Abstract: A method for training a neural network, includes: training a super network to obtain a network parameter of the super network, wherein each network layer of the super network includes multiple candidate network sub-structures in parallel; for each network layer of the super network, selecting, from the multiple candidate network sub-structures, a candidate network sub-structure to be a target network sub-structure; constructing a sub-network based on target network sub-structures each selected in a respective network layer of the super network; and training the sub-network, by taking the network parameter inherited from the super network as an initial parameter of the sub-network, to obtain a network parameter of the sub-network.

Abstract: A search method for a neural network model structure, includes: generating an initial generation population of network model structure based on multi-objective optimization hyper parameters, as a current generation population of network model structure; performing selection and crossover on the current generation population of network model structure; generating a part of network model structure based on reinforcement learning mutation, and generating a remaining part of network model structure based on random mutation on the selected and crossed network model structure; generating a new population of network model structure based on the part of network model structure generated by reinforcement learning mutation and the remaining part of network model structure generated by random mutation; and searching a next generation population of network model structure based on the current generation population of network model structure and the new population of network model structure.

Abstract: A hypernetwork training method includes: acquiring a multipath neural subnetwork based on a preconstructed initial hypernetwork; training the multipath neural subnetwork to update a weight parameter of each substructure in the multipath neural subnetwork; synchronizing the weight parameter of each substructure in the multipath neural subnetwork to the preconstructed initial hypernetwork; and determining whether the preconstructed initial hypernetwork converges, and if it is determined that the preconstructed initial hypernetwork does not converge, re-executing the acquiring, the training, the synchronizing, and the determining, to obtain a target hypernetwork.

Abstract: The present disclosure relates to an image processing method and apparatus based on a super network, and a computer storage medium. The method can include that a pretrained backbone network is merged with a rear end of a target detection network to obtain a merged super network, the merged super network is trained, Neural Architecture Search (NAS) is performed based on the trained super network to obtain a target detection neural architecture, and an image to be processed is processed by using the target detection neural architecture to obtain an image processing result.

Abstract: A hypernetwork training method includes: acquiring a multipath neural subnetwork based on a preconstructed initial hypernetwork; training the multipath neural subnetwork to update a weight parameter of each substructure in the multipath neural subnetwork; synchronizing the weight parameter of each substructure in the multipath neural subnetwork to the preconstructed initial hypernetwork; and determining whether the preconstructed initial hypernetwork converges, and if it is determined that the preconstructed initial hypernetwork does not converge, re-executing the acquiring, the training, the synchronizing, and the determining, to obtain a target hypernetwork.

Abstract: A method for training a neural network, includes: training a super network to obtain a network parameter of the super network, wherein each network layer of the super network includes multiple candidate network sub-structures in parallel; for each network layer of the super network, selecting, from the multiple candidate network sub-structures, a candidate network sub-structure to be a target network sub-structure; constructing a sub-network based on target network sub-structures each selected in a respective network layer of the super network; and training the sub-network, by taking the network parameter inherited from the super network as an initial parameter of the sub-network, to obtain a network parameter of the sub-network.

Abstract: A search method for a neural network model structure, includes: generating an initial generation population of network model structure based on multi-objective optimization hyper parameters, as a current generation population of network model structure; performing selection and crossover on the current generation population of network model structure; generating a part of network model structure based on reinforcement learning mutation, and generating a remaining part of network model structure based on random mutation on the selected and crossed network model structure; generating a new population of network model structure based on the part of network model structure generated by reinforcement learning mutation and the remaining part of network model structure generated by random mutation; and searching a next generation population of network model structure based on the current generation population of network model structure and the new population of network model structure.

Abstract: A method for subnetwork sampling is applicable to a hypernetwork topoloyly. The hypernetwork topology includes n layers, each layer includes at least two substructures, and each substructure includes hatch normalization (BN) modules in one-to-one correspondence with the substructures of a closest upper layer, n>0 and n being a positive integer. The method includes: a substructure A(N) of an N-th layer is selected, 1>N?n; a selected substructure A(N-1) of an (N?1)-th layer is determined; a BN module C(B) in one-to-to correspondence with A(N-1) is determined from the substructure A(N); and the substructure A(N) is added into a subnetwork through the BN module C(B).

