Abstract: Systems and methods are disclosed for detecting hallucinations in large language models (LLMs). An example method includes receiving a first prompt for submission to the first LLM, generating, using the first LLM, a plurality of semantically equivalent prompts to the first prompt, generating, using the first LLM, a first response to the first prompt and a plurality of second responses to the plurality of semantically equivalent prompts, generating, using a second LLM, a plurality of third responses to the semantically equivalent prompts, generating a semantic consistency score for the first response based at least in part on the first prompt, the plurality of semantically equivalent prompts, the plurality of second responses, and the plurality of third responses, and determining whether or not the first response is an accurate response to the first prompt based at least in part on the semantic consistency score.

Abstract: A method may include obtaining first audio data of a first communication session between a first and second device and during the first communication session, obtaining a first text string that is a transcription of the first audio data and training a model of an automatic speech recognition system using the first text string and the first audio data. The method may further include in response to completion of the training, deleting the first audio data and the first text string and after deleting the first audio data and the first text string, obtaining second audio data of a second communication session between a third and fourth device and during the second communication session obtaining a second text string that is a transcription of the second audio data and further training the model of the automatic speech recognition system using the second text string and the second audio data.

Abstract: Systems, methods, computer systems, and computer-readable code for implementing methods for client relation management in which fewer call-in numbers need to be maintained. A method for client relation management preserves a more-complete record of online and offline interactions of a customer with systems provided by a plurality of service providers. A session identifier (SID) is associated with a customer upon interaction of the customer with a website provided by a first service provider. Website interaction information is received from the first service provider and is associated with the SID. Information associated with an incoming call from the customer at a call-in number at a call center maintained by a second service provider is used to associate the incoming call with the SID, creating a record of an offline interaction between the customer and an agent of the call center. The record of the offline interaction is associated with the SID.

Abstract: Disclosed herein are system, method, and computer program product embodiments for machine learning systems to process incoming call-center calls to provide communication summaries that capture effort levels of statements made during interactive communications. For a given call, the system receives a transcript as the input and generates a textual summary as the output. In order to improve a call summary and customize a summarization task to a call center domain, the technology disclosed herein may employ a classifier that predicts an effort level and attention score for individual utterances within a call transcript, ranks the attention scores and uses selected ones of the ranked utterances in the summary.

Abstract: A hearing aid may include a neural network chip having tiles arranged in an array, each tile including memory, 16-128 multiplier-accumulator circuits (MACs), and routing circuitry. The memory of each tile may be configured to store a portion of elements of a matrix A comprising weights of a recurrent neural network. Each tile may be configured to receive and store elements of an activation vector X, and all tiles in a column of the array may be configured to receive the same elements of X. The plurality of tiles may be configured to perform a matrix-vector multiplication A*X by performing multiply-and-accumulate sub-operations in parallel among the plurality of tiles. The routing circuitry from the tiles in each respective row of tiles may be configured to combine results of the multiply-and-accumulate sub-operations.

Abstract: A tensile tester includes a tester main body that executes a material test, an association storage portion that stores a specific word and instruction information to the tester main body in association with each other, a voice receiving portion that receives a voice from a user, an extraction portion that extracts the specific word from the voice by voice recognition processing, a determination portion that determines, with reference to the association storage portion, the instruction information corresponding to the specific word extracted by the extraction portion, and an instruction portion that outputs, to the tester main body, the instruction information determined by the determination portion.

Abstract: Apparatuses, systems, and techniques to transfer grammar between sentences. In at least one embodiment, one or more first sentences are translated into one or more second sentences having different grammar using one or more neural networks.

Abstract: A disclosed method includes receiving at least an indication of audio data generated based on sound captured by a microphone of a monitoring device. A set of audio event labels are generated by processing the received audio data with a model that includes audio recognition patterns. Each audio event label is generated by matching an audio pattern in the received audio data with an audio pattern of the audio recognition patterns. The method further includes identifying a hazard event type occurring within a time window of the time period by processing the set of audio event labels with a hazard detection model. In response to detecting the hazard event type, a hazard alert is generated and at least an indication of the hazard alert is communicated to an electronic device other than the monitoring device, which is authorized to receive communications from the monitoring device.

Abstract: A system for transportation includes a vehicle interface for gathering physiological sensed data of a rider in the vehicle. The system includes an artificial intelligence-based circuit that is trained on a set of outcomes related to rider in-vehicle experience and that induces, responsive to the sensed rider physiological data, variation in one or more of the user experience parameters to achieve at least one desired outcome in the set of outcomes. The inducing variation includes control of timing and extent of the variation.

Abstract: The present disclosure provides a method of training a classification model, which relates to an active learning, neural network and natural language processing technology. A specific implementation scheme includes: selecting, from an original sample set, a plurality of original samples with a class prediction result meeting a preset condition as to-be-labeled samples according to a class prediction result for a plurality of original samples in the original sample set; labeling the to-be-labeled sample as belonging to a class by using the second classification model, so as to obtain a first labeled sample set; and training the first classification model by using the first labeled sample set. The present disclosure further provides a method of classifying a sample, an electronic device, and a storage medium.

