Unlock DREAM 11 app with me using artificial intelligence algorithm and move ahead towards your…

Original article was published by Florina Regius on Artificial Intelligence on Medium

Unlock DREAM 11 app📳 with me using artificial intelligence algorithm and move ahead towards your goal !

Let me spill the beans, it is none other than REINFORCEMENT ALGORITHM! Yes it is…wondering how? Allow me to enlighten you thoroughly! I know this will open up your eyes technically as well as financially!

HEART OF THE APP: In this app, they are using multiple agents/bots for betting the outcome along with real people as user. This is to make sure that best agent/bot must bring highest amount of cash to company as these bots/agents play from the company side. With each of their right prediction, they earn money as reward and learns more. With passing time, these agents/bot becomes powerful. Brainstorming part is — Let’s say 1 lakh user gives Rs 25 for a match, rewards won will be 25 lakhs. All I am saying the amount of money artificial intelligence produces for the company.

This is an adapted version of the “Gambler’s Problem” that I’ve applied to DREAM 11 app’s solution method.

Reference : The Gambler Problem as discussed in Example 4.3 in Reinforcement Learning: An Introduction by Richard S. Sutton and Andrew G. Barto.

Description: Dream 11 app user has the opportunity to make bets on the outcomes of a sequence of match outcomes(wins and loses). If the match comes up leads to victory, user wins as many money as he has staked on that match; if it is leading towards losing, user loses his stake. The game ends when the user wins by reaching his goal of money, or loses by running out of money.

The Method : On each match, the user must decide what portion of his capital to stake, in integer numbers of money. This problem can be formulated as an undiscounted, episodic, finite MDP. The state is the user’s capital, s ∈ {1, 2, . . . , 99} and the actions are stakes, a ∈ {0, 1, . . . , min(s, 100−s)}. The reward is zero on all transitions except those on which the user reaches his goal, when it is +1. The state-value function then gives the probability of winning from each state. A policy is a mapping from levels of capital to stakes. The optimal policy maximizes the probability of reaching the goal. Let x denote the probability of the match coming up in winning terms. If x is known, then the entire problem is known and it can be solved, for instance, by value iteration.

Keyword: MDP: Markov Decision Process

Reference : The Gambler Problem as discussed in Example 4.3 in Reinforcement Learning: An Introduction by Richard S. Sutton and Andrew G. Barto.

Acknowledgement : Jerome Johnson for support as well as guidance and Siraj Raval for explanation

People ! Please feel free to DM me for codes and doubts. I would love to answer. Also, I would like to build something like this app for football match prediction! I know there are Chelsea , Arsenal etc fans out there…Hence, It will be fun to work along !