Stack Exchange Network

Stack Exchange network consists of 175 Q&A communities including Stack Overflow, the largest, most trusted online community for developers to learn, share their knowledge, and build their careers.

Visit Stack Exchange

Questions tagged [reinforcement-learning]

The study of what actions an agent should take in a stochastic environment in order to maximize a cumulative reward.

2
votes
1answer
34 views

What is a sensorimotor connection in plain English?

I have googled, but haven't found any definition simple enough that I understand. I would appreciate it if you could give me an example as well! Thanks, Jack
3
votes
1answer
51 views

What is the fastest way to make an intermittent reinforced behavior become extinct?

Let's suppose I trained my dog to touch a button and win a prize in random sequences. Now I want to make this behavior become extinct. What is the fastest way: Not allow my dog to see the button ...
1
vote
0answers
26 views

How to increase transference on the classroom

During the last times I've been quite interested in the phenomena of transference, which in the context of teaching means transference of knowledge. More than anything, the change of times mainly ...
4
votes
1answer
108 views

How does the brain generate answers to questions?

Firstly, I would say that this question is probably beyond current knowledge. But I would like to hear the latest theories. Given an input "Can you name as many animals as you can until I say 'stop'?...
5
votes
2answers
2k views

Why is conditioned taste aversion an example of classical conditioning (rather than operant)?

The internet seems to be in complete agreement that conditioned taste aversion is an example of classical (Pavlovian) conditioning. My (admittedly limited) understanding of classical conditioning is ...
2
votes
0answers
113 views

How to generate the reward signal in temporal difference (TD) learning algorithm?

With reference to the TD learning algorithm proposed by Sutton and Barto which is given by the equations: $$V_i(t+1) = V_i (t)+ \beta \bigg(\lambda(t+1)+\gamma \bigg[\sum_{j}V_j(t)X_j(t+1)\bigg]-\...
5
votes
1answer
107 views

How does TD-learning explain trace conditioning?

I'm reading Reinforcement learning and causal models by Sam Gershman, which states that TD-learning provides an account of second-order conditioning that provides an explanation for the phenomenon ...
2
votes
1answer
177 views

Does the brain's reward system also work when the reward comes before the experience?

Here's how I understand how the brain rewarding system works, as described by behaviorism: When I eat a chocolate after learning, there's a chance that I'll connect the learning with the pleasurable ...
3
votes
0answers
48 views

Effectiveness of audio / video feedback

Hattie & Timperley (2007) mention a meta-analysis about video or audio feedback with a mean effective size of .64 (Table 2 on page 84). They state that this analysis was part of the meta-analysis ...
2
votes
0answers
90 views

Mathematical models of learning [closed]

Looking for a list of the most common mathematical models of learning. In particular, I am interested in models that explain behavior such as "learning by doing" or "learning from others".
5
votes
2answers
159 views

What's the relation between firing of dopaminergic neurons and dopamine dispersion in terms of neurophysiological processes?

Question: How does the firing of dopaminergic neurons affect the dispersal of dopamine? Evidence of my limited familiarity with dopaminergic neurons and motivation for asking the question: Most of ...
4
votes
2answers
175 views

Cognitive Models of learning Working Memory usage

Recently in deep learning, there's been a surge in learning how to use memories as part of the optimisation process (i.e. LSTM's and Stacks). However, these aren't really analogous to how a cognitive ...
2
votes
1answer
87 views

How do reward signals strengthen synaptic connections in the human brain?

In a vast simplification, the mid-brain sends reward signals (for example through dopaminergic neurons) that tell the rest of the brain whether it succeeded at fulfilling the needs of the organism. If ...
1
vote
0answers
28 views

Reinforcement learning for combining production rules

I'm trying to create a system that learns the Tower of Hanoi puzzle. The system I'm working off of uses a production system (similar to ACT-R), but uses hard-coded production rules. I know that Neil ...
11
votes
1answer
139 views

What's the optimal way to space out rewards in order to maximise motivation?