Abstract: An information-processing apparatus comprises: an identification unit configured to identify at least one of important matters, depending on inquiry contents from a first speaker to a second speaker in a conversation among multiple persons, the at least one of the important matters being matters that the second speaker is to explain to or confirm with the first speaker; and a presentation unit configured to present to the second speaker the at least one of the important matters that has been identified.

Abstract: A computer-implemented method, system and computer program product for identifying a semantic representation for a set of texts. The set of texts is encoded using a language model to obtain a set of corresponding texts' contextualized embeddings. A centroid of the contextualized embeddings is then calculated. A user-designated number of words from a pre-defined vocabulary of the language model that have their non-contextualized embeddings closest to the centroid are then identified. Furthermore, permutations of the identified words are calculated using an n-gram range. The permutations are then encoded using the language model in a contextualized manner. The encoded permutation from the encoded permutations with the greatest similarity to the centroid is identified. The identified encoded permutation is then assigned as the semantic representation of the set of texts.

Abstract: A framework for segmenting a medical text report into sections is disclosed. For each sentence of the report, a first sentence representation is determined by inputting a word-level context representation for each sentence sequentially into a neural network. A second sentence representation is determined by inputting an aggregated representation for each sentence sequentially into another neural network. For each sentence, a third sentence representation is determined based on a combination of the first and second sentence representations, and a section classification for the sentence is determined by inputting the third sentence representation into a section classifier. Each sentence is assigned the section classification determined for the sentence.

Abstract: An exemplary spatial enhancement system performs frequency-specific localization and speech comprehension enhancement. Specifically, the system receives an audio signal presented to a recipient of a hearing device, and generates, based on the audio signal, a first frequency signal and a second frequency signal. The first frequency signal includes a portion of the audio signal associated with a first frequency range, and the second frequency signal includes a portion of the audio signal associated with a second frequency range. Based on the first and second frequency signals, the system generates an output frequency signal that is associated with the first and second frequency ranges and that is configured for use by the hearing device in stimulating aural perception by the recipient. This generating of the output frequency signal includes processing the first frequency signal to apply a localization enhancement and processing the second frequency signal to apply a speech comprehension enhancement.

Abstract: A computing system identifies an incoming voice call from a user device to an agent device associated with the computing system. The computing system generates a transcription of the incoming voice call using one or more natural language processing techniques. The computing system extracts a problem description from the transcription. The problem description indicates a topic for the incoming voice call. A first machine learning model estimates a situation vector from the problem description. A second machine learning model identifies a pre-existing situation vector that closely matches the estimated situation vector. The computing system retrieves a situation description that corresponds to the identified pre-existing situation vector.

Abstract: There is provided a method and a system for generating an interactive search interface in response to a search request by using at least one machine learning (ML) model. A search request such as one of a word, a sentence, a paragraph, and a document is received, and a semantic representation of the search request is received. Semantically similar documents are received by: comparing the search request semantic representation with document representations to obtain semantic similarity scores, and selecting semantically similar documents based on the scores. For each of the set of semantically similar documents, a respective set of similar paragraphs, sentences, and words are determined based on associated representations. An interactive search interface is generated and displayed to a user interface. A selection of a given document is received, and each of the respective set of similar paragraphs, sentences, and similar words associated with the given document are displayed.

Abstract: A voice conference apparatus that provides a plurality of voice conferences via a network, the voice conference apparatus including: a voice conference section that transmits and receives a sound generated in each of the plurality of voice conferences to and from a plurality of user terminals used by a plurality of users participating in the voice conference; a sound analyzing section that analyzes the sound generated in each of the plurality of voice conferences; and a display control section that causes an administrator terminal used by an administrator administering the plurality of voice conferences to display a result of the analysis, by the sound analyzing section, of the sound generated in each of the plurality of voice conferences, in association with the voice conference.

Abstract: An earpiece whose inner cylinder is not crushed when the earpiece is inserted into an ear canal of a user is provided. An earpiece according to the present invention includes an inner cylinder 1 that includes one open end and the other open end and is attached to a sound conduit C of an earphone E, and an outer cylinder 2 connected to the one end of the inner cylinder and extending in an umbrella shape toward the other end of the inner cylinder so as to cover an outer surface 1b of the inner cylinder. The inner cylinder includes a fitting part 11 into which a tip of the sound conduit is fitted, and a sound guide part 12 configured to guide sound wave from the sound conduit to an ear canal.

Abstract: A transportation system and method for managing rewards in the transportation system includes using a merchant interface to a cognitive system for managing the offering or fulfillment of a reward to a rider of a vehicle, where a merchant may specify parameters of the reward that can be earned by the rider as a result of performing an action while in the vehicle.

Abstract: A system for operating a vehicle based on a state of a rider includes an artificial intelligence system, a vehicle control system, and a feedback loop. The artificial intelligence system processes a sensory input from a wearable device in a vehicle to determine a state of a rider and optimizes an operating parameter of the vehicle to improve the state of the rider. The artificial intelligence system includes a neural net with a perceptron to mimic human senses to facilitate determining a state of a rider based on an extent to which at least one of the senses of the rider is stimulated. The vehicle control system adjusts vehicle operating parameters and the feedback loop indicates the change in the state of the rider, where the vehicle control system adjusts at least one of the plurality of vehicle operating parameters responsive to the indication of the change.