generating digital information corresponding to the movement based on the optical signals from the multiple angles, collecting depth information corresponding to the movement in one or more planes perpendicular to an image plane captured by the set of visual sensing devices, producing a reduced set of digital information by removing at least some of the digital information based on the depth information, generating a composite digital representation by aligning at least a portion of the reduced set of digital information, and recognizing the movement based on the composite digital representation.

Abstract: According to various embodiments, provided is a control operation comprising: receiving a first utterance from a device; determining a specific category corresponding to the received first utterance from among a plurality of categories; determining a first domain for processing of the first utterance from among a plurality of domains related to the determined specific category; acquiring information enabling an operation corresponding to the first utterance to be performed based on the determined first domain; receiving a second utterance from the device; identifying whether processing of the second utterance is possible in the determined first domain; based on identifying that the processing of the second user utterance based on the first domain is not possible, checking a second domain among the plurality of domains, and identifying whether processing of the second utterance based on the second domain is possible; and based on identifying that the processing of the second utterance based on the second domai

Abstract: Embodiments described herein provide a cross-lingual sentence alignment framework that is trained only on rich-resource language pairs. To obtain an accurate aligner, a pretrained multi-lingual language model is used, and a classifier is trained on parallel data from rich-resource language pairs. This trained classifier may then be used for cross-lingual transfer with low-resource languages.

Abstract: Systems, methods, and computer-readable media are disclosed for annotating content data such as video data with annotation data (e.g., images, emoji, memes, stylized text, sounds) in near real time. Example methods may include determining transcribed text from the content data, associated annotation data with the transcribed text, and annotating the content data with some or all transcribed text and annotation data. Example methods may further include editing the annotated content data to generate modified annotated content data and sending the annotated content data and/or modified annotated content data to a device.

Abstract: Embodiments of the present invention relate to a Lookup and Decision Engine (LDE) for generating lookup keys for input tokens and modifying the input tokens based on contents of lookup results. The input tokens are parsed from network packet headers by a Parser, and the tokens are then modified by the LDE. The modified tokens guide how corresponding network packets will be modified or forwarded by other components in a software-defined networking (SDN) system. The design of the LDE is highly flexible and protocol independent. Conditions and rules for generating lookup keys and for modifying tokens are fully programmable such that the LDE can perform a wide variety of reconfigurable network features and protocols in the SDN system.

Abstract: A non-transitory computer-readable recording medium stores a program for causing a computer to execute a process, the process includes acquiring training data that includes a first sentence expressed in a first language and a second sentence expressed in a second language, identifying a named entity and parts of speech from the first sentence, and generating, based on the training data, a translation model that includes an attention mechanism for the named entity and the parts of speech.

Abstract: In general, embodiments relate to a method for proactively detecting and filtering vulnerabilities of an application upgrade, comprising: receiving an application upgrade request to upgrade an application to a version from a client device; sending information related to the application upgrade to a vulnerability validator; determining, based on the impact score information, that the version of the application has vulnerabilities and that a second version of the application does not have vulnerabilities; filtering, based on the determining, the version of the application that has vulnerabilities; generating an application upgrade strategy by only considering the second version of the application; and sending information related to the version of the application to a vendor to fix the vulnerabilities.

Abstract: A system and method for determining the locations and types of objects in a plurality of videos. The method comprises pairing each video with one or more sentences describing the activity or activities in which those objects participate in the associated video, wherein no use is made of a pretrained object detector. The object locations are specified as rectangles, the object types are specified as nouns, and sentences describe the relative positions and motions of the objects in the videos referred to by the nouns in the sentences. The relative positions and motions of the objects in the video are described by a conjunction of predicates constructed to represent the activity described by the sentences associated with the videos.

Abstract: One embodiment provides a method, comprising: receiving an input sentence for a classification by a machine-learning model, where the classification is based upon a sentiment of the input sentence; splitting the input sentence into a plurality of tokens, each of the plurality of tokens corresponding to a term within the input sentence; creating a causal subgraph from the plurality of tokens, wherein the creating is based upon a causal relationship identified between tokens of the plurality of tokens; identifying, using the causal subgraph, tokens of the plurality of tokens influencing the classification; and generating, based upon the tokens of the plurality of tokens, a causal explanation for the classification, wherein the causal explanation identifies at least one portion of the input sentence resulting in the classification.

