Abstract: In an approach for generating a user-specific personal corpus, a processor creates a basic corpus for a first user using a first set of data sources, wherein the basic corpus includes one or more basic words and one or more vectors of the one or more basic words. A processor extracts a set of text from a second set of data sources associated with the first user. Responsive to finding an unknown word included in the set of text extracted, a processor updates the basic corpus, wherein the basic corpus is updated by replacing a vector of the unknown word with an average vector of the one or more basic words in the basic corpus created and registering the unknown word in a first personal corpus.

Abstract: Inside a first case outer frame portion as an outer frame of the first case, a plurality of first core pieces and a partition to separate a pair of adjacent first core pieces among the plurality of first core pieces are disposed. The first case has a shape capable of accommodating at least a part of the second core piece. The first case outer frame portion includes a first case accommodating portion as a portion of the first case outer frame portion that is capable of accommodating the plurality of first core pieces, and a first case cover portion to cover a space inside the first case accommodating portion.

Abstract: An object of the present disclosure is to provide a vehicle outside monitoring apparatus capable of appropriately dealing with an attention-requiring state when a specific target object other than the subject vehicle or target person is in the attention-requiring state.

Abstract: A system and method control an automobile by decelerating the automobile at a first deceleration using a braking mechanism in response to detecting that a parking brake switch is turned on, a second deceleration which is smaller than the first deceleration, in response to not detecting that the parking brake switch is turned on and determining that the driver is incapacitated, and a third deceleration which is smaller than the first deceleration, in response to determining that an SOS switch is turned on, detecting that the parking brake switch is changed from on to off, and not determining that the driver is incapacitated.

Abstract: Some embodiments of the present invention are directed to techniques for training teacher neural networks (TNNs) and student neural networks (SNNs). A training data set is received with a lossless set of data and a corresponding lossy set of data. Two branches of a TNN are established, with one branch trained using the lossless data (a lossless branch) and one trained using the lossy data (a lossy branch). Weights for the two branches are tied together. The lossy branch, now isolated from the lossless branch, generates a set of soft targets for initializing an SNN. These generated soft targets benefit from the training of lossless branch through the weights that were tied together between each branch, despite isolating the lossless branch from the lossy branch during soft-target generation.

Abstract: Systems, computer-implemented methods, and computer program products to facilitate multi-task training a recurrent neural network transducer (RNN-T) using automatic speech recognition (ASR) information are provided. According to an embodiment, a system can comprise a memory that stores computer executable components and a processor that executes the computer executable components stored in the memory. The computer executable components can include an RNN-T that can receive ASR information. The computer executable components can include a voice activity detection (VAD) model that trains the RNN-T using the ASR information, where the RNN-T can further comprise an encoder and a joint network. One or more outputs of the encoder can be integrated with the joint network and one or more outputs of the VAD model.

Abstract: The present invention provides a compound represented by formula (I): wherein the dashed line indicates the presence or absence of a bond; R1 is carboxy or the like; L is substituted or unsubstituted non-aromatic carbocyclyldiyl or the like; R2 is substituted or unsubstituted alkyl; R3 is a hydrogen atom or the like; X is ?CRX— or ?N—; Y is ?CRY— or ?N—; U is —CRU? or —N?; V is —CRV? or —N?; W is ?CRW— or ?N—; ZA is —C? or —N—; ZB is —CR5R6— or the like; ZC is —CR7R8— or the like; RX, RY, RV and RW are each independently a hydrogen atom or the like; RU is a hydrogen atom or the like; R5 and R6 are each independently a hydrogen atom or the like; R7 and R8 are each independently a hydrogen atom or the like; R4 is substituted or unsubstituted alkyloxy or the like, or a pharmaceutically acceptable salt thereof, having an antiviral activity; and a pharmaceutical composition comprising the same.

Abstract: There are provided optical bonding particles capable of improving the visibility and controlling gaps with high accuracy. The optical bonding particles according to the present invention include silicone particles and coating particles or a coating layer disposed on the surface of the silicone particles, and in the optical bonding particles, the absolute value of the difference between the refractive index of the silicone particles and the refractive index of the coating particles or the refractive index of the coating layer is 0.03 or less.

