Abstract: The present invention relates to compounds of formula (I), wherein R1 to R4, A and L are as described herein, and their pharmaceutically acceptable salt thereof, and compositions including the compounds and methods of using the compounds.

Abstract: A method is provided for obtaining new distant germplasm of intergeneric hybridized potato based on grafting-mentor, including the following steps: (1) using young branches, which are growing well, free of diseases, non-lignified, and from an upper part of Lycium as rootstock; (2) using seedlings of diploid primitive cultivated potato with 2-3 unfolded leaves as scion; (3) grafting; (4) conducting post-grafting management; (5) obtaining distant hybridized germplasm by using potato flower buds from the grafted plants for pollination, obtaining seeds of potato varieties after self-fertility, performing pre-germination on selfing seeds, obtaining true seed mini-tubers, and using these mini-tubers for further grafting and pollination to obtain distant hybridized seeds, performing tissue culture propagation on the distant hybridized seeds to obtain F1 asexual clone line foundation seeds.

Abstract: Methods, apparatus, systems and articles of manufacture for loss-error-aware quantization of a low-bit neural network are disclosed. An example apparatus includes a network weight partitioner to partition unquantized network weights of a first network model into a first group to be quantized and a second group to be retrained. The example apparatus includes a loss calculator to process network weights to calculate a first loss. The example apparatus includes a weight quantizer to quantize the first group of network weights to generate low-bit second network weights. In the example apparatus, the loss calculator is to determine a difference between the first loss and a second loss. The example apparatus includes a weight updater to update the second group of network weights based on the difference. The example apparatus includes a network model deployer to deploy a low-bit network model including the low-bit second network weights.

Abstract: A method is provided that includes: creating a plurality of first model instances of a first service model to be deployed; allocating an inference service for each of a plurality of first model instances from the plurality of inference services; calling, for each first model instance, a loading interface of the inference service allocated for the first model instance to mount a weight file; determining, in response to a user request for a target service model, a target model instance from a plurality of model instances of the target service model to respond to the user request; and calling a target inference service allocated for the target model instance to use computing resources configured for the target inference service to run, in the target model instance, a base model mounted with a target weight file, and obtain a request result of the user request.

Abstract: An aerosol generation device includes: a chamber configured to removably receive an aerosol generation article including a magnetic material, a heater configurated to heat the aerosol generation article recieved in the chamber to generate an aerosol, a detection circuit including a capacitor connected in series with the heater, and a controller configured to control the detection circuit to have a direct current flowing therethrough the determine, based on a duration for which a potential difference between two ends of the capactior reaches a preset potential difference threshold, that the aerosol generation article is received in the chamber or the aerosol generation article is removed from the chamber. The device determines, based on the duration for which the potential differene between the two ends of the capactiro reaches the preset potentional difference theshold, whether the aerosol generation article is inserted into the chamber, and then controls an action of the heater.

Abstract: Methods, apparatus, systems and articles of manufacture for loss-error-aware quantization of a low-bit neural network are disclosed. An example apparatus includes a network weight partitioner to partition unquantized network weights of a first network model into a first group to be quantized and a second group to be retrained. The example apparatus includes a loss calculator to process network weights to calculate a first loss. The example apparatus includes a weight quantizer to quantize the first group of network weights to generate low-bit second network weights. In the example apparatus, the loss calculator is to determine a difference between the first loss and a second loss. The example apparatus includes a weight updater to update the second group of network weights based on the difference. The example apparatus includes a network model deployer to deploy a low-bit network model including the low-bit second network weights.

Abstract: A delivery robot includes: a mobile chassis and a housing provided on the mobile chassis. The housing comprises a bottom plate, at least two first supporting members, and a top plate; the bottom plate is provided at the top of the mobile chassis, and the at least two first supporting members are provided between the bottom plate and the top plate, and are respectively located on two sides of a front end or two sides of a rear end of the bottom plate, wherein the front end of the bottom plate is a front surface of the housing, and the rear end of the bottom plate is a back surface of the housing.

Abstract: This application discloses an aerosol generation apparatus, including: a liquid storage unit, configured to store a liquid for generating an aerosol; a heating element, configured to heat the liquid; a liquid transfer unit, configured to transfer the liquid stored in the liquid storage unit to the heating element a power supply, configured to supply power to the heating element and a circuit, configured to determine whether the liquid stored in the liquid storage unit has decreased to a threshold after a specific preset time point during a period when the heating element starts for heating, where the preset time point is determined according to a time when a resistance change rate of the heating element decreases to a first resistance change rate during the heating period. Hence, dry burning occurs in the second half of each inhalation can be accurately determined.

Abstract: An electronic atomizing device includes a battery cell, a heating element, a charging socket used for power charging the battery cell, and an airflow sensor used for measuring airflows flowing through the electronic atomizing device and being formed by inhaling of users. The airflow sensor includes a first port and a second port. The first port includes two electrical connections. A first electrical connection is electrically connected with a positive electrode of the charging socket, and a second electrical connection is electrically connected with the heating element via a switch transistor. The second port is electrically connected with a positive electrode of the battery cell. The switch transistor is disposed to be switched on when the positive electrode of the charging socket is at a low voltage level, and disposed to be switched off when the positive electrode of the charging socket is at a high voltage level.

Abstract: An electronic atomizing device includes a battery cell, a heating element, a charging socket used for power charging the battery cell, and an airflow sensor used for measuring airflows flowing through the electronic atomizing device and being formed by inhaling of users. The airflow sensor includes a first port and a second port. The first port includes two electrical connections. A first electrical connection is electrically connected with a positive electrode of the charging socket, and a second electrical connection is electrically connected with the heating element via a switch transistor. The second port is electrically connected with a positive electrode of the battery cell. The switch transistor is disposed to be switched on when the positive electrode of the charging socket is at a low voltage level, and disposed to be switched off when the positive electrode of the charging socket is at a high voltage level.

Abstract: Methods, apparatus, systems and articles of manufacture for loss-error-aware quantization of a low-bit neural network are disclosed. An example apparatus includes a network weight partitioner to partition unquantized network weights of a first network model into a first group to be quantized and a second group to be retrained. The example apparatus includes a loss calculator to process network weights to calculate a first loss. The example apparatus includes a weight quantizer to quantize the first group of network weights to generate low-bit second network weights. In the example apparatus, the loss calculator is to determine a difference between the first loss and a second loss. The example apparatus includes a weight updater to update the second group of network weights based on the difference. The example apparatus includes a network model deployer to deploy a low-bit network model including the low-bit second network weights.

Abstract: A composition and method is described for intracellular delivery of fullerene containing peptides. The composition and method involve fullerene-substituted phenylalanine as part of a peptide based delivery system. The presence of a fullerene-substituted amino acid in a peptide is found to alter the intracellular transport properties of the peptide.