model-free learning
[/ˈmɑːdəl friː ˈlɜːrnɪŋ/]
noun
aprendizagem sem modelo
1. A machine learning approach where an agent learns optimal behavior through trial and error without building or using an explicit model of the environment's dynamics or transition probabilities
In model-free learning, the agent uses Q-learning to discover the best actions based solely on rewards received, without needing to understand how the environment works.
Na aprendizagem sem modelo, o agente utiliza Q-learning para descobrir as melhores ações baseado apenas nas recompensas recebidas, sem necessidade de entender como o ambiente funciona.
2. A reinforcement learning method that learns value functions or policies directly from experience without constructing a predictive model of state transitions
Model-free learning algorithms like SARSA and Deep Q-Networks have shown remarkable success in complex domains such as video games.
Algoritmos de aprendizagem sem modelo como SARSA e Deep Q-Networks demonstraram sucesso notável em domínios complexos como videogames.
This is a technical term primarily used in artificial intelligence, machine learning, and cognitive science communities. It gained prominence in both Brazilian and international research spheres with the rise of deep reinforcement learning in the 2010s. The term reflects a fundamental paradigm in how intelligent systems can learn without explicit environmental models, contrasting with traditional planning approaches.
Look up more words on Fala2Me
The free English-Portuguese dictionary with real Brazilian accents, NYC slang, conjugator and more
Open Fala2Me →