Questions tagged [deep-rl]

For questions related to deep reinforcement learning (DRL), that is, RL combined with deep learning. More precisely, deep neural networks are used to represent e.g. value functions or policies.

164 questions with no upvoted or accepted answers
Filter by
Sorted by
Tagged with
7 votes
0 answers
1k views

Is there a difference in the architecture of deep reinforcement learning when multiple actions are performed instead of a single action?

I've built a deep deterministic policy gradient reinforcement learning agent to be able to handle any games/tasks that have only one action. However, the agent seems to fail horribly when there are ...
user avatar
  • 403
5 votes
0 answers
227 views

How to correctly implement self-play with DQN?

I have an environment where an agent faces an equal opponent, and while I've achieved OK performance implementing DQN and treating the opponent as a part of the environment, I think performance would ...
user avatar
  • 51
4 votes
0 answers
696 views

Why isn't a target network used for the critic in on-policy actor-critic methods?

Based on my research, I've seen so many on-policy AC approaches that utilise a critic network to estimate the value function $V$. The Bellman equation for the value function is as bellow: $$ V_\pi(s_t)...
user avatar
4 votes
0 answers
86 views

Where does this variation of the importance sampling weight come from?

I have seeing a variation in importance sampling (IS) in Prioritized Experience Replay (PER) in some implementations regarding the original paper approach stated as (in section 3.4): $$ w_{i}=\left(\...
user avatar
  • 81
4 votes
1 answer
140 views

How does the optimization process in hindsight experience replay exactly work?

I was reading the following research paper Hindsight Experience Replay. This is the paper that introduces a concept called Hindsight Experience Replay (HER), which basically attempts to alleviate the ...
user avatar
4 votes
0 answers
50 views

How to deal with nonstationary rewards in asymmetric self-play reinforcement learning?

Suppose we're training two agents to play an asymmetric game from scratch using self play (like Zerg vs. Protoss in Starcraft). During training one of the agents can become stronger (discover a good ...
user avatar
  • 141
4 votes
1 answer
200 views

What could be the cause of the drop in the reward in A3C?

The mean episodic reward is generally increasing, but it has spontaneous drops, and I'm not sure of their cause. The problem has a sparse reward, batch size=2000, <...
user avatar
  • 163
4 votes
0 answers
596 views

What could be causing the drastic performance drop of the DQN model on the Pong environment?

I am running a basic DQN (Deep Q-Network) on the Pong environment. Not a CNN, just a 3 layer linear neural net with ReLUs. It seems to work for the most part, but at some point, my model suffers from ...
user avatar
  • 141
4 votes
1 answer
252 views

How did the OpenAI 5 for Dota concatenate units?

I am no expert in the field of AI so I apologize if this is a simple/easy question. I was trying to implement a network similar to OpenAI's for another game and I noticed that I did not fully ...
user avatar
4 votes
1 answer
2k views

How to deal with different actions for different states of the environment?

I'm new to this AI/Machine Learning and was playing around with OpenAI Gym a bit. When looking through the environments, I came across the Blackjack-v0 environment, ...
user avatar
3 votes
0 answers
48 views

How to prove Lemma 1.6 in the book "Reinforcement Learning: Theory and Algorithms"

I am trying to prove the following lemma from Reinforcement Learning: Theory and Algorithms on page 8. Lemma 1.6. We have that: $$ \left[(1-\gamma)\left(I-\gamma P^{\pi}\right)^{-1}\right]_{(s, a),\...
user avatar
3 votes
0 answers
166 views

How to deal with a moving target in the Lunar Lander environment with DDPG?

I have noticed that DDPG does rather well at solving environments with a static target. For example, the default of Lunar Lander, the flags do not change position. So the DDPG model learns how to get ...
user avatar
3 votes
1 answer
114 views

How can I compress the states of a reinforcement learning agent?

I'm working on a problem that involves an RL agent with very large states. These states consist of several pieces of information about the agent. The states are not images, so techniques like ...
user avatar
3 votes
0 answers
193 views

Stack of Planes as the Action Space Representation for AlphaZero (Chess)

I have a question regarding the action space of the policy network used in AlphaZero. From the paper: We represent the policy π(a|s) by a 8 × 8 × 73 stack of planes encoding a probability ...
user avatar
  • 107
3 votes
0 answers
88 views

How to take actions at each episode and within each step of the episode in deep Q learning?

In deep Q learning, we execute the algorithm for each episode, and for each step within an episode, we take an action and record a reward. I have a situation where my action is 2-tuple $a=(a_1,a_2)$. ...
user avatar
  • 247
3 votes
0 answers
131 views

Does the concept of validation loss apply to training deep Q networks?

