Abstract: The disclosure provides a natural language processing (NLP) model arranged to operate on two lexicons, where one lexicon is a sub-set of the other lexicon. The NLP model can be arranged to generate output based on the sub-set lexicon and exit processing of the NLP model, to potentially save computation cycles.

Abstract: The present invention relates to a method and system for verification scoring and automated fact checking. More particularly, the present invention relates to a combination of automated and assisted fact checking techniques to provide a verification score. According to a first aspect, there is a method of verifying input data, comprising the steps of: receiving one or more items of input data; determining one or more pieces of information to be verified from the or each item of input data; determining which of the one or more pieces of information are to be verified automatically and which of the one or more pieces of information require manual verification; determining an automated score indicative of the accuracy of the at least one piece of information which is to be verified automatically; and generating a combined verification score which gives a measure of confidence of the accuracy of the information which forms the or each item of input data.

Abstract: The present disclosure relates to utilizing a deep recurrent neural network for accurately performing punctuation restoration. For example, the disclosed systems can provide a sequence of words to a punctuation restoration neural network having multiple bi-directional recurrent layers and one or more neural attention mechanisms. In one or more embodiments, the punctuation restoration neural network incorporates layer-wise attentions and/or multi-head attention. The disclosed systems can utilize the punctuation restoration neural network to generate probabilities for each word, indicating the likelihood that each possible punctuation mark is associated with that word. Based on these probabilities, the disclosed systems can generate a punctuated transcript that includes punctuation before the appropriate words.

Abstract: The present disclosure provides a sentence generation method as well as a sentence generation apparatus and a smart device. The method includes: obtaining an input sentence; searching for structurally similar sentence(s) of each input sentence, where the structurally similar sentence(s) are structurally similar to the input sentence; finding semantically similar sentence(s) of the structurally similar sentence(s); parsing the input sentence and the structurally similar sentence(s) to obtain a subject block, a predicate block, and an object block to rewrite the semantically similar sentences to generate a new sentence; filtering the new sentence based on a preset filtering condition; and labeling the filtered new sentence as a semantically similar sentence of the input sentence. In this manner, a plurality of new sentences with different sentence patterns can be generated based on the same input sentence, which improves the controllability in generating the sentences and saves the labor cost therein.

Abstract: Techniques for localization of a content presentation in an information processing system are provided. In one example, a method adapts translated content to accommodate an original content design structure (e.g., wireframe) when such translated content is inconsistent with the original content design structure. In another example, a method adapts the original content design structure when translated content is inconsistent with the original content design structure. In yet another example, an image analysis method is used to adapt translated content determined to be inconsistent with the original content design structure.

Abstract: Disclosed is a template-based automatic question and answer method for software bug. An entity relationship triple is extracted from a bug corpus and a natural language pattern is acquired; an entity relationship in the triple is determined; a query template corresponding to the natural language pattern is acquired; an entity in a question q proposed by a user is replaced with an entity type to acquire a question q?; then, the entity type in q? and an entity type in the natural language pattern are compared and searched for and a similarity is calculated; then, a SPARQL query pattern of the question q is acquired according to the similarity and the entity in the question q; and finally, the SPARQL query pattern of the question q is executed so as to acquire an answer to the question q.

Abstract: Embodiments of the present disclosure provide a method and apparatus for processing a speech. The method may include: acquiring an original speech; performing speech recognition on the original speech, to obtain an original text corresponding to the original speech; associating a speech segment in the original speech with a text segment in the original text; recognizing an abnormal segment in the original speech and/or the original text; and processing a text segment indicated by the abnormal segment in the original text and/or the speech segment indicated by the abnormal segment in the original speech, to generate a final speech. A speech segment in the original speech is associated with a text segment in the original text to realize visual processing of the speech.

Abstract: A deep learning based dialog method, apparatus, and device are provided and belong to the field of artificial intelligence. The method includes: obtaining a to-be-replied statement, encoding the to-be-replied statement to obtain a first vector , where the first vector is a representation of the to-be-replied statement; obtaining dialog history information corresponding to the to-be-replied statement, and the attention vector is used to represent search intent; making each dialog statement interact with the attention vector, so as to extract information related to the search intent from the dialog statement to obtain a plurality of result vectors, generating a to-be-decoded vector based on the plurality of result vectors, and decoding the to-be-decoded vector to obtain a next word in the reply statement. In the method, the reply statement refers to a dialog history.

Abstract: Techniques for localization of a content presentation in an information processing system are provided. In one example, a method adapts translated content to accommodate an original content design structure (e.g., wireframe) when such translated content is inconsistent with the original content design structure. In another example, a method adapts the original content design structure when translated content is inconsistent with the original content design structure. In yet another example, an image analysis method is used to adapt translated content determined to be inconsistent with the original content design structure.

Abstract: The invention discloses a device and text representation method applied to sentence embedding, which has determining a parent word and a child words set corresponding to the parent word, obtaining an hidden interaction state of parent word based on the hidden interaction state of all child words; obtaining parent word sequence corresponding to the parent word and obtaining hidden state sequence corresponding to the hidden state of parent word; obtaining the interaction representation sequence of each parent word and other parent word based on the hidden state sequence, generating sentence embeddings. The invention proposes to realize sentence embeddings through a two-level interaction representation.

