Video Experimental Relacionado
Updated: Aug 19, 2025
09:45
New Variations for Strategy Set-shifting in the Rat
Published on: January 23, 2017
8.3K
Dominar el juego de Stratego con aprendizaje de refuerzo multiagente libre de modelos
Julien Perolat1, Bart De Vylder1, Daniel Hennes1
1DeepMind Technologies Ltd., London, UK.
Science (New York, N.Y.)
|December 1, 2022
Resumen
Desarrollamos DeepNash, un agente de inteligencia artificial que domina el complejo juego de Stratego. Esta IA logra jugar a nivel de experto humano en Stratego, un desafiante juego de información imperfecta, utilizando el aprendizaje por refuerzo profundo.
Área de la Ciencia:
- Inteligencia artificial
- Teoría de juegos
- Aprendizaje por refuerzo
Sus antecedentes:
- Stratego es un icónico juego de mesa que históricamente ha desafiado a la inteligencia artificial (IA) debido a su mezcla de profundidad estratégica e información imperfecta.
- El dominio de juegos con información imperfecta, similar al póquer, junto con la planificación estratégica, como el ajedrez, sigue siendo un área de investigación significativa de la IA.
Objetivo del estudio:
- Para presentar DeepNash, un agente de IA capaz de jugar el juego de Stratego a un nivel de experto humano.
- Demostrar un enfoque novedoso para dominar los juegos de información imperfectos utilizando el aprendizaje por refuerzo sin métodos de búsqueda tradicionales.
Principales métodos:
- DeepNash emplea un enfoque de aprendizaje de refuerzo profundo sin modelos teóricos de juegos.
- El agente aprende la estrategia de Stratego de forma autónoma a través del autojuego, comenzando desde cero.
- El método omite notablemente el uso de algoritmos de búsqueda, diferenciándolo de muchas técnicas de juego de IA.
Principales resultados:
- DeepNash logró un rendimiento a nivel de experto humano en el juego de Stratego.
- El agente superó a los métodos de IA de última generación existentes en Stratego.
- DeepNash se aseguró un primer puesto en la plataforma de juegos Gravon en 2022, compitiendo contra expertos humanos.
Conclusiones:
- DeepNash representa un avance significativo en la capacidad de la IA para dominar juegos de información complejos e imperfectos.
- El enfoque de aprendizaje de refuerzo profundo sin búsqueda es efectivo para Stratego.
- Este trabajo empuja los límites de la IA en el juego estratégico, particularmente en dominios con información oculta.
Videos de Conceptos Relacionados
01:29Mechanistic Models: Compartment Models in Algorithms for Numerical Problem Solving
97
Mechanistic models play a crucial role in algorithms for numerical problem-solving, particularly in nonlinear mixed effects modeling (NMEM). These models aim to minimize specific objective functions by evaluating various parameter estimates, leading to the development of systematic algorithms. In some cases, linearization techniques approximate the model using linear equations.
In individual population analyses, different algorithms are employed, such as Cauchy's method, which uses a...
In individual population analyses, different algorithms are employed, such as Cauchy's method, which uses a...
97
01:23Reinforcement
308
Positive and negative reinforcement are key concepts in operant conditioning, a learning process where the consequences of a behavior affect the likelihood of that behavior being repeated.
Positive reinforcement occurs when a behavior is followed by the presentation of a rewarding stimulus, increasing the frequency of that behavior. For example:
Positive reinforcement occurs when a behavior is followed by the presentation of a rewarding stimulus, increasing the frequency of that behavior. For example:
308
01:24Reinforcement Schedules
228
Positive reinforcement is a powerful method for teaching new behaviors to both animals and humans. B.F. Skinner demonstrated this with his experiments using rats in a Skinner box. When a rat pressed a lever, it received a food pellet. This immediate reward encouraged the rat to repeat the behavior. This method, where a reward follows every instance of the behavior, is known as continuous reinforcement. It is highly effective for establishing new behaviors quickly.
Once a behavior is learned,...
Once a behavior is learned,...
228
01:12Observational Learning
263
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...
263
01:22Multi-input and Multi-variable systems
139
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...
In the absence...
139
01:27Associative Learning
507
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...
507