Abstract: A method for acquiring skills through imitation learning by employing a meta imitation learning framework with structured skill discovery (MILD) is presented. The method includes learning behaviors or tasks, by an agent, from demonstrations: by learning to decompose the demonstrations into segments, via a segmentation component, the segments corresponding to skills that are transferrable across different tasks, learning relationships between the skills that are transferrable across the different tasks, employing, via a graph generator, a graph neural network for learning implicit structures of the skills from the demonstrations to define structured skills, and generating policies from the structured skills to allow the agent to acquire the structured skills for application to one or more target tasks.

Abstract: Methods and systems for training a neural network include collecting model exemplar information from edge devices, each model exemplar having been trained using information local to the respective edge devices. The collected model exemplar information is aggregated together using federated averaging. Global model exemplars are trained using federated constrained clustering. The trained global exemplars are transmitted to respective edge devices.

Abstract: A communication device (100) includes a plurality of communication units (103-1, 103-2) that transmit data in different transmission schemes from one another, a route selector (101) that searches for a most appropriate communication route and gives an instruction to transmit data from the communication unit corresponding to the most appropriate communication route, and a converter/sorter (102) that converts data into a format compatible with a transmission scheme of the most appropriate communication route and outputs the converted data to the communication unit indicated by the route selector (101), wherein the communication unit corresponding to the most appropriate communication route transmits the converted data. It is thereby possible to perform communications using the most appropriate communication route among a plurality of transmission schemes even in the case where an application is not compatible with the data format or protocol of the most appropriate communication route.

Abstract: Provided is an object detection system including: a ranging device that emits light so as to pass through a detection region and receives reflected light in which the ranging device is configured to perform a scan by changing an optical path of emitted light; and a reflection member that is provided in a scan range so as to reflect light passing through the detection region toward the ranging device.

Abstract: An information processing system (10) comprises: a storage means (110) for storing face information of a user and re-identification information used for re-identification of the user in association with each other; an acquiring means (120) for acquiring the face information and the re-identification information of a target user; a determining means (130) for performing determination whether or not the face information and the re-identification information of the target user are a pair stored in association with each other; and an output means (140) for outputting according to a result of the determination. According to the information processing system, it is possible to appropriately detect spoofing of a target user.

Abstract: An information processing apparatus includes: an acquisition unit that acquires information about a person including at least an image of the person; a detection unit that detects a face area including a face of the person from the image; an estimation unit that estimates a hidden shield area in a case where at least a part of the face area is hidden; an expression estimation unit that estimates an expression of the person on the basis of the information about the person; an estimated expression image generation unit that generates an estimated expression image of an area corresponding to the shield area, in accordance with the expression estimated by the expression estimation unit; and a composite image generation unit that generates a composite image on the basis of the image and the estimated expression image.

Abstract: A system identification apparatus acquires time series data of input values to an identification target and observed values indicating states of the identification target; constructs a first model obtained by identification of the identification target by training with the time series data; removes noise from the time series data based on the time series data and predicted values of states of the identification target predicted by the first model based on the time series data; and constructs a second model obtained by identification of the identification target by training with the time series data from which the noise has been removed.

Abstract: In the deformation composition data generation apparatus which is implemented by one or more processors, an acquisition unit acquires a distribution of amount of displacement for points in point cloud data, and a composition unit composes the amount of displacement to the points in the point cloud data according to the distribution.

Abstract: Implementations of the present disclosure relate to methods, devices and computer readable media for communications. A method implemented at a terminal device comprises receiving first DCI at a terminal device from a network device. The first DCI comprises a first indication of a first set of cells and scheduling data transmissions in a subset of the first set. The subset comprises more than one of the cells. The method also comprises receiving, from the network device, a second indication of the subset of the first set. The method further comprises performing, based on the first DCI, the data transmissions in the more than one of the cells in the subset.

Abstract: A control apparatus performs: span loss adjustment processing of adjusting a span loss of a first optical signal based on first and third intensities, and adjusting a span loss of a second optical signal based on second and fourth intensities by controlling an optical signal adjustment unit; and intensity difference adjustment processing of adjusting the intensity difference between the first and second optical signals after a span loss adjustment in a second optical communication apparatus by controlling first and second intensity adjustment units. The control apparatus performs the span loss adjustment processing and the intensity difference adjustment processing as an initial setting operation after output of a third optical signal from a first optical communication apparatus starts and performs a monitoring operation of repeating the span loss adjustment processing and the intensity difference adjustment processing on a predetermined cycle after the initial setting operation ends.

