11 results about "Base class" patented technology

This invention provides a machine learning apparatus, method, and program that improve the classification performance of a classifier by utilizing local features of the input data. [Solution] A machine learning device that continuously learns a small number of new classes compared to the base classes, wherein the feature extraction unit 10 extracts features of the input data using weights learned based on the divided features of the input data. The base class classification unit 20 classifies the base classes based on the features of the input data. The new class classification unit 40 classifies the new classes based on the features of the input data. The attention attractor unit 30 regularizes the weights of the base classes and the weights of the new classes.

This invention discloses a few-sample class-variable audio classification method based on a prototype adaptive network. The steps are as follows: 1) Extracting log-Mel spectrum from input audio samples; 2) In the basic stage, training an initial model consisting of a feature extractor, a prototype classifier, and a class-adjustable prototype adaptive network using base class samples; 3) Extracting representations from the base class training samples, calculating the mean of each base class representation as a prototype, and updating the prototype classifier; 4) In the incremental stage, if the number of categories increases, inputting the representations of the incremental class training samples into the class-adjustable prototype adaptive network to obtain the prototypes of each incremental class and updating the class-adjustable prototype adaptive network; if the number of categories decreases, deleting prototypes of unnecessary categories; 5) Then updating the prototype classifier using the class-adjustable prototype adaptive network; 6) Obtaining the classification category of the test sample. This invention can adaptively expand or shrink the prototype classifier, thereby continuously learning new categories while remembering or deleting old categories, achieving few-sample class-variable audio classification.

The application provides a visual-language model prompting method and system based on a coupling prompt field, and belongs to the fields of artificial intelligence and computer vision. The method comprises the following steps: mapping a base class task and a new class task to a shared feature space, defining a coupling prompt field, and enabling the base class and the new class task to form mutual constraints in the shared feature space; performing scale constraints on the coupling prompt field through projection layer norm alignment and coding layer layer-by-layer norm alignment; integrating the projection layer norm alignment loss and the coding layer layer-by-layer norm alignment loss, fusing the alignment loss and a task loss of a visual-language model, forming a model overall training target, and performing training; performing reasoning on the base class task and the new class task based on the trained model, and outputting a classification prediction result. The application solves the problems of existing end-to-end, decoupling prompt learning, base class-new class isolated optimization, easy falling into local optimum, norm drift and entanglement collapse, and improves the cross-task generalization and robustness of the model.

This invention discloses a few-sample incremental action recognition method and apparatus with test-time adaptive capability, relating to the field of video action recognition technology. The invention is based on a two-stage structure of "base class learning—new class adaptation": in the base class stage, the model learns general spatiotemporal features through sufficient samples and introduces a multi-head self-attention mechanism to improve generalization ability; in the new class stage, a hierarchical test-time adaptive mechanism is used to achieve self-updating of the model during inference. Using this invention, not only is the model's responsiveness to new class samples improved, but catastrophic forgetting and bias accumulation caused by noisy samples are also effectively suppressed, thereby achieving a stable and efficient incremental learning process, and thus improving the accuracy and efficiency of few-sample action recognition.

Continual learning techniques are described for extending the capabilities of a base model, which is trained to predict a set of existing or base classes, to generate a target model that is capable of making predictions for both the existing or base classes and additionally for making predictions for new or custom classes. The techniques described herein enable the target model to be trained such that the model can make predictions involving both base classes and custom classes with high levels of accuracy.

A computer-implemented method for semantic segmentation includes constructing a co-occurrence table that includes co-occurrences of predictions of a pre-trained model for base classes and labels for novel classes from the pre-trained model for base classes and from training data with novel classes. Classifiers are trained that associated with a base class and that classify an input into the base class and one of the novel classes that have co-occurrences with the base class according to the co-occurrence. A prediction is fused from the pre-trained model and the trained classifiers to obtain a final prediction result as a fully labeled image.