Abstract: Systems and methods for screenless computerized social-media access may include (1) producing, via an audio speaker that is communicatively coupled to a computing device, a computer-generated verbal description of a social-media post provided via a social-media application, (2) detecting, via a microphone that is communicatively coupled to the computing device, an audible response to the social-media post from a user of the computing device, and (3) digitally responding to the social-media post in accordance with the detected audible response. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: A device includes a memory adapted to store a list in a file or database comprising a plurality of vocabulary words in a first language and, for each vocabulary word, a corresponding word in a second language, a display device, and a processor. The processor is adapted to receive a plurality of words in the first language, select one or more words among the plurality of words, based on one or more predetermined criteria, translate, match or equate the one or more selected words from the first language to words of the second language, and cause the display device to display the plurality of words, wherein one or more first words that are in the plurality of words and are not among the one or more selected words which are displayed in the first language and one or more second words that are in the plurality of words and are among the one or more selected words are displayed in the second language.

Abstract: To improve accuracy in a document search for a document that includes a typographical error. A document retrieval device according to one embodiment of the present invention includes a misidentification table storing a correctly identified character string and a misidentified character string. The document retrieval device includes a document searcher. The document searcher obtains a search character string, and retrieves the search character string from both a document and a character string that is obtained by changing the misidentified character string included in the document to the correctly identified character string.

Abstract: A method may including obtaining a voice transcript corpus and a chat transcript corpus, extracting voice transcript sentences from the voice transcript corpus and chat transcript sentences from the chat transcript corpus, encoding, by a series of neural network layers, the voice transcript sentences to generate voice sentence vectors, encoding, by the series of neural network layers, the chat transcript sentences to generate chat sentence vectors, determining, for each voice sentence vector, a matching chat sentence vector to obtain matching voice-chat vector pairs, and adding, to a parallel corpus, matching voice-chat sentence pairs using the matching voice-chat vector pairs. Each of the matching voice-chat sentence pairs may include a voice transcript sentence and a matching chat transcript sentence. The method may further include training a disfluency remover model using the parallel corpus.

Abstract: Techniques are disclosed for training and/or utilizing an alignments and language model (“ALM”) in automatically determining an ALM score corresponding with natural language text generated using a natural language generation model. The natural language text generated using the natural language generation model can be based on a set of structured data. Additionally or alternatively, the ALM can include a fluency model portion and a semantics model portion. The fluency model portion can be used in determining the fluency and/or grammar of the text. The semantics model portion be used in evaluating the content of the natural language text with respect to the content of the structured data.

Abstract: A method may include retrieving, over a network device, data from a data source, the data formatted according to a relational database schema; converting, using at least one processor, the data from the data source into a knowledge graph; training, using the at least one processor, a neural network using triples from the knowledge graph, wherein a loss function of the neural network is based on a distance between a true triple from the knowledge graph and a corrupt triple; and after the training, storing weights of a hidden layer of the neural network as a vector space representation of entities in the knowledge graph.

Abstract: Embodiments of this application provide a method and a system for generating a note. The method includes detecting, by an electronic device, drawing a plurality of discontinuous underlines by a stylus on a touchscreen of the electronic device; combining, by the electronic device, a content selected by the discontinuous underlines to obtain a to-be-stored content; displaying, by the electronic device, a previous box comprising the to-be-stored content; controlling, by the electronic device, a target device to store the to-be-stored content, where the target device is at least one of the electronic device and the stylus; and outputting, by the target device, the to-be-stored content. In the embodiments of this application, a storage function of the electronic device and/or the stylus is used to output the to-be-stored content required by a user as a separate note, which facilitates convenience for the user to read key notes and improves user experience.

Abstract: A translation pen includes: a pen body, an indication component, an image collector and a first processor. The pen body has a pen tip end. The indication component is arranged on the pen tip end. The image collector is arranged on the pen body, and the image collector is configured to: collect an image including a text to be translated according to a position indicated by the indication component, and send the image collected. The first processor is arranged in the pen body and electrically connected to the image collector, and the first processor is configured to: receive the image sent by the image collector, and recognize the text to be translated in the image.

Abstract: The present invention discloses a system and for sentence completion in an edge device. The method comprises receiving input from a user through a user interface (202); scrutinizing a length of the received input by a smart sentence composer; sanitizing the received input text and restricting the length of the input text to be within a preset threshold number; splitting the received input text from into letters/words (208); predicting several words and a cluster of most probable word's indices by the smart sentence composer (210), and wherein the word's indices have a similarity above the preset threshold number; converting the word's indices to words using index to word mapping (212); choosing the word/words in the cluster having highest score for prediction (214); and the words are continuously predicted until the end of token of word index (216), is determined for completing the sentence prediction.