Abstract: A combined translation request is obtained that includes content of a first language to be translated to a second language combined with translation context information of a user for which translation is to be performed. The content is translated, using machine translation, from the first language to the second language based on the translation context information to provide a translation. The translation is adjusted based on one or more identified translation factors selected based on user feedback to provide an adjusted translation. The adjusted translation is provided to the user.

Abstract: A method performed by a computing device includes identifying a first set of sentiment identigens of a plurality of sets of identigens for a query that includes a string of words, where the first set of sentiment identigens represents different meanings of a first sentiment word of the string of words. The method further includes selecting an entigen group from a knowledge database based on the plurality of sets of identigens, where a first sentiment entigen of the entigen group corresponds to the first set of sentiment identigens. The method further includes generating a response entigen group based on the entigen group, where a response entigen of the response entigen group corresponds to a selected identigen from a set of identigens of the plurality of sets of identigens regarding a word of the string of words.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating a target sequence from an input sequence. In one aspect, a method comprises maintaining a set of current hypotheses, wherein each current hypothesis comprises an input prefix and an output prefix. For each possible combination of input and output prefix length, the method extends any current hypothesis that could reach the possible combination to generate respective extended hypotheses for each such current hypothesis; determines a respective direct score for each extended hypothesis using a direct model; determines a first number of highest-scoring hypotheses according to the direct scores; rescores the first number of highest-scoring hypotheses using a noisy channel model to generate a reduced number of hypotheses; and adds the reduced number of hypotheses to the set of current hypotheses.

Abstract: Systems, methods, and computer-readable media for identifying application modules for accomplishing the predicted required functionality are disclosed. The systems and methods may involve outputting a logical sentence structure template for use in building a new application module, the logical sentence structure template including a plurality of definable variables; receiving at least one input for at least one of the definable variables; performing language processing on the logical sentence structure including the at least one received input to thereby characterize the function of the new application module; comparing the characterized function of the new application module with pre-stored information related to a plurality of predefined application modules to determine at least one similarity to a specific predefined application module; and based on the at least one similarity, presenting the specific predefined application module as an adoptable alternative for accomplishing the function.

Abstract: A method, system and a computer program product are provided for aligning embeddings of multiple languages and domains into a shared embedding space by transforming monolingual embeddings into a multilingual embeddings in a first shared embedding space using a cross-lingual learning process, and then transforming the multilingual embeddings into cross-domain, multilingual embeddings in a second shared embedding space using a cross-domain learning process.

Abstract: Techniques are disclosed for detecting distributed incompetence in text of a conversation using communicative discourse trees and then inserting an automatic response from an autonomous agent (chatbot) or other entity. For example, a computing system generates a communicative discourse tree from utterances from multiple agents to a user. The computing system obtains a prediction of whether the text includes distributed incompetence by applying a trained predictive model to the communicative discourse tree. Based on the detection, the computing system generates an updated response to a user device.

Abstract: A database may contain text strings in a preferred language and in one or more other languages. One or more processors may be configured to: generate a graphical user interface containing the text strings in the preferred language and in the other languages, and a control for dynamic translation, wherein a first set of the text strings in the other languages are displayed within text input controls, and wherein a second set of the text strings in the other languages are not displayed within the text input controls; receive an activation indication of the control for dynamic translation; and generate an update to the graphical user interface that includes translations of the first set into the preferred language appearing adjacent to the first set in the other languages, and also translations of the second set into the preferred language replacing the second set in the other languages.

Abstract: Techniques are disclosed for generating a virtual persuasive dialogue. In an example, a dialogue application receives a selection of a topic from a user device. The application identifies document results that are associated with the topic. Using communicative discourse trees, the application identifies document results that include argumentation, transforms these document results into a dialogue form, and presents the results to a user device as a virtual persuasive dialogue.

Abstract: Techniques for using noisy-robust discourse trees to determine a rhetorical relationship between sentences. In an example, a rhetoric classification application creates a noisy-robust communicative discourse tree. The application accesses accesses a first communicative discourse tree derived from a first sentence, a third sentence, and a fourth sentence and a second communicative discourse tree derived from a second sentence, the third sentence, and the fourth sentence. The application determines that syntactic parse trees cannot be generated for the first sentence and the second sentence. The application identifies a common rhetorical relationship between the first communicative discourse tree and the second communicative discourse tree. The application removes an elementary discourse unit that does not correspond to the common rhetorical relationship from the first communicative discourse tree and the second communicative discourse tree.

Abstract: A text-based egotism level detection system and a process for detecting egotism level in alpha-numeric textual information via artificial intelligence (AI), deep learning, and natural language processing (NLP) are disclosed. The text-based egotism level detection system and process for detecting egotism level in alpha-numeric textual information via AI, deep learning, and NLP detects egotism in text using a convolution neural network (CNN) for deep learning. The text-based egotism level detection system and process for detecting egotism level in alpha-numeric textual information via AI, deep learning CNN, and NLP builds, maintains, utilizes, and updates an egotism detection text language processing model that is generated from a huge amount of text data including sentences that are designated as egotistic sentences or not egotistic sentences based on a deep understanding of egotism and data science.