The application discloses a source code processing method, computer equipment and program product, and relates to the technical field of computer program code analysis, and the method is characterized in that: an abstract syntax tree is converted into a hybrid code tree that integrates data semantics and control logic, unified standard node basic attributes and semantic information of data nodes and control nodes that inherit a unified base class are relied on, multiple independent intermediate representations are avoided to be built, and operation overhead in a code analysis process is reduced; the original parent-child hierarchical relationship of the abstract syntax tree is completely retained in a conversion process by using the hybrid code tree, sibling nodes in the same layer are arranged according to a source code lexical analysis sequence, and syntax, data and control semantic information are integrated into a hierarchical clear tree structure by combining the design of the data nodes and the control nodes in the hybrid code tree, without additionally adding jump edges, control flow edges and data flow edges, so that the query overhead of the analysis process and the system maintenance complexity are reduced, and the method is suitable for various source code analysis application scenarios such as vulnerability detection and compliance audit.

The present application relates to the technical field of artificial intelligence, and proposes a two-stage fine-tuning and decoupling reasoning method and device of a visual language model, which comprises: in the first stage, panoramic view and subject view are obtained through center cropping, and a learnable prompt vector is initialized for each view, the encoder is frozen, and the prompt vector is optimized to decouple view-specific semantics; in the second stage, the prompt vector is frozen, the text embedding stack is stacked as a trainable weight matrix, the linear classifier is optimized to enhance the discriminability of the base class, and the optimal fusion weight is searched on the validation set; the decoupling reasoning of the third stage comprises fusing the logits of the two views using the optimal weight fusion for the base class, adopting Dempster-Shafer evidence theory fusion for new classes, and introducing an uncertainty measure to realize robust prediction. The present application effectively balances the recognition performance of the base class and the generalization ability of the new class, and exhibits superior accuracy and robustness in a multi-modal classification task.

This invention discloses an apparatus and method for building web front-end applications based on object technology, belonging to the field of Internet technology. It solves the technical problems of existing technologies for building web front-end applications, such as the high learning cost caused by the separation of HTML, CSS, and JavaScript in traditional native methods; difficulties in data access between components during runtime in component-based methods; instability of common basic code and redundant code during component reuse; and difficulties in code understanding, porting, and debugging due to significant differences between component-based implementations and traditional native methods. It includes a main application base class, page view base class, storage base class, event base class, and loading and packaging tool base class, encapsulating common and basic functions in these base classes, while the specific page and function-related code required by the web application is implemented by their respective derived classes. This invention is convenient and quick to build, has a low learning cost, overcomes the problems of unstable common basic code and redundant code in existing component-based methods, and reduces the complexity of inter-component access and operation.

This invention discloses a unified intermediate representation data model and construction method for hardware design, belonging to the field of electronic design automation. The model includes: a unified representation base class; structural relationship graph representation and temporal behavior representation, describing the structural composition and runtime behavior of the design, respectively; formal semantic consistency constraints to ensure semantic consistency between structure and behavior; and a semantic association architecture to establish bidirectional semantic mapping between various scenario representations such as HDL code, schematics, and waveforms and the unified representation base class. The construction method based on this model includes an incremental synchronization process, achieving automatic synchronization after any representation modification through change capture, cross-representation mapping indexing, and change propagation rules. This invention also provides an AI-assisted design method based on this model, improving AI interaction efficiency and the automation level of design verification through multi-dimensional context construction, syntax stripping, incremental interaction protocols, and automatic closed-loop verification.

This invention discloses a method, system, and device for implementing undoable operations in 3D object editing, belonging to the field of computer technology applications. The method includes: defining a base class; defining subclasses; defining a command interface, command classes implementing the command interface, and a command manager; and using the command interface, the command manager performs push and pop operations on command objects, where each command object is an instance of the command class. This invention solves the problem of increased code development and maintenance costs due to adding new classes during 3D scene editing, and implements undoable operations in 3D object editing using a command pattern with a fixed number of command classes.