JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us
Esta página ha sido traducida por una máquina. Otras páginas pueden seguir apareciendo en inglés. View in English

PiCCL: Un marco de aprendizaje multiview ligero para la clasificación de imágenes

  • 0Research Center for Biomedical Engineering, Medical Innovation and Research Division, Chinese PLA General Hospital, Beijing, People's Republic of China.
Plos One +

|

25 de agosto de 2025

|

25 de agosto de 2025

Ver abstracta en PubMed

Resumen

Este resumen es generado por máquina.

El Componente Primario de Aprendizaje Contrastivo (PiCCL) es un nuevo marco auto-supervisado que utiliza una red Siamesa múltiple para un aprendizaje eficiente. Se obtienen resultados de vanguardia, especialmente en escenarios de pequeños lotes.

Área De La Ciencia

  • Ciencias de la computación
  • Inteligencia artificial
  • Aprendizaje automático

Sus Antecedentes

  • El aprendizaje auto-supervisado (SSL) es crucial para aprovechar los datos sin etiqueta.
  • Los marcos de aprendizaje de contraste existentes a menudo utilizan estructuras complejas y funciones de pérdida.
  • Se necesitan métodos SSL más simples y eficientes.

Objetivo Del Estudio

  • Introducir PiCCL (Componente Primario de Aprendizaje Contrastivo), un nuevo marco de aprendizaje contrastivo auto-supervisado.
  • Desarrollar un método SSL computacionalmente ligero y generalizable.
  • Demostrar la eficacia de PiCCL en varios conjuntos de datos y escenarios de aprendizaje.

Principales Métodos

  • Utilizó una red siamesa múltiple con múltiples ramas idénticas.
  • Empleó una sencilla estrategia de aumento de imagen para generar múltiples muestras positivas.
  • Diseñó una función de pérdida de peso computacional personalizada (PiCLoss).

Principales Resultados

  • Alcanzó el máximo rendimiento en los conjuntos de datos CIFAR-10 (94%), CIFAR-100 (72%) y STL-10 (97%).
  • Se ha demostrado un rendimiento superior en escenarios de aprendizaje en lotes pequeños (precisión del 93% en STL-10 con tamaño de lote 8).
  • Superó a los algoritmos de última generación auto-supervisados en las pruebas de referencia.

Conclusiones

  • PiCCL ofrece un enfoque simple, ligero y efectivo para el aprendizaje de contraste auto-supervisado.
  • La estructura siamesa múltiple y la función de pérdida personalizada mejoran la eficiencia y el rendimiento del aprendizaje.
  • PiCCL muestra una promesa particular para entornos con recursos limitados y aprendizaje en pequeños lotes.

Videos de Conceptos Relacionados

Aggregates Classification 01:29

381

Aggregate classification is generally based on its size, petrographic characteristics, weight, and source. Size classification ranges from coarse to fine aggregates, defined by the size of the particles. Coarse aggregates are particles that do not pass through ASTM sieve No. 4, and aggregates that pass through the sieve are fine aggregates.
Petrographic classification groups aggregates based on common mineralogical characteristics. Some of the common mineral groups found in aggregates are...

Force Classification 01:22

1.6K

Forces play a crucial role in the study of physics and engineering. They are essential in describing the motion, behavior, and equilibrium of objects in the physical world. Forces can be classified based on their origin, type, and direction of action.
Contact and non-contact forces are two of the most widely used categories of forces. As the name suggests, contact forces require physical contact between two objects to act upon each other. Examples of contact forces include frictional,...

Classification of Systems-I 01:26

296

Linearity is a system property characterized by a direct input-output relationship, combining homogeneity and additivity.
Homogeneity dictates that if an input x(t) is multiplied by a constant c, the output y(t) is multiplied by the same constant. Mathematically, this is expressed as:

Additivity means that the response to the sum of multiple inputs is the sum of their individual responses. For inputs x1(t) and x2(t) producing outputs y1(t) and y2(t), respectively:

Combining homogeneity and...

Classification of Systems-II 01:31

240

Continuous-time systems have continuous input and output signals, with time measured continuously. These systems are generally defined by differential or algebraic equations. For instance, in an RC circuit, the relationship between input and output voltage is expressed through a differential equation derived from Ohm's law and the capacitor relation,

Discrete-time systems have input and output signals at specific intervals, defined at distinct instants by difference equations. An example...

Introduction to Learning 01:18

530

Learning is the process of acquiring knowledge or skills through practice or experience, leading to long-lasting behavioral changes. This acquisition occurs through interaction with the environment and requires practice or experience. For instance, mastering a skill such as surfing requires considerable practice and experience, highlighting the essential role of repeated interactions with the environment in learning.
In contrast to learned behaviors, unlearned behaviors such as crying, sexual...

Multi-input and Multi-variable systems 01:22

149

Cruise control systems in cars are designed as multi-input systems to maintain a driver's desired speed while compensating for external disturbances such as changes in terrain. The block diagram for a cruise control system typically includes two main inputs: the desired speed set by the driver and any external disturbances, such as the incline of the road. By adjusting the engine throttle, the system maintains the vehicle's speed as close to the desired value as possible.
In the absence...