Hi everyone, I was studying reinforcement learning this week and I found something interesting. I’ve my master’s degree in Economics and I always try to connect machine learning with statistics and econometrics to gain a deeper intuition behind the core concept.
I found that the concept of Reinforcement learning is much similar to the concept of ‘Adaptive decision making’ concept of macroeconomics in which an economic agent/Individual maximizes his decisions based on the analysis of past experiences and the overall information available to him at that particular moment.
In reinforcement learning, an agent first takes a random move and then maximizes its next move on the basis of the information gathered and analyzed.
In theory, they are looking similar so I’m just curious whether it is actually similar or just a coincidence?
3 Likes
Hello @Ashish_Raj ,
That’s a great observation!
Reinforcement learning of ML and adaptive decision making of macroeconomics, despite being from different disciplines has a similarity where the main driving principle that is Optimizations based on feedbacks - agents learning from their experiences and use that info for enhanced decision making , both work in dynamic environment, both are iterative. So there are similarities indeed!
This particular principle is also very effective and relatable in our daily life. Just like : we learn from our mistakes. Be it particular decisions or coding errors which we fix by ourselves, we remember them next time when it occurs and debugging gets faster or we don’t make that error again - optimization. Possibly there could be a similar reference to human psychology as well.
So things are relatable.
That’s the beauty of science! 
With regards,
Nilosree Sengupta
2 Likes