Kambria Code Challenge is returning with Quiz 04, which will focus on the AI topic: Reinforcement Learning. If you want to earn generous rewards, you’ll definitely want to join the Kambria Code Challenge! Below we have an intro in reinforcement learning, the topic of our final quiz.
To help you better understand our Quiz 04 AI topic, we will cover an introduction into Reinforcement Learning. Unlike supervised or unsupervised learning, reinforcement learning takes a different approach to machine learning. Rather than utilizing training data or looking for patterns in clusters of data, reinforcement learning looks for effective solutions through taking various actions on an environment.
Made most popular by the infamous AlphaGo project, reinforcement learning uses what is essentially the process of trial and error to learn from the outcomes of a variety of variables and scenarios within a state. As actions are taken within a state, it creates subsequent states in which new actions can be taken and numerical rewards given to the actions, essentially creating a tree of possible outcomes, and then viewing it as a whole to learn what actions are effective and which actions result in failure. The reinforcement agent uses these values to learn how to select the correct actions in order to maximize accumulated rewards over time.
This process is done through the process of exploitation and exploration. In the early stages of the reinforcement learning process, the agent will explore a number of actions to learn which variables lead to the most desired outcomes. Some of these outcomes will lead to a “punishment” or numerical penalties, which are depicted by negative numbers, whereas other actions will lead to “rewards,” which are assigned a positive numerical number. This is the exploration phase.
Over time, the agent will begin to favor actions that it knows lead to desirable outcomes, and avoid actions that it knows will lead to undesired outcomes. Using the numerical values, it begins to learn which actions taken have the highest probability to succeed, and begins to exploit these actions in order to find the solutions more efficiently. As the agent accumulates more data, it is able to make increasingly accurate inferences, thus learning through trial and error.
It is because of this unique process that reinforcement learning is quickly becoming more popular in today’s world. It’s broad applicability is incredibly insightful in solving problems in real-world situations, especially within the fields of robotics, autonomous navigation, logistics, and game-theory. To take a deeper dive into reinforcement learning, read this.
Now that you’ve taken a quick look into reinforcement learning, we hope you’ll join the final Kambria Code Challenge quiz and earn your place at the top of our leader board. Please read the details below to learn how to join. Good luck!
How to Join Quiz 04:
1️⃣ Get access to https://app.kambria.io/bounty
2️⃣ Click on Quiz 04 Bounty
3️⃣ Click on Join This Challenge
4️⃣ Complete your profile page
5️⃣ Receive confirmation email
6️⃣ Take the quiz on time!
Be ready for the quiz to open at 02:00 PM – 02:45 PM on Saturday, August 22, 2020. You will have 30 minutes total to complete the quiz. You can start the quiz anytime after the quiz opens, but you have to submit your responses before the quiz closes at 02:45 PM.
Quiz 04 Includes 2 Bounties With Different Levels & Prizes:
You can only participate in 1 out of 2 bounties:
💎Standard Quiz 04 Bounty: available for everyone to participate. Standard Bounty has 15 questions with the same difficulty as compared to previous quizzes of Kambria Code Challenge.
💎Special Quiz 04 Bounty: only open for participants who have taken at least 2 previous quizzes of Kambria Code Challenge. The Special Bounty will have 20 questions with higher difficulty and bigger prizes.
QUIZ 04 INFORMATION:
⏰ Time: 2 PM – 2:45 PM UTC + 7, Saturday, August 22, 2020
📌 Register here: https://bit.ly/Standard-Quiz-04-Bounty