In deep learning, the concept of validation loss is to ensure that the model being trained is not currently overfitting the data. Is there a similar concept of overfitting in deep q learning? Given ...
user avatar
  • 1,211
3 votes
0 answers
751 views

Why is it hard to prove the convergence of the deep Q-learning algorithm?

Why is it hard to prove the convergence of the DQN algorithm? We know that the tabular Q-learning algorithm converges to the optimal Q-values, and with a linear approximator convergence is proved. ...
user avatar
3 votes
0 answers
92 views

Understanding the TensorFlow implementation of the policy gradient method

I was trying to understand the implementation of a basic policy gradient (REINFORCE) method using TensorFlow. I think I got almost everything. The only thing that still bothers me is the loss function ...
user avatar
  • 31
3 votes
0 answers
103 views

How does normalization of the inputs work in the context of PPO?

What does the normalization of the inputs mean in the context of PPO? At each time step of an episode, I only know the values of this time step and of the previous ones, if I take track of them. This ...
user avatar
  • 151
3 votes
0 answers
1k views

How much time does it take to train DQN on Atari environment?

I am trying to build a DQN model for the Atari Pong game, but I am not sure whether the model is learning at all. I am using the architecture described in the paper Playing Atari with Deep ...
user avatar
  • 151
3 votes
0 answers
89 views

Is this a good approach to solving Atari's "Montezuma's Revenge"?

I'm new to Reinforcement Learning. For an internship, I am currently training Atari's "Montezuma's Revenge" using a double Deep Q-Network with Hindsight Experience Replay (HER) (see also ...
user avatar
3 votes
0 answers
201 views

Deep Q-Network (DQN) to learn the game 2048

I am trying to build a Deep Q-Network (DQN) agent that can learn to play the game 2048. I am orientating myself on other programs and articles that are based on the game snake and it worked well (...
user avatar
  • 31
3 votes
0 answers
34 views

Training a reinforcement learning model with multiple images

I am tentatively trying to train a deep reinforcement learning model the maze escaping task, and each time it takes one image as the input (e.g., a different "maze"). Suppose I have about $10K$ ...
user avatar
3 votes
0 answers
284 views

DQN Agent not learning anymore - what can I do to fix this?

I am trying to use Deep-Q-Learning to learn an ANN which controls a 7-DOF robotic arm. The robotic arm must avoid an obstacle and reach a target. I have implemented a number of state-of-art ...
user avatar
  • 715
3 votes
0 answers
57 views

Deep Q-Learning agent poor performing actions. Need help optimizing

I'm trying to make deep q-learning agent from https://keon.io/deep-q-learning My environment looks like this: https://imgur.com/a/OnbiCtV As you can see my agent is a circle and there is one gray ...
user avatar
  • 31
3 votes
0 answers
205 views

What can be considered a deep recurrent neural network?

In the paper Deep Recurrent Q-Learning for Partially Observable MDPs, the DRQN is described as DQN with the first post-convolutional fully-connected layer replaced by a recurrent LSTM. I have DQN ...
user avatar
  • 61
2 votes
0 answers
26 views

Is Reinforcement Learning with only feedback on a single action possible?

Consider the following case: A reinforcement based web-crawler where: State = current page + 1 out-link (reduced to features of some sort) Action = Whether to visit that out-link or not (n_actions = ...
user avatar
  • 175
2 votes
0 answers
55 views

How to reduce the dimensionality of the actions in RL

I have a single-agent RL model in which the dimension of the dimension of the action space is $70$. This action space is too big and the deep RL agent is not learning properly. The boundaries of the ...
user avatar
  • 49
2 votes
0 answers
36 views

How to deal with variable action ranges in RL for continuous action spaces

I am reading this paper on battery management using RL. The action consist in the charging/discharging power of the battery at timestep $t$. For instance, in the case of the charging power, the ...
user avatar
  • 49
2 votes
1 answer
43 views

How to embed game grid state with walls as an input to neural network

I've read most of the posts on here regarding this subject, however most of them deal with gameboards where there are two different categories of single pieces on a board without walls etc. My game ...
user avatar
2 votes
0 answers
355 views

How to interpret the training loss curves in Soft-Actor-Critic (SAC)?

I am using stable-baseline3 implementation of the Soft-Actor-Critic (SAC) algorithm. The plotted training curves look promising. However, I am not fully sure how to interpret the actor and critic ...
user avatar
  • 35
2 votes
1 answer
105 views

How do I represent sample efficiency of RL rewards in mathematical notation?

