8
This is a great question.
Short answer: No, the evidence does not suggest that positive reinforcement is universally more effective than negative reinforcement or punishment. However, there are still good reasons to focus on rewards over punishment in real-life training/learning situations.
Long answer:
The trouble for folk psychology began with Skinner'...
6
As per the comments to the question, human research observing this distinction does exist. CHCH possibly alludes to an article by Gläscher, Daw, Dayan and O'Doherty (2010) which concisely defines the difference between model-free learning and model-based learning:
Reinforcement learning (RL) uses sequential experience with situations (“states”) and ...
answered Apr 22 '15 at 21:56
Christian Hummeluhr
6,6774 gold badges43 silver badges113 bronze badges
5
The task you're referring to is called the Verbal Fluency Test. In "A Biologically Constrained Model of Semantic Memory Search" by Kajic et al. wherein a neural model of this task is described.
The model assumes words were previously encoded with relations between them according to lived experience. Consequently, the words are represented as N-dimensional ...
5
The lateral amygdala appears to be involved in representing fear memories after extinction (Hobin, Goosens and Maren, 2003). The extent of the lateral amygdala's involvement in representing these appears to revolve around the context of the {CS, UCS} pair. The authors state the following in their abstract:
Similarly, the majority of LA neurons exhibited ...
answered Apr 22 '15 at 21:14
Christian Hummeluhr
6,6774 gold badges43 silver badges113 bronze badges
4
To add gaming element to education, you have two options:
Look at currently successful examples of products that gamify learning
Check most successful mobile games (short engagement times)
Here are a couple examples of language learning products that gamify learning (Duolingo and ChineseSkill) They use:
Very short lessons which produce instant sense of ...
4
There is a lot of research here so there is a lot to cover. Please bear with me.
There are what is known as 3 forces of psychology
Behavioural Theory (First Force) – (Short & Thomas, 2014, p. 203)
Psychodynamic Theory (Second Force) – (Short & Thomas, 2014, p. 139)
Humanistic Theory (Third Force) – (Short & Thomas, 2014, p. 79)
...
4
First a couple of definitions:
Positive reinforcement occurs when an event or stimulus is presented as a consequence of a behavior and the behavior increases.
Negative reinforcement occurs when the rate of a behavior increases because an aversive event or stimulus is removed or prevented from happening.
Learning may occur as a result of negative or ...
4
The term I was looking for is "concurrent activities". Some research in the domain of hierarchical learning has been done in this domain by Rohanimanesh and Mahadevan.
According to this literature review on hierarchical learning, basically what they did was determine how multiple tasks can be managed without interfering with one another and how they should ...
4
I don't think anyone has ever bothered (though Ralph Miller might disagree), since many of the 'failures' are outside of the model's purview. The model expresses as simply as possible the profound insight that we learn most when our expectations are not met. Many features of learning won't conform to this general principle (spontaneous recovery) but it doesn'...
3
The Merriam Webster Dictionary defines the word sensorimotor to be:
of, relating to, or functioning in both sensory and motor aspects of bodily activity
and the following relating to the tags you used may help to understand the word.
The idea that we learn by interacting with our environment is probably the first to occur to us when we think about the ...
3
Short answer
Operant conditioning through positive reinforcement is always established by applying a stimulus after the behavior. Negative reinforcers can, however, be deployed before the wanted behavior is initiated.
Background
The reward system can be seen as positive reinforcement.
In experimental paradigms (and everyday life in general), ...
3
I'm not sure I fully understand your design; perhaps you can clarify what you want your network to learn, why TD-learning "isn't cutting it", and what you mean by 'reinforcement' and 'prediction' learning. In particular, TD-learning is a reinforcement learning model, and it does reward based on predicted (and not just observed) outcomes. However, you seem ...
3
One explanation for the development of 'acquired tastes' is a form of reinforcement learning called flavor-flavor learning. The name sounds a bit odd on its own, but should best be seen in relation to the concept of flavor-nutrient learning.
