Abstract: A computer-implemented method, according to one embodiment, includes obtaining information about a plurality of remote storage systems. Each of the remote storage systems includes a cache. The method further includes generating a routing table based on the information, the routing table indicating potential write data migration paths from a cache of a local storage system to the caches of the remote storage systems. A write data migration plan is generated based on the routing table. The write data migration plan specifies which of the caches of the remote storage systems to migrate at least some write data originally intended to be stored on the cache of the local storage system to. In response to a determination that a first predetermined condition associated with the cache of the local storage system is met, the method includes causing the write data migration plan to be performed.

Abstract: A monolithic integrated multi-segment cascade optical frequency comb and its chip are disclosed, which belongs to the technical field of sensing detection, quantum information and optical communication technology. The optical frequency comb includes a first semiconductor passive mode-locking laser, a semiconductor optical amplifier, and a second semiconductor passive mode-locking laser sequentially integrated and connected; the first semiconductor passive mode-locking laser includes a first reverse bias absorption area integrated with an optogalvanic distribution grating and a first gain cavity length extender (coupled multi-ring or multi-disk); the second semiconductor passive mode-locking laser includes a second reverse bias absorption area and a second gain cavity length extender (coupled multi-ring or multi-disk); each structure is connected to each other by electrical isolation grooves.

Abstract: This application relates to a slot negotiation method and a device. The method includes: A transmitter sends a first FlexE overhead frame to a receiver, to request active/standby calendar switching. When the receiver is in a restart state, the receiver does not respond to the received first FlexE overhead frame. In addition, the RX sends a routine update second FlexE overhead frame to the transmitter. Determining that the second FlexE overhead frame is not a response to the first FlexE overhead frame, the transmitter sends a third FlexE overhead frame to request active/standby calendar switching again. According to the method in this application, incorrect calendar switching on the transmitter side caused by a mistaken response of the receiver can be avoided. This reduces the likelihood of a service interruption caused by the existing slot negotiation mechanism.

Abstract: A wiring board includes a base substrate and first connection pads disposed on the base substrate. The first connection pads each include electrical connection layer(s); each electrical connection layer includes a main material layer and protective layer(s) disposed on a side of the main material layer away from the base substrate; the protective layer(s) include a first reference protective layer, which is a protective layer farthest away from the base substrate in the protective layer(s); and a material of the main material layer includes copper. The electrical connection layer(s) includes a first electrical connection layer, which is an electrical connection layer farthest away from the base substrate in the electrical connection layer(s); and in protective layer(s) in the first electrical connection layer, at least a material of the first reference protective layer is capable of forming a first intermetallic compound with a first solder.

Abstract: A wall mountable sensor module includes a housing defining an internal space that is segmented into a first internal space and a second internal space. The first internal space defines an air channel that extends from an air inlet to an air outlet. Two or more sensors are configured to be exposed to the air flow channel. A first sensor is configured to detect a first air parameter and a second sensor is configured to detect a second different air parameter, wherein the second sensor is situated downstream of the first sensor in the air flow channel. The sensor module includes a fan housed by the housing, the fan configured to cause an airflow to flow in through the air inlet, through the air flow channel thereby exposing each of the sensors to the airflow, and out through the air outlet.

Abstract: A method of performing speech recognition using a two-pass deliberation architecture includes receiving a first-pass hypothesis and an encoded acoustic frame and encoding the first-pass hypothesis at a hypothesis encoder. The first-pass hypothesis is generated by a recurrent neural network (RNN) decoder model for the encoded acoustic frame. The method also includes generating, using a first attention mechanism attending to the encoded acoustic frame, a first context vector, and generating, using a second attention mechanism attending to the encoded first-pass hypothesis, a second context vector. The method also includes decoding the first context vector and the second context vector at a context vector decoder to form a second-pass hypothesis.

Abstract: A method includes receiving audio data encoding an utterance spoken by a native speaker of a first language, and receiving a biasing term list including one or more terms in a second language different than the first language. The method also includes processing, using a speech recognition model, acoustic features derived from the audio data to generate speech recognition scores for both wordpieces and corresponding phoneme sequences in the first language. The method also includes rescoring the speech recognition scores for the phoneme sequences based on the one or more terms in the biasing term list, and executing, using the speech recognition scores for the wordpieces and the rescored speech recognition scores for the phoneme sequences, a decoding graph to generate a transcription for the utterance.

Abstract: A method of performing speech recognition using a two-pass deliberation architecture includes receiving a first-pass hypothesis and an encoded acoustic frame and encoding the first-pass hypothesis at a hypothesis encoder. The first-pass hypothesis is generated by a recurrent neural network (RNN) decoder model for the encoded acoustic frame. The method also includes generating, using a first attention mechanism attending to the encoded acoustic frame, a first context vector, and generating, using a second attention mechanism attending to the encoded first-pass hypothesis, a second context vector. The method also includes decoding the first context vector and the second context vector at a context vector decoder to form a second-pass hypothesis.

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 method of performing speech recognition using a two-pass deliberation architecture includes receiving a first-pass hypothesis and an encoded acoustic frame and encoding the first-pass hypothesis at a hypothesis encoder. The first-pass hypothesis is generated by a recurrent neural network (RNN) decoder model for the encoded acoustic frame. The method also includes generating, using a first attention mechanism attending to the encoded acoustic frame, a first context vector, and generating, using a second attention mechanism attending to the encoded first-pass hypothesis, a second context vector.

