Abstract: Disclosed in some examples are methods, systems, and machine-readable mediums for utilizing context information to create decoding feedback information to improve decoder accuracy and/or performance. In some examples, the context information is from layers of a network stack above the layers in which the decoders are present. The context information may be or be based upon information about previously received and decoded data and/or information about the sender to provide decoding feedback information to the decoder that is used either to correct a previous decoding error or to inform the decoder on which of a plurality of decoding choices is more likely to be correct. This may increase decoding performance by decreasing errors and in some examples, reducing the complexity of choices by eliminating certain decoding possibilities and thus increasing decoder efficiency.

Abstract: A system and methods for latent contextual threading for personalized dialogue through geometric manifold-based conversation management. The system maintains a personalized cognitive manifold as a geometric manifold in latent space that encodes user-specific dialogue patterns as navigable geometric structures. Multiple dialogue contexts are maintained as geometric trajectories within the manifold, with dialogue responses generated through manifold traversal rather than discrete context retrieval. A bidirectional adaptation system modifies the manifold's geometric structure based on user interactions. The system preserves dialogue continuity across session boundaries by serializing manifold geometry during session termination and restoring geometric positioning during session resumption. Dialogue coherence is evaluated through geometric analysis including curvature calculations and geodesic deviation measurements.

Abstract: A method and a system for interacting with one or more computer resource assets at a location. The system includes a processor, a storage, and an interface suite including a first interface configured to communicate with a user device and a second interface configured to interact with at least one computing resource asset at the location.

Abstract: In one aspect, a method includes detecting a fingerprint match between query fingerprint data representing at least one audio segment within podcast content and reference fingerprint data representing known repetitive content within other podcast content, detecting a feature match between a set of audio features across multiple time-windows of the podcast content, and detecting a text match between at least one query text sentences from a transcript of the podcast content and reference text sentences, the reference text sentences comprising text sentences from the known repetitive content within the other podcast content. The method also includes responsive to the detections, generating sets of labels identifying potential repetitive content within the podcast content. The method also includes selecting, from the sets of labels, a consolidated set of labels identifying segments of repetitive content within the podcast content, and responsive to selecting the consolidated set of labels, performing an action.

Abstract: The present disclosure is directed toward systems, methods, and non-transitory computer-readable media for generating responses to multi-order text queries using a context orchestration engine. For example, the disclosed systems generate context-defining query subcomponents from a multi-order text query, where the context-defining query subcomponents indicate contextual data sources pertaining to their respective portions of the multi-order text query. In addition, the disclosed systems provide or transmit the context-defining query subcomponents to a large language model for domain-specific computer code pertaining to each respective context-defining query subcomponent. The disclosed systems can further execute the generated computer code for each context-defining query subcomponent to access indicated contextual data sources for generating component-specific results. The disclosed systems can also generate a multi-order result to the multi-order text query from the component-specific results.

Abstract: A computer-implemented method and computer system improvements for automatically generating a summary note related to a digitally-recorded interlocutor conversation, such as a chat transcript, including accessing a data corpus having at least one digitally-recorded conversation of text-based interlocutory conversation(s) including at least one labeled value, extracting at least one Conversation Feature from the labeled value(s), creating at least one Summary Feature from the extracted Conversation Feature(s), generating at least one Summary Note by combining one or more Narrative Structures with the Summary Feature(s), and digitally outputting the at least one Summary Note.

Abstract: A method of suppressing noise may include receiving a sequence of audio frames representing a multi-channel audio signal. The method may include determining a likelihood of speech in a first audio frame of the sequence of audio frames based on a Gaussian mixture model. Further, the method may include generating a first audio signal based on the likelihood of speech in the first audio frame and a second audio signal representing a first speech component of a second audio frame. The second audio frame follows the first audio frame in the sequence of audio frames. The method may also include determining, using a neural network model, a likelihood of speech in the second audio frame based on the first audio signal, and filtering a noise component of the second audio frame based at least in part on the likelihood of speech in the second audio frame.

Abstract: This disclosure provides methods, devices, and systems for audio signal processing. The present implementations more specifically relate to speech enhancement techniques that combine statistical signal processing with neural network inferencing. In some aspects, a speech enhancement system may include a linear filter, a deep neural network (DNN), and a nonlinear post-filter. The linear filter and the nonlinear post-filter are configured to suppress noise in audio signals using statistical signal processing techniques. More specifically, the linear filter denoises an input audio signal based on a temporal correlation between successive frames of the audio signal. The DNN infers a speech signal and a noise signal (representing a speech component and a noise component, respectively, of the audio signal) based on the denoised audio signal.

Abstract: The present disclosure is generally directed to a tangible, non-transitory machine-readable medium that includes machine-readable instructions that, when executed, cause processing circuitry to receive a first indication of multimedia content and a second indication of whether the multimedia content is to be transcribed. The instructions cause the processing circuitry to send content generated from the multimedia content for transcription. The content includes an identifier associated with the multimedia content. Additionally, the instructions cause the processing circuitry to send a request for the content to be transcribed. The request includes or is indicative of the identifier. Moreover, the instructions cause the processing circuitry to receive a transcript for at least a portion of the content and generate transcript metadata that includes timing data and is indicative of text of the transcript.