I'm working on an educational product, and we're thinking of introducing some gamified aspects to motivate students to use the product. One of these will be the concept of 'levels' - when a student ...
2
votes
0answers
22 views

Evidence for different learning applied to negative and positive feedback [duplicate]

I was consulting a friend of mine while building a cognitive model for the Wisconsin Card Sorting Task and he mentioned that it's important to note that people respond differently to to positive ...
6
votes
1answer
297 views

What is an acquired taste?

Some foods are delicious. People love eating them, and the experience has a positive reinforcement. Thus, people will eat the food again. Some foods are instead "challenging", at least initially. For ...
9
votes
1answer
1k views

Evidence that positive rewards are learnt faster than negative rewards?

Its folk psychology wisdom that its easier to reward positive behavior than punish negative behavior (e.g. any book on parenting or dog training), but is there any evidence in the cognitive science ...
4
votes
0answers
68 views

How do mammals explore state spaces in reinforcement learning tasks?

Reinforcement learning is the act of learning how to preform a task given punishment and reward. A "state-space" is the space of choices in a context. When performing a reinforcement learning task, is ...
1
vote
1answer
45 views

Brain areas active while learning hierarchical structure of a problem

There are multiple examples in the machine learning literature of trying to learn the hierarchical structure of a reinforcement learning problem, however have there been any papers tying this learning ...
11
votes
1answer
2k views

Why has behaviourism fallen out of favour?

The reasons for behaviourism as a philosophy and school of psychology to have fallen out of favour are well known and documented. However, when Newton's view of gravity was replaced by general ...
5
votes
1answer
154 views

Evolutionary motivation for the existence of pleasure?

It seems like pleasure goes beyond needs. What is its purpose? Is it supposed to make you not do anything, because of how good you feel? Are there animals that don't feel pleasure, but only have needs ...
4
votes
0answers
53 views

Frequency estimation in binary prediction task

In a binary prediction task, people often match choice probabilities to outcome probabilities (a phenomenon known as probability matching). However, under certain circumstances (eg, existence of a ...
4
votes
2answers
77 views

How do humans learn to combine tasks?

I've been reading about hierarchical learning (a variant of reinforcement learning from what I understand) and how it is shown to allow learning of a higher-level task (the main example is assembly). ...
0
votes
1answer
126 views

What evidence is there to suggest that delayed gratification is taught and learned and not genetic?

Every evening I spend considerable time and energy trying to get my kids to eat their dinner, holding off treats such as chips or dessert. My experience was that my parents did the same. I have ...
0
votes
1answer
1k views

What causes behavioural inhibition?

There are obvious consequences that prevent people from behaving anti-socially or criminally. However there are many behaviours that are within the bounds of social norms, yet there seems to be some "...
2
votes
1answer
66 views

Is it possible to create an inhibition against some activities?

There is an unexplained psychological reason that causes one not to be able to do some random activities (for instance, not being able to eat several days because one's mouth just won't open to take ...
9
votes
1answer
253 views

'Model-free' learning in humans

In reinforcement learning, there is a stark distinction between model-based and model-free learning algorithms, where model-free methods don't make use any explicit information about the dynamics of ...
10
votes
1answer
134 views

What salient features of a {conditioned stimulus,unconditioned stimulus} pair are represented in the lateral amygdala?

In classical conditioning, a conditioned stimulus (CS, e.g., a tone) is presented just before an unconditioned stimulus (UCS, e.g., a mild toe pinch) in repeated trials, such that the CS will ...
10
votes
1answer
517 views

Refinements of Rescorla-Wagner model of classical conditioning

The Rescorla-Wagner model is one of the most commonly discussed mathematical models of classical conditioning. It was wildly popular when it came out in 1972, and very successful. The same math, is ...
14
votes
0answers
2k views

How does goal-tracking and sign-tracking behaviour vary across species?

In Pavlonian (classical) conditioning, conditioned responses of an animal may vary. Some animals focus on the unconditioned stimulus (ie. food/location of food) while others may focus on the ...
10
votes
1answer
264 views

Are there any connectionist models that integrate reinforcement and fully supervised learning?

I've been working on modeling some phenomena involving real-time control in an environment with inherent rewards (specifically, playing a 'pong'-like game), and it's increasingly looking like ...