Abstract: A supernet construction method includes: setting a linear connection unit in at least one layer of a supernet, wherein an input end of the linear connection unit is connected to an upper layer of a home layer of the linear connection unit, and an output end is connected to a lower layer of the home layer of the linear connection unit; an output and an input of the linear connection unit form a linear relationship, where the linear relationship includes a linear relationship other than that the output is equal to the input.

Abstract: A super network training method includes: performing sub-network sampling on a super network for multiple rounds to obtain a plurality of sub-networks, wherein for any layer of the super network, different sub-structures are selected when sampling different sub-networks, and training the plurality of sub-networks obtained by sampling and updating the super network.

Abstract: A method for displaying notification information on an electronic device is disclosed. The method includes: receiving notification information in a lockscreen state; determining a user type of a user to view the notification information; determining a display mode of the notification information according to the user type; and displaying the notification information on the electronic device according to the display mode.

Abstract: A method for displaying notification information on an electronic device is disclosed. The method includes: receiving notification information in a lockscreen state; determining a user type of a user to view the notification information; determining a display mode of the notification information according to the user type; and displaying the notification information on the electronic device according to the display mode.

Abstract: A method for displaying notification information on an electronic device is disclosed. The method includes: receiving notification information in a lockscreen state; determining a user type of a user to view the notification information; determining a display mode of the notification information according to the user type; and displaying the notification information on the electronic device according to the display mode.

Abstract: The present disclosure relates to a method and a device for displaying information flows in a social network. The method includes judging whether pieces of target information with a number of pieces being greater than or equal to a set number exist in information flows posted within a set time period by a user of the social network, aggregating the pieces of target information into one or more target information groups according to a preset rule when the pieces of target information exist, and displaying the target information group.

Abstract: A voice prompting method includes detecting whether an unread communication message is received by an electronic device, acquiring, a signal intensity of a wearable device coupled with the electronic device if the unread communication message is received, and performing voice prompt for the unread communication message according to a prompting manner corresponding to the signal intensity.

Abstract: A calling method for use in a terminal configured to communicate with a device, includes: establishing a data channel with the device for assisting a call, when the terminal is conducting the call with the device through a voice channel; and exchanging interaction information with the device through the data channel, the interaction information being configured to assist a first user of the terminal and a second user of the device in coordinating and interacting when conducting the call.

Abstract: A method for telephone interaction includes establishing a voice channel for voice calls with a telephone agent system, establishing a data channel with the telephone agent system if the telephone agent system is in a busy state, and acquiring information provided by the telephone agent system via the data channel. The information is associated with an entertainment service. The method further includes providing the entertainment service to a user based on the information.

Abstract: The present disclosure relates to a method and device for displaying a dynamic icon. The method comprises: detecting an icon display instruction; obtaining a dynamic icon resource pack corresponding to an icon to be displayed based on the icon display instruction; analyzing the dynamic icon resource pack to obtain an icon resource information and a dynamic description information; generating a dynamic icon based on the icon resource information and the dynamic description information; and displaying the dynamic icon. The method and device for displaying the dynamic icon provided by the present disclosure may generate the dynamic icon corresponding to each application program based on the dynamic icon resource pack provided by the application program, so that the icon has a dynamic effect.

Abstract: A method for incoming call notification is disclosed. The method may comprise, when receiving an incoming call, determining voice identification information of the incoming call according to number information of the incoming call, and playing the voice identification information of the incoming call.

Abstract: The present disclosure discloses a method and device for obtaining a voice service. The method includes: obtaining information on a service process corresponding to a service number of a call, wherein the information on the service process comprises input prompt information for each stage of the service process, and the input prompt information is used to prompt a user for input; displaying, after getting through the call, the input prompt information of the service process; receiving an input signal from the user in response to the input prompt information; and sending the input signal to an automatic voice service system to obtain a service therefrom. The method and device avoid the problems that the user does not know what kind of response signal should be input to continue the query until he or she hears the whole voice prompt information.