Related Experiment Video
Updated: Aug 22, 2025
09:01
The Double-H Maze: A Robust Behavioral Test for Learning and Memory in Rodents
Published on: July 8, 2015
12.7K
Acquisition of chess knowledge in AlphaZero
Thomas McGrath1, Andrei Kapishnikov2, Nenad Tomašev1
1DeepMind, London, United Kingdom.
Summary
AlphaZero, a chess AI, learns human-like strategies by playing itself. Analysis reveals how its neural network represents these concepts, offering insights into artificial intelligence learning.
Area of Science:
- Artificial Intelligence
- Computational Neuroscience
- Game Theory
Background:
- AlphaZero, a neural network engine, learns chess independently.
- It achieves superhuman performance without human game data.
Purpose of the Study:
- To analyze the knowledge acquired by AlphaZero.
- To understand how AlphaZero represents human-analogous chess concepts.
- To investigate the internal workings of a self-learning AI.
Main Methods:
- Utilized linear probes on AlphaZero's internal network states.
- Quantified concept representation within the neural network.
- Conducted a behavioral analysis of AlphaZero's opening game strategies.
Main Results:
- Identified specific internal states where chess concepts are represented.
- Demonstrated that AlphaZero learns concepts similar to human players.
- Opening analysis, with expert commentary, revealed strategic depth.
Conclusions:
- AlphaZero's self-play learning leads to the acquisition of human-analogous chess knowledge.
- Internal network analysis provides a window into AI concept representation.
- The findings contribute to understanding AI learning and strategic development.
Related Concept Videos
01:05Introduction to z Scores
463
A z score (or standardized value) is measured in units of the standard deviation. It indicates how many standard deviations the value x is above (to the right of) or below (to the left of) the mean, μ. Values of x that are larger than the mean have positive z scores, and values of x that are smaller than the mean have negative z scores. If x equals the mean, then x has a zero z score. It is important to note that the mean of the z scores is zero, and the standard deviation is one.
z scores...
z scores...
463
01:12Observational Learning
269
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...
269
01:27Associative Learning
510
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...
510
01:21Cognitive Learning
476
Cognitive learning is based on purposive behavior, incidental learning, and insight learning.
E. C. Tolman's theory of purposive behavior emphasizes that much behavior is goal-directed. He argued that to understand behavior, we must look at the entire sequence of actions leading to a goal. For instance, high school students study hard, not just due to past reinforcement but also to achieve the goal of getting into a good college.
Tolman introduced the idea that behavior is influenced by...
E. C. Tolman's theory of purposive behavior emphasizes that much behavior is goal-directed. He argued that to understand behavior, we must look at the entire sequence of actions leading to a goal. For instance, high school students study hard, not just due to past reinforcement but also to achieve the goal of getting into a good college.
Tolman introduced the idea that behavior is influenced by...
476
01:22Purposive Learning
183
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...
183
01:22Machines: Problem Solving I
388
A toggle clamp is a mechanical device commonly used for holding and clamping objects in various applications, such as woodworking, metalworking, and assembly operations. Consider a toggle clamp subjected to a force of 200 N at the handle. The vertical clamping force can be calculated, provided the dimensions of the toggle clamp are known.
The toggle clamp system is a machine structure consisting of movable, pin-connected multi-force members that form a stabilized system to transmit forces. The...
The toggle clamp system is a machine structure consisting of movable, pin-connected multi-force members that form a stabilized system to transmit forces. The...
388