Abstract: A method includes receiving a training dataset that includes one or more spoken training utterances for training an automatic speech recognition (ASR) model. Each spoken training utterance in the training dataset paired with a corresponding transcription and a corresponding target sequence of auxiliary tokens. For each spoken training utterance, the method includes generating a speech recognition hypothesis for a corresponding spoken training utterance, determining a speech recognition loss based on the speech recognition hypothesis and the corresponding transcription, generating a predicted auxiliary token for the corresponding spoken training utterance, and determining an auxiliary task loss based on the predicted auxiliary token and the corresponding target sequence of auxiliary tokens. The method also includes the ASR model jointly on the speech recognition loss and the auxiliary task loss determined for each spoken training utterance.

Abstract: A video processing method includes obtaining motion information of a neighboring block of a current image block, dividing the current image block into a plurality of sub-blocks in response to the neighboring block satisfying a preset condition, determining, in a time-domain reference image of the current image block, related blocks of the plurality of sub-blocks according to a motion vector of the neighboring block, and performing prediction on the current image block according to motion vectors of the related blocks.

Abstract: A video image processing method includes determining a related block of a current image block according to a motion vector of a target neighboring block, the current image block, and a collocated frame of the current image block, decoding the current image block according to a motion vector of the related block of the current image block, constructing motion vectors of part of control points of the current image block according to neighboring blocks of the part of the control points of the current image block, and adding the motion vectors of the part of the control points of the current image block to a motion vector candidate list of the current image block.

Abstract: A method includes receiving training data that includes a set of unspoken textual utterances. For each respective unspoken textual utterance, the method includes, tokenizing the respective textual utterance into a sequence of sub-word units, generating a first higher order textual feature representation for a corresponding sub-word unit tokenized from the respective unspoken textual utterance, receiving the first higher order textual feature representation generated by a text encoder, and generating a first probability distribution over possible text units. The method also includes training an encoder based on the first probability distribution over possible text units generated by a first-pass decoder for each respective unspoken textual utterance in the set of unspoken textual utterances.

Abstract: An automated speech recognition (ASR) model includes a first encoder, a first encoder, a second encoder, and a second decoder. The first encoder receives, as input, a sequence of acoustic frames, and generates, at each of a plurality of output steps, a first higher order feature representation for a corresponding acoustic frame in the sequence of acoustic frames. The first decoder receives, as input, the first higher order feature representation generated by the first encoder, and generates a first probability distribution over possible speech recognition hypotheses. The second encoder receives, as input, the first higher order feature representation generated by the first encoder, and generates a second higher order feature representation for a corresponding first higher order feature frame. The second decoder receives, as input, the second higher order feature representation generated by the second encoder, and generates a second probability distribution over possible speech recognition hypotheses.

Abstract: A method includes generating, using an audio encoder, a higher-order feature representation for each acoustic frame in a sequence of acoustic frames; generating, using a decoder, based on the higher-order feature representation, a plurality of speech recognition hypotheses, each hypotheses corresponding to a candidate transcription of an utterance and having an associated first likelihood score; generating, using an external language model, for each speech recognition hypothesis, a second likelihood score; determining, using a learnable fusion module, for each speech recognition hypothesis, a set of fusion weights based on the higher-order feature representation and the speech recognition hypothesis; and generating, using the learnable fusion module, for each speech recognition hypothesis, a third likelihood score based on the first likelihood score, the second likelihood score, and the set of fusion weights, the audio encoder and decoder trained using minimum additive error rate training in the presence of t

Abstract: A method of text-only and semi-supervised training for deliberation includes receiving training data including unspoken textual utterances that are each not paired with any corresponding spoken utterance of non-synthetic speech, and training a deliberation model that includes a text encoder and a deliberation decoder on the unspoken textual utterances. The method also includes receiving, at the trained deliberation model, first-pass hypotheses and non-causal acoustic embeddings. The first-pass hypotheses is generated by a recurrent neural network-transducer (RNN-T) decoder for the non-causal acoustic embeddings encoded by a non-causal encoder. The method also includes encoding, using the text encoder, the first-pass hypotheses generated by the RNN-T decoder, and generating, using the deliberation decoder attending to both the first-pass hypotheses and the non-causal acoustic embeddings, second-pass hypotheses.

Abstract: A computer-implemented method includes receiving a sequence of acoustic frames as input to an automatic speech recognition (ASR) model. Here, the ASR model includes a causal encoder and a decoder. The method also includes generating, by the causal encoder, a first higher order feature representation for a corresponding acoustic frame in the sequence of acoustic frames. The method also includes generating, by the decoder, a first probability distribution over possible speech recognition hypotheses. Here, the causal encoder includes a stack of causal encoder layers each including a Recurrent Neural Network (RNN) Attention-Performer module that applies linear attention.

Abstract: A method includes receiving an initial alignment for a candidate hypothesis generated by a transducer decoder model during a first pass. Here, the candidate hypothesis corresponds to a candidate transcription for an utterance and the initial alignment for the candidate hypothesis includes a sequence of output labels. Each output label corresponds to a blank symbol or a hypothesized sub-word unit. The method also include receiving a subsequent sequence of audio encodings characterizing the utterance. During an initial refinement step, the method also includes generating a new alignment for a rescored sequence of output labels using a non-autoregressive decoder. The non-autoregressive decoder is configured to receive the initial alignment for the candidate hypothesis and the subsequent sequence of audio encodings.

