Scribd Logo
Upload

Welcome to Scribd!

  • 6 views

    Reinforcement Learning

    Uploaded by

    crazybruce2024

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd

    Reinforcement Learning

    Uploaded by

    crazybruce2024
    6 views10 pages

    Copyright

    © © All Rights Reserved

    Available Formats

    PPTX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    6 views10 pages

    Reinforcement Learning

    Uploaded by

    crazybruce2024

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    of 10

    Unit V Reinforcement Learning

    Dr.M.Thamarai
    Professor, ECE Department,
    SVEC
    Reinforcement learning (RL)
    • Reinforcement learning (RL) is a type of
    machine learning where an agent learns how
    to make decisions by performing actions in an
    environment to maximize some notion of
    cumulative reward.
    • Unlike supervised learning, where the model
    learns from labeled data, RL is based on trial
    and error, and it uses feedback from its
    actions to improve its performance over time.
    Supervised Vs Reinforcement Learning
    Key Concepts in Reinforcement Learning

    • 1.Agent: The learner or decision-maker that


    interacts with the environment.
    • 2. Environment: The world or system within
    which the agent operates and learns.
    • 3. State (S): A representation of the current
    situation of the environment.
    • 4.Action (A): Any set of moves or decisions the
    agent can make in a given state.
    Key Concepts in Reinforcement Learning
    • 5. Reward (R) : A feedback signal the agent receives after
    taking an action in a state, indicating the immediate gain or
    loss.
    • 6.Policy (π): A strategy or function that maps states to
    actions and determines the agent's behavior.
    • 7. Value Function (V): A function that estimates the expected
    cumulative reward (or "value") of being in a certain state.
    • 8.Q-Function (Q): A function that estimates the expected
    cumulative reward of taking a particular action in a given
    state, helping the agent decide between different actions.
    How RL Works?
    • 1.Exploration vs. Exploitation: The agent must balance exploring new
    actions to discover potentially better rewards (exploration) with
    exploiting known actions that have previously yielded good rewards
    (exploitation).

    • 2. Learning Process: The agent goes through a cycle of observing the


    current state, choosing and performing an action, receiving a reward,
    and moving to the next state. The reward serves as feedback, which
    helps the agent learn which actions yield the highest cumulative
    reward over time.

    • 3. Goal: The agent’s goal is to learn a policy that maximizes


    cumulative reward over time, sometimes discounted over the future.
    Types of RL Algorithms
    • 1.Model-Free vs. Model-Based RL:
    - Model-Free: The agent learns purely from interaction without
    understanding the environment’s underlying model (e.g., Q-
    learning, SARSA).
    -Model-Based: The agent learns or uses a model of the
    environment to plan actions (e.g., Dyna-Q, AlphaGo).
    • 2. Value-Based Methods: Focus on estimating the value of states or
    state-action pairs, typically with Q-learning and Deep Q Networks
    (DQNs).
    • 3. Policy-Based Methods: Directly optimize the policy function
    without estimating values, usually through gradient-based methods
    (e.g., REINFORCE, Proximal Policy Optimization).
    • 4.Actor-Critic Methods: Combine value-based and policy-based
    approaches. The “actor” updates the policy, and the “critic”
    estimates the value function to critique the actions taken (e.g.,
    Applications of RL
    • -Games: RL is used in board games like chess and Go (e.g.,
    AlphaGo), video games, and other competitive environments.
    • - Robotics: Teaching robots to navigate, manipulate objects,
    and perform tasks autonomously.
    • - Healthcare: Assisting in personalized treatment plans,
    optimizing resource allocation, and managing healthcare
    operations.
    • - Finance: Portfolio optimization, trading strategies, and risk
    management.
    • -Self-Driving Cars: Decision-making in complex environments,
    like lane changing, braking, and accelerating.

    You might also like