Abstract: An abstractive technique and an extractive technique are used to generate concise natural-language summaries of related text documents. The abstractive step generates a machine summary by constructing a graph with nodes representing unique pairs of tokens and corresponding parts-of-speech (POS), and with edge sequences representing token/POS pairs comprising sentences of a corresponding topic group from the text documents. Ranked candidate summary sentences are generated using subgraphs of the graph having initial and final nodes corresponding with valid sentence start and end pairs. The machine summary includes representative summary sentence(s) selected from each topic group's ranked candidates. The extractive step generates a natural-language summary from the machine summary by computing, for each topic group, numerical suitability measures providing comparisons between the representative summary sentence and sentences of the topic group.

Abstract: A system includes a memory and at least one processor to receive text from a client computing device, the text received one character at a time, as each character of the text is received, determine a recommendation in real-time to be added to the text based on at least one of a list of rules, word embedding, an n-gram model, and a co-occurrence model, the recommendation comprising at least one of a word, a list of hashtags, a quotation, and a list of emojis, and send the recommendation to the client computing device.

Abstract: Systems and methods are provided for displaying received publishing content on a web page along with one or more user elements by which one or more users may submit sentiment input or textual input in relation to the received publishing content or a subportion of the received publishing content; receiving a user input related to the displayed publishing content displayed on the web page, the user input including an identification of a subportion of the displayed publishing content and a sentiment input or a textual input; analyzing any user input received from each of one or more of the plurality of users in relation to the subportion; computing a sentiment score based on analysis of the analyzed user inputs received from each of one or more of the plurality of users in relation to the subportion; and displaying indicia representing the sentiment score computed for the subportion.

Abstract: A method for context-based natural language processing is disclosed herein. The method comprises maintaining a plurality of dialog system rules, receiving a user request from a Dialog System Interface, receiving one or more attributes associated with the user request from the Dialog System Interface or a user device, and identifying a type of context associated with the user request based on the user request and the one or more attributes. A context label is assigned to the user request associated with the type of context. Based on the context label and the user request, a particular dialog system rule is selected from the plurality of dialog system rules. A response to the user request is generated by applying the dialog system rule to at least a part of the user request.

Abstract: Content generation leverages an unsupervised, generative pre-trained language model (e.g., a generative-AI). In this approach, a model derived by applying to given content relevant competitive content and one or more optimization targets is received. Based on optimization criteria encoded as embedding signals in the model, a determination is made regarding whether a template suitable for use as an input to the generative-AI exists in a set of templates. If so, the model embedding signals are merged into the template, or the template itself is transformed using the embedding signals, in either case creating a modified template. If, however, no template suitable as the input exists, the model and other information are input to a natural language processor to generate a generative-AI input. Either the modified template or the generative-AI input, as the case may be, is then applied through the generative-AI to generate an output competitively-optimized with respect to the optimization targets.

Abstract: A method for example-driven machine learning is disclosed herein. The method comprises maintaining a plurality of dialog system rules and a knowledge database including a plurality of intent objects and a plurality of entity objects. The plurality of intent objects and the plurality of entity objects are associated with at least one dialog system rule. An exemplary phrase is received and one or more linguistic elements are retrieved from the exemplary phrase. It is determined that at least one of the linguistic elements is directed to at least one of the plurality of intent objects of the plurality of entity objects and the at least one of the linguistic elements in association with the at least one dialog system rule is added to the knowledge database.

Abstract: A communication broker service of a microservices platform may direct a distributed queue management service to monitor multiple data distribution outboxes of a plurality of microservices of the microservices platform for message data. The communication broker service may receive an indication from the distributed queue management service that the distributed queue management service has retrieved message data from a data distribution outbox of a first microservice. The communication broker service may apply a business logic to identify one or more second microservices that are configured to receive the message data from the first microservice. The communication broker server may then direct the distributed queue management service to distribute the message data of the first microservice from a data store of the distributed queue management service to one or more corresponding data distribution inboxes of the one or more second microservices.

Abstract: A rule integration device includes determination means for converting, to natural language sentences, each of multiple management rules used by a management device for managing a management target, and determining whether or not the multiple management rules are combinable based on grammar relating to the converted natural language sentences; and combination means for generating a post-combination rule by combining the multiple management rules that have been determined to be combinable by the determination means.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, that facilitate generating stable real-time textual translations in a target language of an input audio data stream that is recorded in a source language. An audio stream that is recorded in a first language is obtained. A partial transcription of the audio can be generated at each time interval in a plurality of successive time intervals. Each partial transcription can be translated into a second language that is different from the first language. Each translated partial transcription can be input to a model that determines whether a portion of an input translated partial transcription is stable. Based on the input translated partial transcription, the model identifies a portion of the translated partial transcription that is predicted to be stable. This stable portion of the translated partial transcription is provided for display on a user device.

