Nomura Research Institute, Ltd.

There are provided an application that is installed in an information processing terminal used by a user, hooks a user prompt input by the user, and suspends an input of the user prompt to an LLM service provider, and a diagnostics unit that detects whether content of the user prompt transferred from the application is relevant to a predetermined condition related to security, in which, when the diagnostics unit detects that the content of the user prompt is relevant to the predetermined condition, the diagnostics unit responds to the application that the detection has been made as a diagnostics result, and the application displays a warning screen, and inputs the user prompt to the LLM service provider when an instruction from the user is received via the warning screen.

An information processing method is a method executed by an information processing system including: a vehicle including an imaging unit and a communication unit; and a server device. The information processing method includes: requesting, by the server device, the vehicle positioned in an information collecting area including a stop position of a train that moves on a railroad track to provide a peripheral image when the train stops between stations; and transmitting, by the communication unit of the vehicle positioned in the information collecting area, a stoppage factor image that is the peripheral image in which at least one of the train that has stopped in the stop position and the railroad track near the train appears to the server device.

An information processing device executes learning a generative model that generates data belonging to a designated class on the basis of a noise vector and the designated class by executing a learning algorithm of a conditional generative adversarial network, learning a classification model that classifies input data based on whether the input data is in the designated class, and determining whether a property is satisfied when the classification model classifies an output of the generative model by executing a formal verification algorithm. The property indicates that the generative model does not generate data classified into a class different from a first class designated for the generative model, within a range of a certain norm of a noise vector input to the generative model.

An NFT issuing system that issues an NFT related to digital content generated by inference by AI, includes an NFT issuing unit that issues an NFT including information of a learned model used to generate the digital content and a parameter given to the learned model at the time of inference.

A natural language processing system including: a weight array acquisition unit that acquires a weight array related to a weight for determining a label corresponding to a character string, the weight array being generated by learning based on at least one learning character string in which a feature and a label of at least one character string including at least one character are set; a feature extraction unit that extracts a feature corresponding to a target character string; and a label determination unit that determines the label of the target character string on the basis of the learned model generated by learning, the weight array, and the extracted at least one feature.

A vehicle transmits images in which garbage collection sites appear to a server. The server performs: identifying, among the plurality of images acquired from the vehicle, a yet-to-be-visited collection site image in which a yet-to-be-visited collection site appears that is a collection site by which a garbage truck has not yet passed in a travel route of the garbage truck; based on the identified yet-to-be-visited collection site image, calculating an estimated amount of garbage brought together at the yet-to-be-visited collection site; calculating an expected time of arrival at a collection site by which the garbage truck passes after the yet-to-be-visited collection site in the travel route such that the expected time of arrival at that collection site becomes later when the estimated garbage amount is large than when the estimated garbage amount is small; and notifying a user terminal of the expected time of arrival.

A failure notification system mounted on a vehicle includes an imaging device, a processing device, and a transmission device. The processing device acquires video data of a preceding vehicle captured by the imaging device. The processing device performs failure determination based on the acquired video data. As a result of the failure determination by the processing device, when a failure of a brake light of the preceding vehicle is detected, the processing device causes the transmission device to transmit information indicating that the brake light is failed to a reception device of the preceding vehicle.

An information processing apparatus acquires a user setting related to a feature of adversarial training and performs the adversarial training that trains a neural network by using training data including an adversarial sample and correct answer data indicating an original classification class, and the user setting, the adversarial training training the neural network that outputs a misclassification class in a case where the adversarial sample is input so as to output the original classification class in a case where the adversarial sample is input. The apparatus determines, by executing a format validation algorithm, that the adversarial sample does not exist within a predetermined range of noise of specific data in the neural network using a weighting factor obtained by the adversarial training.

In a navigation system, a data center and a plurality of vehicles are connected via a communication network. The data center accepts specific section information, which can identify a specific section for which the capturing of road conditions is requested, from a capturing requester. The data center proposes capturing-requested routes, each of which goes through a specific section, to a user who performed a route search with a destination set using a navigation device, based on the specific section information. The data center stores, in a storage device, a video that was uploaded by a user, that was captured by a dashcam while the vehicle was traveling in the specific section, and that shows the road conditions of the specific section, along with capturing position information that can identify the position where the video was captured.

The dialogue unit 11 grasps a business instruction by a dialogue with a user 2 by the generative AI, and stores a content of the dialogue as a dialogue log in a short-term storage 16 , the solution unit 12 creates a task list by decomposing the business instruction into tasks by the generative AI and passes an execution instruction of each task to the execution unit 13 , the execution unit 13 executes a work on a corresponding data source 3 by the generative AI corresponding to the task related to the execution instruction, and passes an execution result to the solution unit 12 , and the monitoring unit 14 refers to the dialogue log at any time, grasps a context of the dialogue by the generative AI, predicts a content to be dealt with next, stores the content as a summary in the short-term storage 16.