Abstract: A system of recommending an item, a method of recommending an item by a system of recommending an item, a computer system, and a computer-readable storage medium are provided. The system includes: an item expansion module configured to expand a content input by a user in response to the content input by the user, to generate an item set of interest to the user, the item set includes one or more items; a price radar module configured to monitor a discount information of the item in the item set; and a price monitoring module configured to calculate an actual price of the item based on the discount information of the item, maintain a price change record of the item in the item set, and determine whether to push a prompt information to the user or not according to the calculated actual price of the item and the price change record.

Abstract: System and methods for identifying a text word from a spoken utterance are provided. An ensemble BPE system that includes a phone BPE system and a character BPE system receives a spoken utterance. Both BPE systems include a multi-level language model (LM) and an acoustic model. The phone BPE system identifies first words from the spoken utterance and determine a first score for each first word. The first words are converted into character sequences. The character BPE model converts the character sequences into second words and determines a second score for each second word. For each word from the first words that matches a word in the second words the first and second scores are combined. The text word is the word with a highest score.

Abstract: Embodiments described herein combine both masked reconstruction and predictive coding. Specifically, unlike contrastive learning, the mutual information between past states and future states are directly estimated. The context information can also be directly captured via shifted masked reconstruction—unlike standard masked reconstruction, the target reconstructed observations are shifted slightly towards the future to incorporate more predictability. The estimated mutual information and shifted masked reconstruction loss can then be combined as the loss function to update the neural model.

Abstract: System and methods for identifying a text word from a spoken utterance are provided. An ensemble BPE system that includes a phone BPE system and a character BPE system receives a spoken utterance. Both BPE systems include a multi-level language model (LM) and an acoustic model. The phone BPE system identifies first words from the spoken utterance and determine a first score for each first word. The first words are converted into character sequences. The character BPE model converts the character sequences into second words and determines a second score for each second word. For each word from the first words that matches a word in the second words the first and second scores are combined. The text word is the word with a highest score.

Abstract: An audio event detection system that processes audio data into audio feature data and processes the audio feature data using pre-configured candidate interval lengths to identify top candidate regions of the feature data that may include an audio event. The feature data from the top candidate regions are then scored by a classifier, where the score indicates a likelihood that the candidate region corresponds to a desired audio event. The scores are compared to a threshold, and if the threshold is satisfied, the top scoring candidate region is determined to include an audio event.

Abstract: An audio event detection system that subsamples input audio data using a series of recurrent neural networks to create data of a coarser time scale than the audio data. Data frames corresponding to the coarser time scale may then be upsampled to data frames that match the finer time scale of the original audio data frames. The resulting data frames are then scored with a classifier to determine a likelihood that the individual frames correspond to an audio event. Each frame is then weighted by its score and a composite weighted frame is created by summing the weighted frames and dividing by the cumulative score. The composite weighted frame is then scored by the classifier. The resulting score is taken as an overall score indicating a likelihood that the input audio data includes an audio event.

Abstract: The present application relates to a method of treating a fabric article so that it has a characteristic smell normally associated with garments that have been exposed to natural sunlight, the method comprising positioning a fabric article (1) to be treated in an enclosure (8) and irradiating said fabric article positioned in said enclosure with ultraviolet light with a wavelength of between 280 nm and 400 nm and so that it is subjected to a predetermined radiant exposure. A device for treating a fabric article to replicate the characteristic effect of exposing said fabric article to natural sunlight is also disclosed.

Abstract: A pH monitoring device includes a chamber for containing a solution; and a polymer immersed in the solution. The physical state of the polymer is changeable in dependence on whether the pH of the solution exceeds a threshold value. The pH monitoring device includes further includes a detector configured to detect the change of the physical state of the polymer and thus determine the pH of the solution.

Abstract: The invention relates to an air purification apparatus. The apparatus is disposed in a window for separating an indoor space and an outdoor space. The apparatus comprising: an inlet chamber having a first inlet, a second inlet and an outlet, wherein the first inlet is operatively open to the outdoor space and the second inlet is operatively open to the indoor space; an air pumping unit for pumping air from the inlet chamber to the indoor space through the outlet, wherein the air is pumped into the inlet chamber through the first inlet or through the second inlet; and a filtering unit disposed upstream of the air pumping unit, for filtering pollutants in the air.

Abstract: The invention relates to an air purification apparatus. The apparatus is disposed in a window for separating an indoor space and an outdoor space. The apparatus comprising: an inlet chamber having a first inlet, a second inlet and an outlet, wherein the first inlet is operatively open to the outdoor space and the second inlet is operatively open to the indoor space; an air pumping unit for pumping air from the inlet chamber to the indoor space through the outlet, wherein the air is pumped into the inlet chamber through the first inlet or through the second inlet; and a filtering unit disposed upstream of the air pumping unit, for filtering pollutants in the air.

Abstract: The present application relates to a method of treating a fabric article so that it has a characteristic smell normally associated with garments that have been exposed to natural sunlight, the method comprising positioning a fabric article (1) to be treated in an enclosure (8) and irradiating said fabric article positioned in said enclosure with ultraviolet light with a wavelength of between 280 nm and 400 nm and so that it is subjected to a predetermined radiant exposure. A device for treating a fabric article to replicate the characteristic effect of exposing said fabric article to natural sunlight is also disclosed.