Abstract: This application provides a method for training a semantic understanding model, including: obtaining a first training sample set; performing denoising processing on the first training sample set to form a corresponding second training sample set; processing the second training sample set by using a semantic understanding model, to determine initial parameters of the semantic understanding model; processing the second training sample set by using the semantic understanding model in response to the initial parameters of the semantic understanding model, to determine update parameters of the semantic understanding model; and iteratively updating a semantic representation layer network parameter and a task-related output layer network parameter of the semantic understanding model by using the second training sample set and according to the update parameters of the semantic understanding model.

Abstract: Apparatuses, systems, and methods described include receiving data related to an availability for a shift, automatically triggering initiation of a communications session related to the shift, conducting the communications session, and receiving and storing a plurality of audio or audio and visual signals from the communications session. Machine learning (ML) sentiment analysis is performed on data of the communications session and based on the sentiment analysis, a reliability score is determined.

Abstract: Disclosed herein are systems and method for providing a File System (FS) without redundancy for one or more services. In one aspect, an exemplary method comprises, mounting a base image of microservices to a directory, for each of the one or more services, union-mounting a service image on top of the base image, identifying all dependencies associated with the service image, and creating one or more sub-directories for each dependency associated with the service image, for each identified dependency, creating a link between the dependency and the union-mounted service image and base image, and creating, one or more micro-services.

Abstract: The present invention relates to a method for generating and causing display of a communication interface that facilitates the sharing of emotions through the creation of 3D avatars, and more particularly with the creation of such interfaces for displaying 3D avatars for use with mobile devices, cloud based systems and the like.

Abstract: The present disclosure provides a method and an apparatus for information extraction, an electronic device, and a storage medium. The method for information extraction includes: first obtaining text data, and then inputting the text data into an information extraction model obtained through pre-training to obtain triple information contained in the text data, wherein the triple information includes a subject, a predicate and an object in the text data. The information extraction model includes a binary classification sub-model and a multi-label classification sub-model, wherein the binary classification sub-model is configured to extract the subject in the text data, and the multi-label classification sub-model is configured to extract the predicate and the object corresponding to the subject in the text data according to the subject and the text data.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for disambiguating join paths for natural language queries. One of the methods includes, obtaining a natural language query from a user; parsing the natural language query into structured operations to be performed on APIs of a knowledge base, including: responsive to detecting a parsing ambiguity in which the natural language query can be parsed in two or more ways: providing, through a user interface, to the user one or more information items identifying the parsing ambiguity; responsive to a user interaction with an information item: modifying the parsing in accordance with the user interaction to generate one or more structured operations; performing the one or more structured operations on the structured APIs of the knowledge base to determine one or more search results; and providing one or more search results to the user.

Abstract: A machine translation system, a ChatOps system, a method for a context-aware language machine identification, and computer program product. One embodiment of the machine translation system may include a density calculator. The density calculator may be adapted to calculate a part of speech (POS) density for a plurality of word tokens in an input text, calculate a knowledge density for the plurality of word tokens, and calculate an information density for the plurality of word tokens using the POS density and the knowledge density. In some embodiments, the machine translation system may further comprise a sememe attacher and a context translator.

Abstract: Computer-readable media, systems and methods may improve classification of content based on a machine-learning natural language processing (ML-NLP) classifier. The system may train a general language model based on a general corpus, further train the general language model based on a domain-specific corpus to generate a domain-specific language model, and conduct supervised machine-learning based on the domain-specific language using topic-specific corpus labeled as relating to topics of interest to generate the ML-NLP classifier. Accordingly, the ML-NLP classifier may be trained on a general corpus, further trained on a domain-specific corpus, and fine-tuned on a topic-specific corpus. In this manner, domain-specific content may be classified into topics of interest. The ML-NLP classifier may classify content into the topics of interest.

Abstract: A method includes detecting, in a written electronic communication, an input sentence satisfying a readability metric threshold, and processing, by a sentence transformer model responsive to the input sentence satisfying the readability metric threshold, the input sentence to output a suggested set of sentences. The method further includes evaluating the first suggested set of sentences along a set of acceptability criteria, and determining, based on the evaluating, that the set of acceptability criteria is satisfied. The method further includes modifying, based on determining that the set of acceptability criteria is satisfied, the written electronic communication with the suggested set of sentences to obtain a modified written electronic communication, and storing the modified written electronic communication.