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 method includes receiving a sequence of acoustic frames and generating, by a first encoder, a first higher order feature representation for a corresponding acoustic frame in the sequence of acoustic frames. The method also includes generating, by a first pass transducer decoder, a first pass speech recognition hypothesis for a corresponding first higher order feature representation and generating, by a text encoder, a text encoding for a corresponding first pass speech recognition hypothesis. The method also includes generating, by a second encoder, a second higher order feature representation for a corresponding first higher order feature representation. The method also includes generating, by a second pass transducer decoder, a second pass speech recognition hypothesis using a corresponding second higher order feature representation and a corresponding text encoding.

Abstract: A method includes receiving a speech recognition result, and using a confidence estimation module (CEM), for each sub-word unit in a sequence of hypothesized sub-word units for the speech recognition result: obtaining a respective confidence embedding that represents a set of confidence features; generating, using a first attention mechanism, a confidence feature vector; generating, using a second attention mechanism, an acoustic context vector; and generating, as output from an output layer of the CEM, a respective confidence output score for each corresponding sub-word unit based on the confidence feature vector and the acoustic feature vector received as input by the output layer of the CEM. For each of the one or more words formed by the sequence of hypothesized sub-word units, the method also includes determining a respective word-level confidence score for the word. The method also includes determining an utterance-level confidence score by aggregating the word-level confidence scores.

Abstract: A method includes receiving a speech recognition result, and using a confidence estimation module (CEM), for each sub-word unit in a sequence of hypothesized sub-word units for the speech recognition result: obtaining a respective confidence embedding that represents a set of confidence features; generating, using a first attention mechanism, a confidence feature vector; generating, using a second attention mechanism, an acoustic context vector; and generating, as output from an output layer of the CEM, a respective confidence output score for each corresponding sub-word unit based on the confidence feature vector and the acoustic feature vector received as input by the output layer of the CEM. For each of the one or more words formed by the sequence of hypothesized sub-word units, the method also includes determining a respective word-level confidence score for the word. The method also includes determining an utterance-level confidence score by aggregating the word-level confidence scores.

Abstract: A method includes receiving audio data encoding an utterance spoken by a native speaker of a first language, and receiving a biasing term list including one or more terms in a second language different than the first language. The method also includes processing, using a speech recognition model, acoustic features derived from the audio data to generate speech recognition scores for both wordpieces and corresponding phoneme sequences in the first language. The method also includes rescoring the speech recognition scores for the phoneme sequences based on the one or more terms in the biasing term list, and executing, using the speech recognition scores for the wordpieces and the rescored speech recognition scores for the phoneme sequences, a decoding graph to generate a transcription for the utterance.

Abstract: A method includes receiving audio data encoding an utterance spoken by a native speaker of a first language, and receiving a biasing term list including one or more terms in a second language different than the first language. The method also includes processing, using a speech recognition model, acoustic features derived from the audio data to generate speech recognition scores for both wordpieces and corresponding phoneme sequences in the first language. The method also includes rescoring the speech recognition scores for the phoneme sequences based on the one or more terms in the biasing term list, and executing, using the speech recognition scores for the wordpieces and the rescored speech recognition scores for the phoneme sequences, a decoding graph to generate a transcription for the utterance.

Abstract: A method of performing speech recognition using a two-pass deliberation architecture includes receiving a first-pass hypothesis and an encoded acoustic frame and encoding the first-pass hypothesis at a hypothesis encoder. The first-pass hypothesis is generated by a recurrent neural network (RNN) decoder model for the encoded acoustic frame. The method also includes generating, using a first attention mechanism attending to the encoded acoustic frame, a first context vector, and generating, using a second attention mechanism attending to the encoded first-pass hypothesis, a second context vector.

Abstract: An electrical connector is provided for a mating connector to be inserted thereto. The electrical connector includes an insulating body having a tongue, and multiple terminals. A middle grounding sheet is located between the terminals in upper and lower rows, and has at least one first through hole. Each of left and right sides of the middle grounding sheet has a latch slot and a notch. The insulating body has a first insulating post entering the first through hole and a second insulating post entering the notch. A dimension of the second insulating post is larger than a dimension of the first insulating post. A metal shell frames outside the insulating body to form an insertion space. A first clearance between a first protruding portion and a shielding casing of the mating connector is more than or equal to 0.001 mm and less than or equal to 0.0449 mm.

Abstract: A method includes receiving audio data encoding an utterance spoken by a native speaker of a first language, and receiving a biasing term list including one or more terms in a second language different than the first language. The method also includes processing, using a speech recognition model, acoustic features derived from the audio data to generate speech recognition scores for both wordpieces and corresponding phoneme sequences in the first language. The method also includes rescoring the speech recognition scores for the phoneme sequences based on the one or more terms in the biasing term list, and executing, using the speech recognition scores for the wordpieces and the rescored speech recognition scores for the phoneme sequences, a decoding graph to generate a transcription for the utterance.

Abstract: According to embodiments, a derived cache that is derived from a first instance of particular data is used to speed up queries and other operations over a second instance of the particular data. Traditionally, a DBMS generates and uses derived cache data only for the database data from which the derived data was derived. However, according to embodiments, derived cache data associated with a first instance of database data is relocated to the location of a second, newly created, instance of the database data. Since the derived cache data is derived from an identical copy of the database data, the cache data derived for the first instance can successfully be used to speed up applications running over the second instance of the database data.