Along with unsupervised machine learning and supervised learning, another common form of AI creation is reinforcement learning. Beyond regular reinforcement learning, deep reinforcement learning can lead to astonishingly impressive results, thanks to the fact that it combines the best aspects of both deep learning and reinforcement learning. Lets take a look at precisely how deep reinforcement learning operates. Note that this article wont delve too deeply into the formulas used in deep reinforcement learning, rather it aims to give the reader a high level intution for how the process works.
Before we dive into deep reinforcement learning, it might be a good idea to refresh ourselves on how regular reinforcement learning works. In reinforcement learning, goal-oriented algorithms are designed through a process of trial and error, optimizing for the action that leads to the best result/the action that gains the most reward. When reinforcement learning algorithms are trained, they are given rewards or punishments that influence which actions they will take in the future. Algorithms try to find a set of actions that will provide the system with the most reward, balancing both immediate and future rewards.
Reinforcement learning algorithms are very powerful because they can be applied to almost any task, being able to flexibly and dynamically learn from an environment and discover possible actions.
Photo: Megajuice via Wikimedia Commons, CC 1.0 (https://commons.wikimedia.org/wiki/File:Reinforcement_learning_diagram.svg)
When it comes to deep reinforcement learning, the environment is typically represented with images. An image is a capture of the environment at a particular point in time. The agent must analyze the images and extract relevant information from them, using the information to inform which action they should take. Deep reinforcement learning is typically carried out with one of two different techniques: value-based learning and policy-based learning.
Value-based learning techniques make use of algorithms and architectures like convolutional neural networks and Deep-Q-Networks. These algorithms operate by converting the image to greyscale and cropping out unnecessary parts of the image. Afterward, the image undergoes various convolutions and pooling operations, extracting the most relevant portions of the image. The important parts of the image are then used to calculate the Q-value for the different actions the agent can take. Q-values are used to determine the best course of action for the agent. After the initial Q-values are calculated, backpropagation is carried out in order that the most accurate Q-values can be determined.
Policy-based methods are used when the number of possible actions that the agent can take is extremely high, which is typically the case in real-world scenarios. Situations like these require a different approach because calculating the Q-values for all the individual actions isnt pragmatic. Policy-based approaches operate without calculating function values for individual actions. Instead, they adopt policies by learning the policy directly, often through techniques called Policy Gradients.
Policy gradients operate by receiving a state and calculating probabilities for actions based on the agents prior experiences. The most probable action is then selected. This process is repeated until the end of the evaluation period and the rewards are given to the agent. After the rewards have been dealt with the agent, the networks parameters are updated with backpropagation.
Because Q-Learning is such a large part of the deep reinforcement learning process, lets take some time to really understand how the Q-learning system works.
The Markov Decision Process
A markov decision process. Photo: waldoalvarez via Pixabay, Pixbay License (https://commons.wikimedia.org/wiki/File:Markov_Decision_Process.svg)
In order for an AI agent to carry out a series of tasks and reach a goal, the agent must be able to deal with a sequence of states and events. The agent will begin at one state and it must take a series of actions to reach an end state, and there can be a massive number of states existing between the beginning and end states. Storing information regarding every state is impractical or impossible, so the system must find a way to preserve just the most relevant state information. This is accomplished through the use of a Markov Decision Process, which preserves just the information regarding the current state and the previous state. Every state follows a Markov property, which tracks how the agent change from the previous state to the current state.
Deep Q-Learning
Once the model has access to information about the states of the learning environment, Q-values can be calculated. The Q-values are the total reward given to the agent at the end of a sequence of actions.
The Q-values are calculated with a series of rewards. There is an immediate reward, calculated at the current state and depending on the current action. The Q-value for the subsequent state is also calculated, along with the Q-value for the state after that, and so on until all the Q-values for the different states have been calculated. There is also a Gamma parameter that is used to control how much weight future rewards have on the agents actions. Policies are typically calculated by randomly initializing Q-values and letting the model converge toward the optimal Q-values over the course of training.