Abstract: An amorphous silica powder that is suitable for obtaining a solid sealant that exhibits superior fluidity and filling properties, and a resin composition obtained by using the amorphous silica powder as a filler. An amorphous silica powder is prepared so as to have a modal diameter within the range of 1 to 10 ?m in the particle diameter frequency distribution and have a frequency of particles having particle diameters of 0.50 to 1.83 ?m of less than 3.0% in the particle diameter frequency distribution.

Abstract: An amorphous silica powder that is suitable for obtaining a liquid sealant that exhibits superior filling properties and preservation properties, and a resin composition obtained by using the amorphous silica powder as a filler. More specifically, an amorphous silica powder is prepared so as to have a modal diameter within the range of 1 to 10 ?m and a frequency of particles having particle diameters of less than 0.50 ?m of 1.0% or more in the particle diameter frequency distribution and have a specific surface area of 1 to 12 m2/g.

Abstract: Provided is a target substance detection device that can detect a target substance accurately and efficiently and can be manufactured at low costs.

Abstract: A student neural network may be trained by a computer-implemented method, including: inputting common input data to each teacher neural network among a plurality of teacher neural networks to obtain a soft label output among a plurality of soft label outputs from each teacher neural network among the plurality of teacher neural networks, and training a student neural network with the input data and the plurality of soft label outputs.

Abstract: A computer-implemented method for preparing training data for a speech recognition model is provided including obtaining a plurality of sentences from a corpus, dividing each phoneme in each sentence of the plurality of sentences into three hidden states, calculating, for each sentence of the plurality of sentences, a score based on a variation in duration of the three hidden states of each phoneme in the sentence, and sorting the plurality of sentences by using the calculated scores.

Abstract: A foam molding method includes: sealing a core material and a skin material with a sealing part; and molding a foam layer on the inside of the sealing part. The sealing of the core material and the skin material with the sealing part includes sealing the core material and the skin material so that the sealing part includes a self sealing part for which an edge part of the skin material is housed in a recess provided in the core material, a clamp sealing part sealed by clamping the core material and the edge part of the skin material by foam molding molds, and a crimp sealing part arranged in a switchover part between the clamp sealing part and the self sealing part, and sealed by crimping the edge part of the skin material to the core material.

Abstract: Embodiments are directed to a system, computer program product, and method for text mining, and dynamic facet and facet value management and application to a document collection. Two or more words from a first document collection are extracted, with the extracted words being associated with an applied annotation. At least one word is selected from the extracted words, designated as a facet, and a value is selectively added to the facet. An analysis of the added value is dynamically performed, and a dictionary with the annotation, facet, and values is constructed and the dictionary is applied to the document collection. A targeted list of documents is returned from the dictionary application to the document collection.

Abstract: A computer-implemented method for training a neural transducer for speech recognition is provided. The method includes initializing the neural transducer having a prediction network and an encoder network and a joint network. The method further includes expanding the prediction network by changing the prediction network to a plurality of prediction-net branches. Each of the prediction-net branches is a prediction network for a respective specific sub-task from among a plurality of specific sub-tasks. The method also includes training, by a hardware processor, an entirety of the neural transducer by using training data sets for all of the plurality of specific sub-tasks. The method additionally includes obtaining a trained neural transducer by fusing the plurality of prediction-net branches.

Abstract: A computer-implemented method for preparing training data for a speech recognition model is provided including obtaining a plurality of audio data sets, each audio data set having a different acoustic feature and sorting sentences from the plurality of audio data sets so that similar sentences from different audio data sets are positioned closely, while imposing a weak constraint on audio length, to train the speech recognition model.

Abstract: There is provided an adhesive capable of improving the visibility and controlling gaps with high accuracy. The adhesive according to the present invention is an adhesive containing a curable component and a gap material, in which the 10% K value of the gap material is 10000 N/mm2 or less, and the absolute value of the difference between the refractive index of a cured product obtained by curing the curable component at 23° C. for 1 hour and the refractive index of the gap material is 0.1 or less.