11 results about "Feature discovery" patented technology

The embodiment of the invention discloses a cooperative security protection method and device for a vehicle-mounted communication link, a medium and a product. The method comprises the following steps: when a communication link of a vehicle is attacked, collecting attack characteristics; acquiring a comprehensive score of an attacked communication link, determining a target strategy template according to the comprehensive score, and uploading an attack feature and the target strategy template to a cloud; when the comparison between the attack feature and the cloud historical attack feature fails, determining that the attack feature is a novel attack feature and performing depersonalization processing to obtain a general attack feature; obtaining result data that the novel attack feature discovery vehicle uses the target strategy template to protect the novel attack feature so as to obtain a general protection strategy; and synchronizing the general attack feature and the general protection strategy to an associated vehicle to perform safety protection. According to the method, identification and depersonalization of novel attacks and generation of universal protection strategies can be realized, and the strategies are synchronized to associated vehicles, so that various attacks can be timely and effectively handled, and the overall protection capability is improved.

The invention discloses an original message binary feature extraction method based on a time sequence convolutional network. The method comprises the following steps: generating an embedded vector sequence of a message; constructing a time sequence convolutional network architecture; pre-training and classifier training are carried out through self-supervised pre-training, supervised fine tuning and multi-task learning; model interpretation is carried out, and quantized behavior features are extracted based on interpretation results; converting the features into a Snort / Suricata rule format, and establishing a mapping relation from the features to original message segments; and integration to an IDS / IPS engine is realized. The invention also provides an original message binary feature extraction system based on the time sequence convolutional network. A closed loop from feature discovery to automatic rule deployment is constructed, and the detection response efficiency and accuracy of complex network threats are remarkably improved.

This invention belongs to the field of intelligent connected vehicles and vehicle-road-cloud integration technology, and discloses a method for traffic emergence recognition and new function discovery for IVCPS. This invention collects multi-dimensional traffic states, constructs univariate and joint probability densities, and achieves early and accurate recognition of traffic emergence based on distribution distance; then, it constructs a novelty index through mutual information to quantify the novelty of system functions; finally, it uses novelty as a reward-guided parameter to automatically discover new functions. This invention can effectively capture early emergence, quantify new functions, and form an integrated closed loop of recognition and discovery, improving the autonomous cognition and proactive optimization capabilities of IVCPS, and is applicable to intelligent road networks, vehicle-road cooperation, and proactive traffic management.

The application provides a universal domain adaptation method based on a unified optimal transport framework, relates to the fields of deep learning and computer vision, and comprises the following steps: feeding source domain and target domain image samples into a feature extractor to obtain regularized features; performing similarity calculation on the target domain features and source domain prototypes; feeding a similarity matrix into an unbalanced optimal transport solver to obtain an allocation matrix; performing adaptive padding and then solving an optimization problem; using statistical mean to screen out high-confidence samples as common categories, and giving pseudo-labels to calculate a domain alignment loss function; wherein, the target domain samples are spliced with the nearest neighbor features in the memory queue, similarity calculation is performed on the spliced features and target domain prototypes, a similarity matrix is fed into an optimal transport solver to obtain an allocation matrix, the obtained allocation matrix is used as a pseudo-label to calculate a target domain feature discovery loss function, and network training is performed to solve the universal domain adaptation problem.

The invention discloses a daily load characteristic clustering-based short-term power load prediction method. The method comprises the following steps of 1, acquiring power load original data; 2, carrying out standardization processing on the original data; step 3, carrying out daily load characteristic classification processing; step 4, performing causal feature discovery based on a PC (Personal Computer) algorithm; step 5, inputting a PC-TCN prediction model based on daily load characteristic classification; and step 6, measuring and outputting a prediction result. Carrying out data dimension reduction by measuring and calculating daily load characteristics, and replacing original date characteristics with obtained clustering results; meanwhile, based on a causal discovery algorithm, the influence index set is screened, and extraction of causal features is enhanced, so that the operation efficiency and precision are improved.