Deep Q-Networks
One of the fundamental problems involving the use of Q-learning for reinforcement learning is that the amount of memory required to store data rapidly expands as the number of states increases. Deep Q Networks solve this problem by combining neural network models with Q-values, enabling an agent to learn from experience and make reasonable guesses about the best actions to take. With deep Q-learning, the Q-value functions are estimated with neural networks. The neural network takes the state in as the input data, and the network outputs Q-value for all the different possible actions the agent might take.
Deep Q-learning is accomplished by storing all the past experiences in memory, calculating maximum outputs for the Q-network, and then using a loss function to calculate the difference between current values and the theoretical highest possible values.
Deep Reinforcement Learning vs Deep Learning
One important difference between deep reinforcement learning and regular deep learning is that in the case of the former the inputs are constantly changing, which isnt the case in traditional deep learning. How can the learning model account for inputs and outputs that are constantly shifting?
Essentially, to account for the divergence between predicted values and target values, two neural networks can be used instead of one. One network estimates the target values, while the other network is responsible for the predictions. The parameters of the target network are updated as the model learns, after a chosen number of training iterations have passed. The outputs of the respective networks are then joined together to determine the difference.
Policy-based learning approaches operate differently than Q-value based approaches. While Q-value approaches create a value function that predicts rewards for states and actions, policy-based methods determine a policy that will map states to actions. In other words, the policy function that selects for actions is directly optimized without regard to the value function.
Policy Gradients
A policy for deep reinforcement learning falls into one of two categories: stochastic or deterministic. A deterministic policy is one where states are mapped to actions, meaning that when the policy is given information about a state an action is returned. Meanwhile, stochastic policies return a probability distribution for actions instead of a single, discrete action.
Deterministic policies are used when there is no uncertainty about the outcomes of the actions that can be taken. In other words, when the environment itself is deterministic. In contrast, stochastic policy outputs are appropriate for environments where the outcome of actions is uncertain. Typically, reinforcement learning scenarios involve some degree of uncertainty so stochastic policies are used.
Policy gradient approaches have a few advantages over Q-learning approaches, as well as some disadvantages. In terms of advantages, policy-based methods converge on optimal parameters quicker and more reliably. The policy gradient can just be followed until the best parameters are determined, whereas with value-based methods small changes in estimated action values can lead to large changes in actions and their associated parameters.
Policy gradients work better for high dimensional action spaces as well. When there is an extremely high number of possible actions to take, deep Q-learning becomes impractical because it must assign a score to every possible action for all time steps, which may be impossible computationally. However, with policy-based methods, the parameters are adjusted over time and the number of possible best parameters quickly shrinks as the model converges.
Policy gradients are also capable of implementing stochastic policies, unlike value-based policies. Because stochastic policies produce a probability distribution, an exploration/exploitation trade-off does not need to be implemented.
In terms of disadvantages, the main disadvantage of policy gradients is that they can get stuck while searching for optimal parameters, focusing only on a narrow, local set of optimum values instead of the global optimum values.
Policy Score Function
The policies used to optimize a models performance aim to maximize a score function J(). If J() is a measure of how good our policy is for achieving the desired goal, we can find the values of that gives us the best policy. First, we need to calculate an expected policy reward. We estimate the policy reward so we have an objective, something to optimize towards. The Policy Score Function is how we calculate the expected policy reward, and there are different Policy Score Functions that are commonly used, such as: start values for episodic environments, the average value for continuous environments, and the average reward per time step.
