Related Experiment Video
Updated: Sep 13, 2025
09:01
The Double-H Maze: A Robust Behavioral Test for Learning and Memory in Rodents
Published on: July 8, 2015
12.7K
Int-HRL: towards intention-based hierarchical reinforcement learning
Anna Penzkofer1, Simon Schaefer2, Florian Strohm1
1Institute for Visualisation and Interactive Systems, University of Stuttgart, Pfaffenwaldring 5A, 70569 Stuttgart, Germany.
Neural Computing & Applications
|August 4, 2025
Summary
This study introduces Int-HRL, a new hierarchical reinforcement learning (RL) method. By using human eye gaze to predict intentions, it automatically creates sub-goals, improving sample efficiency in challenging RL tasks.
Area of Science:
- Artificial Intelligence
- Machine Learning
- Robotics
Background:
- Deep reinforcement learning (RL) agents excel at tasks but require vast data for training.
- Hierarchical RL (HRL) improves sample efficiency using structural information but relies on human-annotated sub-goals.
- Discovering effective sub-goals is a major challenge in HRL for complex, long-horizon tasks.
Purpose of the Study:
- To develop a novel HRL method that reduces the need for human-annotated sub-goals.
- To leverage human intention prediction from eye gaze for automated sub-goal generation.
- To enhance sample efficiency in challenging RL environments like Montezuma's Revenge.
Main Methods:
- Predicting human player intentions from eye gaze data.
- Developing an automatic sub-goal extraction pipeline based on predicted intentions.
- Implementing Intention-based Hierarchical Reinforcement Learning (Int-HRL).
Main Results:
- Human intentions can be robustly predicted from eye gaze in long-horizon, sparse-reward tasks.
- The proposed automatic sub-goal extraction pipeline effectively replaces manual annotation.
- Int-HRL demonstrates significantly improved sample efficiency compared to previous HRL methods.
Conclusions:
- Eye gaze-based intention prediction offers a viable alternative to manual sub-goal annotation in HRL.
- Int-HRL significantly enhances sample efficiency, making complex RL tasks more tractable.
- This approach paves the way for more autonomous and efficient learning agents.
Related Concept Videos
01:24Reinforcement Schedules
242
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,...
242
01:23Reinforcement
342
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:
342
01:22Purposive Learning
207
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...
207
01:06Law of Effect
1.6K
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...
1.6K
01:27Associative Learning
576
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...
576
01:24Operant Conditioning Intervention
114
Operant conditioning serves as a foundational principle in therapeutic interventions aimed at modifying maladaptive behaviors. Central to this approach is the notion that behaviors, both adaptive and maladaptive, are learned through reinforcement. By analyzing the environmental factors that reinforce problematic behaviors, clinicians can design interventions to weaken these reinforcements and replace maladaptive behaviors with healthier alternatives.
In operant conditioning, behaviors that are...
In operant conditioning, behaviors that are...
114