Reinforcement Learning Algorithms
You open your eyes, look the environment and scan everything (place, objects, illumination, people, their actions and behaviors), sending in fractions of seconds all this information to your brain. In parallel your skin and organs have sensors and nerves everywhere that will also inform about the temperature, vital functions, the body feeling, how you are today.
The brain will use all this information combined with the hippocampus’ memory, which holds key memories from what you have lived and what you have testified, to take decisions, as simple as move your arm, hand and fingers to scratch your head, open your mouth and move your lips and lungs to yawn or throw adrenaline into your neuron system so you can run faster if a threat is coming towards you. We don’t use prompts, but take decisions, deal with results and learn everyday trying to map as many patterns as possible to increase success rate.
Like many parts of the brain's limbic system, the hippocampus is involved in?memory, learning, and emotions. Its largest job is to hold short-term memories and transfer them to long-term storage in our brains. But also recovers this memory. It also plays a role in emotional processing, including anxiety and avoidance behaviors. This combined with the frontal cortex that helps to take decisions, triggers creativity and innovation, makes humans a powerful machine, very strategic, that can come with sophisticated alternatives to resolve complex situations and act. ?
If action was successful, if result is positive, reward system comes into picture. Indeed, reward-related activities (e.g., feeding, exercise, substance use, and social interactions), lead to an elevated level of dopamine, alters rhythms in the suprachiasmatic nucleus (yellow dot in the second picture) and the brain’s reward system. The opposite is also true, amygdala will be there to process the avoidance system (fear, conditioning, extinction...).
?AI will get closer to how humans behave soon. Q-Start, Q* or Q-Learning is a model free reinforcement learning algorithm. This is the beginning of AI focus as a vehicle to understand the environment, scanning from multiple places and by deduction take decisions. These mechanisms, which tend to improve more and more, also have a heavy emphasis in the reward avoidance systems, as humans do.
领英推荐
Future of AI is much more than a smart prompt with great communication skills and good answers. Goal is to mimic human behaviors having systems fully autonomous, where sources are not pre-defined, including the environment, to take decisions based on policies, judgement, deductions, memories, facts, and consequences.
To learn basics on Q* idea click link below (good article from Chathurangi Shyalika).
More about the brain reward system:
If you like this topic and want to see more content about this, click the like button or add a comment so we can engage further!
?