Policy Gradient Ascent
Gradient ascent aims to move the parameters until they are at the place where the score is highest. Photo: Public Domain (https://commons.wikimedia.org/wiki/File:Gradient_ascent_(surface).png)
After the desired Policy Score Function is used, and an expected policy reward calculated, we can find a value for the parameter which maximizes the score function. In order to maximize the score function J(), a technique called gradient ascent is used. Gradient ascent is similar in concept to gradient descent in deep learning, but we are optimizing for the steepest increase instead of decrease. This is because our score is not error, like in many deep learning problems. Our score is something we want to maximize. An expression called the Policy Gradient Theorem is used to estimate the gradient with respect to policy .
In summary, deep reinforcement learning combines aspects of reinforcement learning and deep neural networks. Deep reinforcement learning is done with two different techniques: Deep Q-learning and policy gradients.
Deep Q-learning methods aim to predict which rewards will follow certain actions taken in a given state, while policy gradient approaches aim to optimize the action space, predicting the actions themselves. Policy-based approaches to deep reinforcement learning are either deterministic or stochastic in nature. Deterministic policies map states directly to actions while stochastic policies produce probability distributions for actions.
See the article here:
AI Used to Monitor Health of Coral Reefs and Detect Ocean Trash Pollution - Unite.AI
- Chinese national arrested and charged with stealing AI trade secrets from Google - NPR - March 8th, 2024 [March 8th, 2024]
- President Biden Calls for Ban on AI Voice Impersonations During State of the Union - Variety - March 8th, 2024 [March 8th, 2024]
- Revolutionize Your Business with AWS Generative AI Competency Partners | Amazon Web Services - AWS Blog - March 8th, 2024 [March 8th, 2024]
- Broadcom Expects AI Demand to Help Offset Weakness Elsewhere - Yahoo Finance - March 8th, 2024 [March 8th, 2024]
- Micron Hits Record High With Analysts Calling It an 'Under-Appreciated AI Beneficiary' - Investopedia - March 8th, 2024 [March 8th, 2024]
- The Adams administration quietly hired its first AI czar. Who is he? - City & State New York - March 8th, 2024 [March 8th, 2024]
- AI likely to increase energy use and accelerate climate misinformation report - The Guardian - March 8th, 2024 [March 8th, 2024]
- This Artificial Intelligence (AI) Stock Could Double, and It Is Way Cheaper Than Nvidia - Yahoo Finance - March 8th, 2024 [March 8th, 2024]
- Fake images made to show Trump with Black supporters highlight concerns around AI and elections - The Associated Press - March 8th, 2024 [March 8th, 2024]
- Artificial intelligence and illusions of understanding in scientific research - Nature.com - March 8th, 2024 [March 8th, 2024]
- Analysis | House AI task force leaders take long view on regulating the tools - The Washington Post - March 8th, 2024 [March 8th, 2024]
- Don't Give Your Business Data to AI Companies - Dark Reading - March 8th, 2024 [March 8th, 2024]
- NIST, the lab at the center of Bidens AI safety push, is decaying - The Washington Post - March 8th, 2024 [March 8th, 2024]
- Essay | AI is Coming! Tips for Staying Calm and Carrying On - The Wall Street Journal - March 8th, 2024 [March 8th, 2024]
- AI can be easily used to make fake election photos - report - BBC.com - March 8th, 2024 [March 8th, 2024]
- 5 Artificial Intelligence (AI) Stocks That Could Make You a Millionaire - Yahoo Finance - March 8th, 2024 [March 8th, 2024]
