Related Experiment Video
Updated: Jan 14, 2026
11:53
The Modular Design and Production of an Intelligent Robot Based on a Closed-Loop Control Strategy
Published on: October 14, 2017
12.1K
Discovering state-of-the-art reinforcement learning algorithms
Junhyuk Oh1, Greg Farquhar2, Iurii Kemaev2
1Google DeepMind, London, UK. junhyuk@google.com.
Nature
|October 22, 2025
Summary
Machines can now discover advanced reinforcement learning (RL) rules, outperforming human-designed ones. This breakthrough in artificial intelligence was achieved through meta-learning from agent experiences.
Area of Science:
- Artificial Intelligence
- Machine Learning
- Reinforcement Learning
Background:
- Biological systems utilize evolved reinforcement learning (RL) mechanisms.
- Current artificial agents rely on manually designed learning rules.
- Discovering autonomous RL algorithms has been a long-standing challenge.
Purpose of the Study:
- To demonstrate that machines can autonomously discover state-of-the-art reinforcement learning rules.
- To develop a method for discovering RL rules through meta-learning.
Main Methods:
- Meta-learning from the collective experiences of a population of agents.
- Training agents across a diverse range of complex environments.
- Discovering the specific RL rule governing policy and prediction updates.
Main Results:
- The discovered RL rule outperformed all existing rules on the Atari benchmark.
- The discovered rule surpassed state-of-the-art RL algorithms on unseen challenging benchmarks.
- This represents a significant advancement in reinforcement learning algorithm discovery.
Conclusions:
- It is possible for machines to autonomously discover powerful reinforcement learning algorithms.
- Future artificial intelligence may rely on automatically discovered RL algorithms.
- This approach shifts from manual design to experience-driven discovery of AI learning rules.
Related Concept Videos
01:23Reinforcement
826
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:
826
01:24Reinforcement Schedules
453
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,...
453
01:12Observational Learning
824
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...
824
01:27Associative Learning
1.2K
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...
1.2K
01:06Law of Effect
2.5K
B.F. Skinner, a prominent figure in behavioral psychology, introduced operant conditioning by emphasizing the role of consequences in shaping behavior. This theory builds upon the law of effect proposed by Edward Thorndike, which posits that behaviors followed by satisfying outcomes are likely to be repeated. In contrast, those followed by unsatisfying outcomes are less likely to recur.
Edward Thorndike's foundational work involved studying learning in animals, particularly using puzzle...
Edward Thorndike's foundational work involved studying learning in animals, particularly using puzzle...
2.5K
01:14Avoidance Learning and Learned Helplessness
2.5K
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...
2.5K