Abstract: An image processing device 100 of the present invention includes: a first processing means 121 that performs a first process of recognizing a first target from an image; an identification means 122 that identifies power consumption required for the first process; and a determination means 123 that determines whether to perform a second process of recognizing a second target from the image according to the identified power consumption.

Abstract: In an information processing device, a data acquisition means acquires change history data concerning a change of software. A prompt acquisition means acquires a prompt which is described and instructs to create a comprehension table based on the change history data. A table creation means interprets the prompt using a natural language model and creates a comprehension table based on the change history data. A storage means stores the comprehension table in a storage unit.

Abstract: The present disclosure provides an information processing apparatus including: at least one memory storing instructions; and at least one processor configured to execute the instructions to: generate a first random number key, output the first random number key, and record the first random number key in the memory; encrypt, with the first random number key recorded in the memory, a second random number key used to encrypt first data to be uploaded, to thereby generate second data, record the second data in the memory, and erase the second random number key and the first random number key recorded in the memory; and decrypt the second random number key that decrypts the first data to be downloaded based on the acquired first random number key and the second data recorded in the memory.

Abstract: An information processing apparatus includes a projection image acquisition unit, a determination unit, and a control signal output unit. The projection image acquisition unit acquires a projection image obtained by projecting a predetermined guidance image onto a hand of a user presented on a biometric information reading apparatus that reads predetermined biometric information from the user's hand. The determination unit determines whether the guidance image and the image of the hand included in the projection image correspond to a predetermined condition. The control signal output unit outputs a predetermined control signal to the biometric information reading apparatus based on the determination.

Abstract: An optical network system includes: a transmitter configured to compensate a first nonlinear distortion; a receiver configured to compensate a second nonlinear distortion; an optical repeater configured to compensate a third nonlinear distortion; and a controller configured to control the transmitter, the receiver, and the optical repeater, and determines a distortion compensation section of the optical repeater in a transmission line in which the optical repeater performs nonlinear distortion compensation.

Abstract: The image processing device 1X includes an acquisition means 30X, a determination means 31X, and a detection means 33X. The acquisition means 30X is configured to acquire an endoscopic image obtained by photographing an examination target by a photographing unit provided in an endoscope. The determination means 31X is configured to determine, based on the endoscopic image, a state of the examination target or a photographed part of the examination target. The detection means 33X is configured to detect, based on the state or the photographed part and the endoscopic image, a part of interest to be noticed in the examination target.

Abstract: Embodiments of the present disclosure relate to communications. According to embodiments of the present disclosure, a network device transmits at least one configuration associated with precoding matrix to a terminal device. The network device also transmits downlink control information (DCI) for scheduling a physical uplink shared channel (PUSCH) to the terminal device. The DCI comprises an indication of a precoding matrix index. The terminal device determines a precoding matrix based on the at least one configuration and the indication. The terminal device transmits the PUSCH based on the precoding matrix. In this way, the precoding matrix can be selected properly.

Abstract: A vector decomposer includes: a phase rotation signal generator configured to generate a first phase rotation signal and a second phase rotation signal each having a phase rotated through 90° in a positive direction and a negative direction with respect to an input signal; and a combiner configured to combine the input signal and the first phase rotation signal and generate a first outphasing signal, and combine the input signal and the second phase rotation signal and generate a second outphasing signal.

Abstract: A data providing apparatus acquires a plurality of range proof data, generates a digital signature for each of the range proof data, and transmits a data set, a sanitizable signature, the range proof data, and the digital signature to a data processing apparatus. The data processing apparatus performs a process for performing generalizing processing for processing-target data, performs a process for the sanitizable signature, and selects range proof data corresponding to a range of a generalized attribute value. The data processing apparatus transmits the processed data set, the processed sanitizable signature, the selected range proof data, and the digital signature corresponding to this range proof data to a data receiving apparatus. The data receiving apparatus verifies the sanitizable signature, verifies the digital signature corresponding to the range proof data, and verifies the range proof data.

Abstract: A multi-party computation method includes: using a relay apparatus that relays communication between a user apparatus and MPC operation apparatuses in a multi-party computation system including at least one user apparatus and two or more MPC operation apparatuses; having the relay apparatus receive a first encryption key generated by each MPC operation apparatus and further transmit the first encryption key to the user apparatus; having the relay apparatus receive secret shares generated by the user apparatus and encrypted thereby using the first encryption key and a second encryption key generated by the user apparatus and transmit the encrypted secret shares and the second encryption key to each MPC operation apparatus; and having the relay apparatus further receive shares, encrypted using the user apparatus's second encryption key, of the results of operations performed by the MPC operation apparatuses and transmit the shares to the user apparatus.