The invention discloses an automatic analysis and report generation method and device for hydropower station operation data and a medium, and the method comprises the steps: S1, collecting and sorting multiple types of operation data of a hydropower station, collecting various types of operation data, and carrying out the time sequence sorting and synchronous processing of the collected data; s2, operating state identification and reason tracing: distinguishing the current operating state of each device in the hydropower station, searching and sorting out various reasons causing the change according to the state change, and establishing a data-driven causal chain; s3, key index integration and feature discovery: performing integration analysis on each performance index of each device and system, and adaptively adjusting each parameter weight according to an actual operation condition; s4, dynamically generating a report template, automatically judging differentiated requirements of different users on report contents, and automatically selecting corresponding content modules; and S5, report content generation and accurate pushing: aiming at users of different roles, automatically pushing corresponding report contents to the users.

Feature discovery in computer applications is a significant challenge. Existing approaches often lack personalization and are of limited utility. Described herein are approaches for providing personalized feature tours to users of computer applications. In some implementations, tours are recommended based at least in part on prioritizations defined by application developers. In some implementations, tours are recommended based at least in part on past user behavior, the behavior of other users, or both. In some implementations, a personalized feature tour system tailors tours, for example by omitting steps, based on a user's past usage of an application. In some implementations, a personalized feature tour system is configured for particular frameworks or is framework-agnostic. The personalized feature tour system can be used for web applications, mobile applications, and / or desktop applications.

The invention belongs to the technical field of network security detection, and provides an artificial intelligence-driven network intrusion detection method, system and device. The method comprises the steps of multi-source metadata extraction and preprocessing, causal feature discovery and selection, meta learning detection model reasoning, known attack screening, causal comparative analysis and branch judgment, adaptive threshold adjustment and interpretable report generation. According to the method, initial parameters of three layers of MLP are optimized by adopting an MAML framework, so that the model can be quickly adapted only by a small amount of fine tuning in a zero-day attack and few-sample scene, and the limitation that traditional machine learning depends on a large number of historical samples is broken through; meanwhile, by constructing a baseline normal causal graph and an abnormal causal graph, adopting GED to quantify the similarity of the two graphs and accurately distinguishing zero-day attack and concept drift based on a preset threshold value, the industrial pain point that zero-day attack detection is difficult is solved, misinformation caused by concept drift is avoided, it is ensured that resources are only used for real attack analysis, and the detection efficiency is improved.

The application provides a service feature recognition method and device, a recognition equipment and a readable storage medium, and relates to the technical field of communication. The method comprises the following steps: obtaining a first message corresponding to a measured application, wherein the first message is obtained by performing stress testing on the measured application; performing message analysis on the first message to obtain message analysis data; performing service feature recognition processing on the message analysis data based on a machine learning algorithm to obtain the service feature of the measured application. The scheme of the application solves the problem that the existing feature discovery method is prone to missing service characteristics.

Systems and methods for a discovery controller configured with feature discovery for exposed data storage devices. The discovery controller may receive a discovery command from a host system that includes a token value, where the token values map to different storage device configurations among the exposed storage devices. The discovery controller compares the received token value to device token values for each of the storage devices to select the storage devices with matching token values. The discovery controller then generates a discovery log that includes device entries for only those devices with the matching token values and sends it to the host system for selecting storage devices to connect to.

The application discloses a lightweight text classification method and system based on large language model feature discovery, and relates to the technical field of artificial intelligence, and the method comprises the following steps: acquiring training samples corresponding to multiple text classification intents, and training a lightweight classifier based on the training samples; acquiring text to be classified, performing feature extraction on the text to be classified to obtain text features, constructing a sparse binary vector based on the text features, inputting the sparse binary vector into the trained lightweight classifier to perform classification reasoning, and outputting a classification result corresponding to the text to be classified, which helps to solve the problem that the prior art cannot realize lightweight text classification based on large language model features.