- AI could be an extraordinary force for good. So why do our politicians still not have a plan? - The Guardian - March 8th, 2024 [March 8th, 2024]
- Mapping Disease Trajectories from Birth to Death with AI - Neuroscience News - March 8th, 2024 [March 8th, 2024]
- India plans 10,000-GPU sovereign AI supercomputer - The Register - March 8th, 2024 [March 8th, 2024]
- SAP enhances Datasphere and SAC for AI-driven transformation - CIO - March 8th, 2024 [March 8th, 2024]
- Jim Cramer names companies and sectors poised to rally on the AI wave - CNBC - March 8th, 2024 [March 8th, 2024]
- The job applicants shut out by AI: The interviewer sounded like Siri - The Guardian - March 8th, 2024 [March 8th, 2024]
- Microsoft confirms Surface and Windows AI event for March 21st - The Verge - March 8th, 2024 [March 8th, 2024]
- Adobes new Express app brings Firefly AI tools to iOS and Android - The Verge - March 8th, 2024 [March 8th, 2024]
- A Google AI Watched 30,000 Hours of Video GamesNow It Makes Its Own - Singularity Hub - March 8th, 2024 [March 8th, 2024]
- Palantir CEO Karp on TITAN, AI Warfare Technology - Bloomberg - March 8th, 2024 [March 8th, 2024]
- Elliptic Curve Murmurations Found With AI Take Flight - Quanta Magazine - March 8th, 2024 [March 8th, 2024]
- 5 AI Stocks to Buy in March 2024, According to Analysts - TipRanks.com - TipRanks - March 8th, 2024 [March 8th, 2024]
- Wix's new AI chatbot builds websites in seconds based on prompts - The Verge - March 8th, 2024 [March 8th, 2024]
- Amid record high energy demand, America is running out of electricity - The Washington Post - March 8th, 2024 [March 8th, 2024]
- AI Crypto Tokens in 5 Minutes: What to Know and Where to Start - Inc. - February 26th, 2024 [February 26th, 2024]
- 'The Worlds I See' by AI visionary Fei-Fei Li '99 selected as Princeton Pre-read - Princeton University - February 26th, 2024 [February 26th, 2024]
- AI is having a 1995 moment, analyst says - Business Insider - February 26th, 2024 [February 26th, 2024]
- Vatican research group's book outlines AI's 'brave new world' - National Catholic Reporter - February 26th, 2024 [February 26th, 2024]
- Honor's Magic 6 Pro launches internationally with AI-powered eye tracking on the way - The Verge - February 26th, 2024 [February 26th, 2024]
- Google explains Gemini's embarrassing AI pictures of diverse Nazis - The Verge - February 26th, 2024 [February 26th, 2024]
- Google cut a deal with Reddit for AI training data - The Verge - February 26th, 2024 [February 26th, 2024]
- What's the point of Elon Musk's AI company? - The Verge - February 26th, 2024 [February 26th, 2024]
- AI agents like Rabbit aim to book your vacation and order your Uber - NPR - February 26th, 2024 [February 26th, 2024]
- Announcing Microsofts open automation framework to red team generative AI Systems - Microsoft - February 26th, 2024 [February 26th, 2024]
- After Nvidia's latest blowout, here are 20 AI stocks expected to rise as much as 44% - Yahoo Finance - February 26th, 2024 [February 26th, 2024]
- 1 Exceptional AI Chip Stock Investors Need to Know About in 2024 - The Motley Fool - February 26th, 2024 [February 26th, 2024]
- Nvidia briefly hits $2 trillion valuation as AI frenzy grips Wall Street - Reuters - February 26th, 2024 [February 26th, 2024]
- AI Chatbots Can Guess Your Personal Information From What You ... - WIRED - October 18th, 2023 [October 18th, 2023]
- Harvard IT Launches Pilot of AI Sandbox to Enable Walled-Off Use ... - Harvard Crimson - October 18th, 2023 [October 18th, 2023]
- Advancing policing through AI: Insights from the global law ... - Police News - October 18th, 2023 [October 18th, 2023]
- Hochul announces new SUNY, IBM investments in AI - Olean Times Herald - October 18th, 2023 [October 18th, 2023]
