Video Experimental Relacionado
Updated: Jan 29, 2026
08:28
Synthesis of Phase-shift Nanoemulsions with Narrow Size Distributions for Acoustic Droplet Vaporization and Bubble-enhanced Ultrasound-mediated Ablation
Published on: September 13, 2012
11.6K
Aprendizaje Federado Bajo Cambios de Distribución Evolutivos
Xuwei Tan1, Tian Xie1, Xue Zheng2
1Department of Computer Science and Engineering, The Ohio State University, Columbus, OH 43210, USA.
Entropy (Basel, Switzerland)
|January 28, 2026
Resumen
Los modelos de aprendizaje federado (FL) ahora pueden adaptarse a las distribuciones de datos cambiantes de los clientes con el tiempo. Los nuevos algoritmos, FedEvolve y FedEvp, aseguran que los modelos se generalicen a datos futuros a pesar de los patrones evolutivos.
Área de la Ciencia:
- Machine Learning
- Sistemas Distribuidos
- Inteligencia Artificial
Sus antecedentes:
- El aprendizaje federado (FL) permite el entrenamiento colaborativo de modelos sin centralizar datos brutos.
- Los métodos FL existentes a menudo asumen distribuciones de datos de clientes estáticas, lo cual no es realista.
- Los escenarios FL del mundo real implican cambios dinámicos y no triviales en los datos de los clientes con el tiempo, incluso entre el entrenamiento y la prueba.
Objetivo del estudio:
- Desarrollar algoritmos de FL capaces de entrenar modelos con datos de clientes que evolucionan temporalmente.
- Mejorar la robustez del sistema FL contra los cambios de distribución de datos evolutivos.
- Lograr la generalización a datos objetivo futuros en entornos FL dinámicos.
Principales métodos:
- Algoritmo FedEvolve propuesto: Modela explícitamente la evolución temporal aprendiendo transiciones de representación entre dominios de datos de clientes consecutivos.
- Algoritmo FedEvp propuesto: Aprende una representación invariante al dominio en evolución alineando los datos actuales con prototipos continuamente actualizados de todos los dominios pasados.
- Se realizaron experimentos exhaustivos en conjuntos de datos sintéticos y del mundo real.
Principales resultados:
- FedEvolve y FedEvp demostraron mejoras significativas de rendimiento sobre las líneas de base tradicionales de FL.
- Los algoritmos propuestos capturaron efectivamente los patrones evolutivos en las distribuciones de datos de los clientes.
- Los métodos mostraron robustez y sólidas capacidades de generalización bajo cambios de distribución evolutivos.
Conclusiones:
- Los algoritmos propuestos FedEvolve y FedEvp abordan con éxito el desafío de las distribuciones dinámicas de datos de clientes en el aprendizaje federado.
- Estos enfoques novedosos permiten que los sistemas FL se generalicen de manera efectiva a datos futuros a pesar de los cambios temporales.
- Los hallazgos resaltan la importancia de tener en cuenta la evolución de los datos en aplicaciones FL realistas.
Videos de Conceptos Relacionados
01:14Avoidance Learning and Learned Helplessness
2.6K
Avoidance learning and learned helplessness are critical concepts in understanding behavioral responses to negative stimuli.
Avoidance learning occurs when an organism learns that a specific behavior can prevent an unpleasant outcome. For example, a student who receives a bad grade may start studying harder to avoid future poor grades. This behavior persists even when the negative outcome is no longer present. Avoidance learning is powerful because it maintains behavior in the absence of the...
Avoidance learning occurs when an organism learns that a specific behavior can prevent an unpleasant outcome. For example, a student who receives a bad grade may start studying harder to avoid future poor grades. This behavior persists even when the negative outcome is no longer present. Avoidance learning is powerful because it maintains behavior in the absence of the...
2.6K
01:25Drug Distribution: Volume of Distribution
7.4K
The volume of distribution refers to the theoretical volume necessary to contain the entire amount of an administered drug at the same concentration observed in the blood plasma. The body's intracellular fluid compartment, which makes up two-thirds of the total body water, is contrasted with the extracellular fluid compartment—comprising plasma and interstitial fluid—that accounts for one-third. The volume of distribution can vary depending on the characteristics of the drug.
7.4K
01:19F Distribution
10.7K
The F distribution was named after Sir Ronald Fisher, an English statistician. The F statistic is a ratio (a fraction) with two sets of degrees of freedom; one for the numerator and one for the denominator. The F distribution is derived from the Student's t distribution. The values of the F distribution are squares of the corresponding values of the t distribution. One-Way ANOVA expands the t test for comparing more than two groups. The scope of that derivation is beyond the level of this...
10.7K
01:27Associative Learning
1.3K
Associative learning is a fundamental concept in behavioral psychology, wherein a connection is established between two stimuli or events, leading to a learned response. This process is critical in understanding how behaviors are acquired and modified. Conditioning, the mechanism through which associations are formed, can be divided into two main types: classical conditioning and operant conditioning, each elucidating different aspects of associative learning.
Classical conditioning, also known...
Classical conditioning, also known...
1.3K
01:22Purposive Learning
470
E. C. Tolman emphasized the purposiveness of behavior — the idea that much of our behavior is goal-directed. For instance, employees who aim for a promotion work diligently to meet their targets. Tolman argued that when classical conditioning and operant conditioning occur, the organism acquires certain expectations. In classical conditioning, a child might fear a dog because they expect it to bite. In operant conditioning, a person might consistently work overtime because they expect a...
470
01:12Observational Learning
893
Albert Bandura's observational learning, also known as imitation or modeling, occurs when a person observes and imitates another's behavior. It is a quicker process than operant conditioning. A well-known example is the Bobo doll study, where children who saw an adult acting aggressively towards the doll were more likely to act aggressively when left alone, compared to those who observed a nonaggressive adult. Many psychologists view observational learning as a form of latent learning...
893