A computer may train a single-class machine learning using normal speech recordings. The machine learning model or any other model may estimate the normal range of parameters of a physical speech production model based on the normal speech recordings. For example, the computer may use a source-filter model of speech production, where voiced speech is represented by a pulse train and unvoiced speech by a random noise and a combination of the pulse train and the random noise is passed through an auto-regressive filter that emulates the human vocal tract. The computer leverages the fact that intentional modification of human voice introduces errors to source-filter model or any other physical model of speech production. The computer may identify anomalies in the physical model to generate a voice modification score for an audio signal. The voice modification score may indicate a degree of abnormality of human voice in the audio signal.

A device may be configured to parse a syntax element specifying the number of available languages within a presentation associated with an audio stream. A device may be configured to parse one or more syntax elements identifying each of the available languages and parse an accessibility syntax element for each language within the presentation.

Many headsets include automatic noise cancellation (ANC) which dramatically reduces perceived background noise and improves user listening experience. Unfortunately, the voice microphones in these devices often capture ambient noise that the headsets output during phone calls or other communication sessions to other users. In response, many headsets and communication devices provide manual muting circuitry, but users frequently forget to turn the muting on and/or off creating further problems as they communicate. To address this, the present inventors devised, among other things, an exemplary headset that detects the absence or presence of user speech, automatically muting and unmuting the voice microphone without user intervention. Some embodiments leverage relationships between feedback and feedforward signals in ANC circuitry to detect user speech, avoiding the addition of extra hardware to the headset. Other embodiments also leverage the speech detection function to activate and deactivate keyword detectors, and/or sidetone circuits, thus extending battery.

A home appliance includes an electrical equipment compartment disposed in an upper portion of the home appliance, and including an upper side that is open, an electrical equipment compartment cover to cover the open upper side of the electrical equipment compartment, and including a speaker hole, a microphone accommodating portion which protrudes upward from an upper side of the electrical equipment compartment cover and including an accommodating space and a front portion that includes microphone holes laterally spaced apart from each other and which face toward a front of the home appliance, a microphone unit including a printed circuit board (PCB) disposed in the accommodation space behind the microphone holes and including microphone chips mounted on the PCB, and a speaker unit disposed in the electrical equipment compartment to correspond to the speaker hole.

A method, apparatus and computer program, the method comprising: receiving a plurality of input signals representing a sound space; using the received plurality of input signals to obtain spatial metadata corresponding to the sound space; using the received plurality of input signals to obtain a first spatial audio signal corresponding to the spatial metadata; and associating the first spatial audio signal with the spatial metadata to enable the spatial metadata to be used to process the first spatial audio signal to obtain a second spatial audio signal.

In some cases, a verbal harassment detection system may use machine learning models to detect verbal harassment in real-time or near real-time. The system may receive an audio segment comprising a portion of audio captured by a microphone located within a vehicle. Further, the system may convert the audio segment to a text segment. The system may provide at least the text segment to a prediction model associated with verbal harassment detection to obtain a harassment prediction. Further, the system may provide the audio segment to an emotion detector to obtain a detected emotion of a speaking user that made an utterance included in the audio segment. Based at least in part on the harassment prediction and the detected emotion, the system may automatically, and without user intervention, determine whether a user is being harassed.

Methods and systems are provided for generating automatic program recommendations based on user interactions. In some embodiments, control circuitry processes verbal data received during an interaction between a user of a user device and a person with whom the user is interacting. The control circuitry analyzes the verbal data to automatically identify a media asset referred to during the interaction by at least one of the user and the person with whom the user is interacting. The control circuitry adds the identified media asset to a list of media assets associated with the user of the user device. The list of media assets is transmitted to a second user device of the user.

A method for evaluating a text content, which may include: after splitting a to-be-evaluated text into a plurality of clauses arranged in sequence according to punctuation information of the to-be-evaluated text, determining a first clause of the plurality of clauses as an actual tune name; then, determining actual prosodic information based on a Chinese phonetic alphabet text of a third clause to a last clause in response to that a number of clauses, whose numbers of Chinese characters satisfy character count requirements of clauses corresponding to the actual tune name, from the third clause to the last clause exceeds a number threshold; and finally, in response to the actual prosodic information being consistent with a standard prosodic information of the actual tune name, evaluating the to-be-evaluated text as a Ci-poetry text.

A method for detecting and categorizing topics in a plurality of interactions includes: extracting, by a processor, a plurality of fragments from the plurality of interactions; filtering, by the processor, the plurality of fragments to generate a filtered plurality of fragments; clustering, by the processor, the filtered fragments into a plurality of base clusters; and clustering, by the processor, the plurality of base clusters into a plurality of hyper clusters.

A system and method for determining whether a textual work submitted for publishing is machine generated or non-machine generated by identifying and quantifying various aspects of the textual work and comparing those aspects to known works. For example, the system and method may identify aspects of a textual work, including, a relationship between the sentences within the textual work, a writing style of the author of the textual work, a grammatical structure of the sentences within the textual work, a quality of the textual work, and other aspects of the textual work. Upon determining that the textual work is machine generated the textual work may be rejected for publishing.