- Nvidia's banking on TensorRT to expand its generative AI dominance - The Verge - October 18th, 2023 [October 18th, 2023]
- AI expands from MRFs to vehicles - Plastics Recycling Update - October 18th, 2023 [October 18th, 2023]
- AI Reads Ancient Scroll Charred by Mount Vesuvius in Tech First - Scientific American - October 18th, 2023 [October 18th, 2023]
- A DEEPer (squared) dive into AI Harvard Gazette - Harvard Gazette - October 18th, 2023 [October 18th, 2023]
- Florida bar weighs whether lawyers using AI need client consent - Reuters - October 18th, 2023 [October 18th, 2023]
- Cognizant and Vianai Systems Announce Strategic Partnership to ... - PR Newswire - October 18th, 2023 [October 18th, 2023]
- How AI could speed up scientific discoveries, from proteins to ... - NPR - October 18th, 2023 [October 18th, 2023]
- AI challenge to deliver better healthcare | Western Australian ... - Government of Western Australia - October 18th, 2023 [October 18th, 2023]
- Henry Kissinger: The Path to AI Arms Control - Foreign Affairs Magazine - October 18th, 2023 [October 18th, 2023]
- Stability AI releases StableStudio in latest push for open-source AI - The Verge - May 18th, 2023 [May 18th, 2023]
- Google CEO Sundar Pichai Predicts That This Profession Will Be ... - The Motley Fool - May 18th, 2023 [May 18th, 2023]
- Frances privacy watchdog eyes protection against data scraping in AI action plan - TechCrunch - May 18th, 2023 [May 18th, 2023]
- Investing in Hippocratic AI - Andreessen Horowitz - May 18th, 2023 [May 18th, 2023]
- As Alphabet flexes its AI prowess, there's a 'new elephant in the room' for Google - MarketWatch - May 18th, 2023 [May 18th, 2023]
- The Boring Future of Generative AI | WIRED - WIRED - May 18th, 2023 [May 18th, 2023]
- OpenAI readies new open-source AI model, The Information reports - Reuters.com - May 18th, 2023 [May 18th, 2023]
- What every CEO should know about generative AI - McKinsey - May 18th, 2023 [May 18th, 2023]
- AI creates images of the 'perfect' man and woman - Sky News - May 18th, 2023 [May 18th, 2023]
- Audit AI search tools now, before they skew research - Nature.com - May 18th, 2023 [May 18th, 2023]
- 3 Reasons C3.ai Stock Could Be Your Golden Ticket to the AI ... - InvestorPlace - May 18th, 2023 [May 18th, 2023]
- Zoom makes a big bet on AI with investment in Anthropic - VentureBeat - May 18th, 2023 [May 18th, 2023]
- AI voice phone scams are on the rise. Here's how to avoid them - USA TODAY - May 18th, 2023 [May 18th, 2023]
- Amazon is building an AI-powered conversational experience for ... - The Verge - May 18th, 2023 [May 18th, 2023]
- AI speculators need to 'differentiate between actual spending and investment' and hype: Strategist - Yahoo Finance - May 18th, 2023 [May 18th, 2023]
- AI Can Be Both Accurate and Transparent - HBR.org Daily - May 18th, 2023 [May 18th, 2023]
- You're Probably Underestimating AI Chatbots | WIRED - WIRED - May 18th, 2023 [May 18th, 2023]
- AI presents political peril for 2024 with threat to mislead voters - The Associated Press - May 18th, 2023 [May 18th, 2023]
- We need AI to help us face the challenges of the future - The Guardian - May 18th, 2023 [May 18th, 2023]
- End Of Googles Dominance? Stock Gets Rare Analyst Downgrade Over AI Fears - Forbes - May 18th, 2023 [May 18th, 2023]
- Watch 44 million atoms simulated using AI and a supercomputer - New Scientist - May 18th, 2023 [May 18th, 2023]
- AI Is The New Electricity: Bank Of America Picks 20 Stocks To Cash In On ChatGPT Hype - Forbes - March 2nd, 2023 [March 2nd, 2023]
- Tech Giants Are Barreling Headfirst Into an AI Arms Race - February 20th, 2023 [February 20th, 2023]
- Bing's AI Is Threatening Users. That's No Laughing Matter - TIME - February 20th, 2023 [February 20th, 2023]