policy gradients

CellStrat > Research/Blog > policy gradients

Jun

#CellStratAILab #disrupt4.0 #WeCreateAISuperstars I presented a session on Multi-Agent RL recently at the CellStrat AI Lab. Introduction :- In the normal Reinforcement Learning setup, you have one agent which interacts with the environment. It uses the Observation from the environment, performs actions and observes the rewards. In real life, many applications will involve several agents […]

Posted in: Reinforcement Learning,

Tags: Actor Critic, DDPG, Deep Deterministic Policy Gradients, MADDPG, Markov Games, Multi Agent Actor Critic, Multi Agent DDPG, Multi Agent RL, Policy Ensembles, policy gradients, Policy Inference, reinforcement learning, Unity ML,

Apr

DDPG and TD3

This post assumes that you have a strong understanding of the basics of Reinforcement Learning, MDP, DQN and Policy Gradient Algorithms. You can go through Policy Gradients to understand the derivation for Stochastic Policies In the previous post on Actor Critic, we saw the advantage of merging Value based and Policy based methods together. The […]

Posted in: Reinforcement Learning, Robotics,

Tags: Actor Critic, Actor Critic method, Bellman equation, DDPG, Deep Deterministic Policy Gradients, deterministic policy, Double DQN, DQN, Experience Replay, exploration vs exploitation, Fixed D Targets, policy gradients, Q learning, TD3 RL, Twin Delayed Double Deterministic Policy Gradients,

Apr

A Summary of Model-Free RL Algorithms

#CellStratAILab #disrupt4.0 #WeCreateAISuperstars #AlwaysUpskilling Reinforcement Learning (RL) refers to training agents with help of incentive-driven environments. RL typically involves a tuple of <state, action, reward> paradigm, which means that the agent has action choices to make in various states, and each action entails a potential reward. This also means that each state has a “value” […]

Posted in: Reinforcement Learning, Robotics,

Tags: Actor Critic, Actor Critic method, AI lab, cartpole, DDPG, deep Q learning, deep Q network, Deep Reinforcement Learning, deterministic policy, DQN, DRL course, gaming, markov decision process, markov process, Markov Reward Process, mdp, model-based RL, model-free RL, monte carlo, off-policy, on-policy, policy based methods, policy gradients, PPO, Q learning, Q table, Rainbow method, Reinforcement learning course, reward function, RL course, RL for gaming, RL training, SAC, stochastic policy, TD Learning, TD3, training in reinforcement learning, TRPO, Value-based methods,

Mar

RL with Actor-Critic Methods

#CellStratAILab #disrupt4.0 #WeCreateAISuperstars #AlwaysUpskilling Minutes from Saturday 14th March 2020 AI Lab Workshop at BLR :- Session Presenter : SHUBHA M., Deep Reinforcement Learning Researcher, CellStrat AI Lab Last Saturday, our Reinforcement Learning Team Lead Shubha M. presented a fantastic presentation and workshop on Actor-Critic method used in RL. She also demonstrated a demo of this technique for Stock Market predictions. […]

Posted in: Reinforcement Learning, Robotics,

Tags: Actor Critic, deep Q learning, Deep Reinforcement Learning, DRL, markov decision process, Markov Reward Process, mdp, MRP, policy based methods, policy gradients, Q Actor Critic, Q learning, REINFORCE algorithm, reinforcement learning, Value-based methods,

Mar

Face Recognition with MTCNN and FaceNet; RL with Proximal Policy Optimization

#CellStratAILab #disrupt4.0 #WeCreateAISuperstars #AlwaysUpskilling Minutes from Saturday 7th March 2020 AI Lab meetup at BLR :- Last Saturday, we had excellent sessions in the AI Lab meetup. Face Recognition with MTCNN and FaceNet :- First Amit Kumar presented a detailed overview of Face Recognition with MTCNN and FaceNet. Face Recognition involves a pipeline of Face […]

Posted in: Computer Vision, Deep Learning, Reinforcement Learning, Security,

Tags: Advantage Function, Face Recognition, facenet, Important Sampling, MTCNN, MTCNN face detector, O-Net, PG algo, policy based methods, policy gradients, PPO, proximal policy optimization, R-Net, reinforcement learning, Triplet Loss, Value-based methods, Vanilla Policy Gradient, VPG,

Mar

Proximal Policy Optimization

In my previous post, we discussed the simplest Policy Gradient REINFORCE. We saw, how Policy based methods are better than value based methods, a derivation of the Gradient of Score(Cost) function, and an implementation of simple Policy Gradient to train Gym’s Acrobot-v0. We then saw, how introducing a baseline reduces variance which leads to the […]

Posted in: Reinforcement Learning,

Tags: clipped PPO, Importance Sampling, PG, policy gradients, PPO, proximal policy optimization, reinforcement learning, RL, TRPO, trust region policy optimization,

Feb

NLP Embeddings with GLoVe, RL Policy Gradients, Quantum Neural Networks

#CellStratAILab #disrupt4.0 #WeCreateAISuperstars Last Saturday (1st Feb 2020), our AI Lab researchers delivered superb presentations on advanced topics. First Shreyas S K started with detailed algorithmic discussion with GLoVe and T-SNE algos. Then he showed how these can be used for Sarcasm Detection in news articles. Detection of sarcasm is of great importance and beneficial […]

Posted in: Natural Language Processing, Quantum Computing, Quantum Machine Learning, Reinforcement Learning,

Tags: GLoVe, policy gradients, qGAN, quantum neural network, reinforcement learning, sarcasm detection, t-sne, TFIDF, word embedding,

Feb

Policy Gradients – An Introduction

I conducted an Introductory session on Reinforcement Learning Policy Gradients (PG) at CellStrat AI Lab on 1st Feb 2020. The goal of this session was to explain the basic underlying principle of Policy Gradients. The session started off with a quick recap of Reinforcement Learning, so that the audience is well aware of the definitions […]

Posted in: Reinforcement Learning,

Tags: Acrobot, deep Q network, deepmind, DQN, experience replay buffer, fixed Q targets, gradient ascent, openAI Gym, policy based methods, policy gradients, Q learning, reinforcement learning, reward function, rewards based learning, value iteration algorithm,