Abstract: The techniques disclosed herein manage computing environments associated with radio access networks using a natural language interface. This is achieved through utilizing natural language processing to analyze user generated inputs and generate robust large language model queries. In various examples, the queries can include radio access network documentation, diagnostic data, and past interactions to provide custom context to the large language model. Accordingly, the query can cause the large language model to generate an operation sequence comprising a plurality of commands to interface with a resource management tool and control computing resources and supporting components. In this way, the present techniques can alleviate the technical burden on end users and minimize the risk of errors.

Abstract: Systems and methods include acquisition of operation data associated with a plurality of operations, generate a set of feature values for each operation based on associated operation data, determine a first set of the operations associated with an error and, for each operation of the first set, automatically determine a rejection reason based on operation data associated with the operation, determine a rejection label associated with the rejection reason, and associate the rejection label with the set of feature values generated for the operation.

Abstract: A method, computer program product, and computer system for generation of training examples for training an automatic speech recognizer. Embodiments of the present invention can receive a training dataset of original audio signals and generate training examples for training an automatic speech recognizer based, at least in part, on a constructed imperceptible space for an original audio signal of the original audio signals and adversarial audio examples in the constructed imperceptible space. Embodiments of the present invention can then generate an imperceptible and adversarial audio example to an adversarial trainer for the automatic speech recognizer.

Abstract: A management system (1090) includes an air conditioner (1062), a sound input unit (1064), an operation sound analyzing unit (1016), a first preprocessing unit (1011), a command extracting unit (1014), and a control unit (1017). The sound input unit (1064) is provided in a room where the air conditioner is provided and obtains sound data (SD). The operation sound analyzing unit (1016) analyzes the sound data by a predetermined analysis method so as to obtain a determination result (RE) regarding whether an abnormality exists in the air conditioner or details of the abnormality. The first preprocessing unit (1011) performs first preprocessing on the sound data. The command extracting unit (1014) extracts a command (CM) of a user from the sound data that has been subjected to the first preprocessing. The control unit (1017) controls the air conditioner on the basis of the command.

Abstract: A knowledge accelerator platform automatically builds and maintains a knowledge index of assets stored across a plurality of sources to allow consolidated searching of assets in the plurality of sources. The platform automatically processes the assets to extract metadata information and record labels for each asset in the index, without storing the content of an asset itself in the index, where a semantic dictionary maps labels to terminology in different domains. The knowledge accelerator platform further provides an internal user interface that allows internal users to evaluate the knowledge contained in the assets at a high level and thus identify knowledge gaps that can be mitigated, and promote quality assets within targeted user communities. The knowledge accelerator platform also provides an external user interface that recommends assets to users based on various factors such as their experience, status, and domain interests.

Abstract: Method, apparatus, and non-transitory storage medium for training a deep neural-network model jointly for acoustic echo suppression and acoustic howling suppression are provided. The method may include generating a teacher speech signal for training the deep neural-network model based on a input speech from a speech system and at least one reference signal. The deep neural-network model is trained jointly for both acoustic echo suppression and acoustic howling suppression by using the teacher speech signal and a correlation loss. During training of the deep neural-network model, the training task formulates a recurrent feedback suppression process as an instantaneous speech separation task using the teacher-forced training strategy.

Abstract: Techniques are disclosed herein for using named entity recognition to resolve entity expression while transforming natural language to a meaning representation language. In one aspect, a method includes accessing natural language text, predicting, by a first machine learning model, a class label for a token in the natural language text, predicting, by a second machine-learning model, operators for a meaning representation language and a value or value span for each attribute of the operators, in response to determining that the value or value span for a particular attribute matches the class label, converting a portion of the natural language text for the value or value span into a resolved format, and outputting syntax for the meaning representation language. The syntax comprises the operators with the portion of the natural language text for the value or value span in the resolved format.

Abstract: Described herein is a method of processing an audio signal using a neural network or using a first and a second neural network. Described is further a method of training said neural network or of jointly training a set of said first and said second neural network. Moreover, described is a method of obtaining and transmitting a latent feature space representation of a perceptual domain audio signal using a neural network and a method of obtaining an audio signal from a latent feature space representation of a perceptual domain audio signal using a neural network. Described are also respective apparatuses and computer program products.

Abstract: According to one implementation, a system includes a computing platform having processing hardware, and a system memory storing a software code. The processing hardware is configured to execute the software code to receive a vocabulary, identify words from the vocabulary for use in extending the vocabulary, pair each of those words with every other of those words to provide word pairs, and output the word pairs to a vocabulary administrator. The software code also receives word pair characterizations identifying each of the word pairs as one of similar, dissimilar, or neither similar nor dissimilar, configures, based on the word pair characterizations, a multi-dimensional vector space including multiple embedding vectors each corresponding respectively to one of the identified words, and cross-references each of those words with its corresponding embedding vector to produce an extended vocabulary corresponding to the received vocabulary.

Abstract: A microphone device, comprising: a microphone configured to generate an audio signal; a natural speech detection module for detecting natural speech in the first audio signal; wherein, on detection of the natural speech in the first audio signal, the natural speech detection module is configured to output a trigger signal to a speech processing module to process the natural speech in the audio signal.

Abstract: An audio watermark addition method carried out on a playback terminal is described. The playback terminal obtains first audio in real time and embeds an audio watermark within the first audio. The audio watermark is associated with the playback terminal, and the playback terminal plays the first audio including the embedded audio watermark.