What flavor-flavor learning is
When a previously unexperienced flavor is encountered with familiar flavors, the ...
answered Mar 22 '15 at 15:55
Christian Hummeluhr
6,6774 gold badges43 silver badges113 bronze badges
3
I've mostly figured it out now, and I've a better idea of the words and phrases I need to search to get the information I was asking for.
The short answer comes from wikipedia: "The neurons' somata produce the enzymes that synthesize dopamine, and they are then transmitted via the projecting axons to their synaptic destinations, where most of the dopamine ...
2
ACT-R is a complete cognitive model that incorporates Working Memory, Declarative Memory and Procedural Memory, but also incorporates input (visual and auditory) and output (manual) buffers. It is a really interesting model about the human in its entirety but, it is at a high level of abstraction.
The paper about memory you want is "REFLECTIONS OF THE
...
2
This is an excellent question. The difference between Classical conditioning (also called Pavlovian conditioning) and operant (instrumental) conditioning is subtle for the new student, but can be quite profound when fully appreciated.
Pavlovian conditioning is learning a response that you have no control over. In this context, a conditioned taste aversion (...
2
The conditioning of aversion to taste is part of learning studies in which there is an association between food and flavors with positive reactions such as food preferences or satiety and negative reactions such as lack of appetite, gastric discomfort, etc.
For example, it has been shown that patients with gastric discomfort acquire aversion to food, or ...
2
Dopaminergic neurons synthesize dopamine and release it at their synapses with target neurons (see David Sulzer's review article. The activity of a dopaminergic neuron (its instantaneous firing rate) is one of the main determinants of how much dopamine is released per unit time at its many synaptic terminals. The amount of dopamine synthesized is also ...
1
Operant Conditioning
The general topic you are referring to is called operant conditioning.
Positive/negative refers to whether you are adding something or taking it away.
Reinforcement/punishment refers to whether you are causing a behavior to increase or decrease.
Positive punishment is adding (positive) a bad thing (punishment) like yelling at or ...
1
The fastest way to extinct any behavior is negative reinforcement. But except few specific cases (e.g. self-biting), it's pretty much unethical. Regarding your 2 hypotheses: hypothesis 2 is the classical extinction paradigm. It will first get worse: the animal will repeatedly, almost obsessively, touch the button hoping for a reward. But it will eventually ...
1
See my comment above, but it is possible to assume that B acquires the same reward value as the original reward, so B can then be treated as a reward, which in turn brings A closer to the reward event (now at time B). Does that make sense? In other words, the addition of B allows the reward value to accrue to B and thus brings A closer to reward.
1
I could only find a single instance of learning to leverage a memory. "The Origin of Epistemic Structures and Proto-Representations" by Chandrasekharan and Stewart shows how to include the option to save the current state (by training a neural network to output '1' given the current state input) and input this memory into the state representation of a ...
1
dopamine release changes the nature of synaptic long term potentiation (ltp) thought to underly long term memory. see http://www.sciencedirect.com/science/article/pii/S0896627305003971 for a paper i am familiar with, for an entry point to that field
for a review article:
http://www.nature.com/nrn/journal/v5/n6/abs/nrn1406.html
see section titled "Dopamine ...
1
According to "Frontal Cortex and the Discovery of Abstract Action Rules" by David Badre, Andrew S. Kayser, and MarkD'Esposito:
... the rostro-caudal axis of frontal cortex support rule learning at
higher levels of abstraction. Moreover, these results indicate that
when humans confront new rule learning problems, this rostro-caudal
division of labor ...
1
whilst the nature-nurture issue in mammalian (and particularly human) behaviour is a perennially-fascinating topic, we have learned to be cautious about attempting to construct any strict dichotomies between learned and innate behaviour. suppose you had asked: "did I learn English or just have it?".
successful delayed gratification most likely involves ...
Only top voted, non community-wiki answers of a minimum length are eligible