I define sample efficiency as the area under the curve/graph, where $x$-axis is the number of episodes while y-axis is the cumulative reward for that episode. I would like to formally define it with a ...
user avatar
2 votes
0 answers
780 views

Optimal episode length in reinforcement learning

I have a custom environment for stock trading where an episode can be as long as 2000-3000 steps. I've run several experiments with td3 and sac algorithms, average reward per episode flattens after ...
user avatar
  • 167
2 votes
1 answer
211 views

How does sharing parameters between the policy and value functions help in PPO?

The PPO objective may include a value function error term when parameters are shared between the policy and value functions. How does this help, and when to use a neural network architecture that ...
user avatar
  • 167
2 votes
0 answers
227 views

Update Rule with Deep Q-Learning (DQN) for 2-player games

I am wondering how to correctly implement the DQN algorithm for two-player games such as Tic Tac Toe and Connect 4. While my algorithm is mastering Tic Tac Toe relatively quickly, I cannot get great ...
user avatar
  • 31
2 votes
0 answers
31 views

What is the name of this algorithm that estimates the gradient with an average by sampling from a distribution?

Consider maximizing the function $R(w)$ with parameter $w$ using gradient ascent. However, we don't know the gradient $\nabla_wR(w)$ formula. Now suppose $w$ is sampled from a probability distribution ...
user avatar
  • 121
2 votes
0 answers
264 views

Is (log-)standard deviation learned in TRPO and PPO or fixed instead?

After having read Williams (1992), where it was suggested that actually both the mean and standard deviation can be learned while training a REINFORCE algorithm on generating continuous output values, ...
user avatar
  • 705
2 votes
0 answers
23 views

Can DQN outperform DoubleDQN?

I found a similar post about this issue, but unfortunately I did not find a proper answer. Are there any references where DQN is better than DoubleDQN, that is DoubleDQN does not improve DQN ?
user avatar
  • 889
2 votes
0 answers
148 views

Policy gradient: Does it use the Markov property?

To derive the policy gradient, we start by writing the equation for the probability of a certain trajectory (e.g. see spinningup tutorial): $$ \begin{align} P_\theta(\tau) &= P_\theta(s_0, a_0, ...
user avatar
  • 121
2 votes
0 answers
75 views

Is Deep SARSA learning a feasible approach?

I noticed that SARSA has been rarely used in the deep RL setting. Usually, the training for DQN is done off-policy. I think one of the major reasons for this is due to greater sample efficiency in ...
user avatar
  • 1,211
2 votes
0 answers
56 views

How to combine two differently equally important signals into the reward function, that have different scales?

I have two signals that I want to use to model my reward. The first one is the CPU TIME: running mean from this diagram: The second one is the MAX RESIDUAL from this diagram: Since they are both ...
user avatar
  • 362
2 votes
0 answers
161 views

If the minimum Q value is decreasing and the maximum Q value increasing, is this a sign that dueling double DQN is diverging?

I'm training a dueling double DQN agent with prioritized replay buffer and notice that the min Q values are decreasing, while the max Q values are increasing. Is this a sign that it is diverging? ...
user avatar
2 votes
1 answer
197 views

How to implement RAM versions of Atari games

I have coded the breakout RAM version, but, unfortunately, its highest reward was 5. I trained it for about 2 hours and never reached a higher score. The code is huge, so I can't paste here, but, in ...
user avatar
2 votes
0 answers
79 views

How can the target rely on untrained parameters?

I'm trying to understand DQN. I understand where the loss function comes from. I'm just unsure about why the target function works in practice. Given the loss function $$ L_i(\theta_i) = [(y_i - Q(s,a;...
user avatar
2 votes
0 answers
84 views

Are the final states not being updated in this $n$-step Q-Learning algorithm?

I am reading this paper and in algorithm 3 they describe an $n$-step Q-Learning algorithm. Below is the pseudo-code. From this pseudo-code, it looks as though the final tuples that they would ...
user avatar
2 votes
0 answers
122 views

How to prevent deep Q-learning algorithms to overfit?

I have recently solved the Cartpole problem using double deep Q-learning. When I saw how the agent was doing, it used to go right every time, never left, and it did similar actions all the time. Did ...
user avatar
2 votes
0 answers
58 views

Can you find another reason for sample inefficiency of model-free on-policy Deep Reinforcement Learning?

The following mindmap gives an overview of multiple reasons for sample inefficiency. The list is definitely not complete. Can you see another reason not mentioned so far? Some related links: ...
user avatar
2 votes
0 answers
132 views

NaNs after a while in training of PPO

My problem is that every time I am trying to train my PPO agent I get NaN values after a while. The diagnostic that I get is the following: ...
user avatar
  • 151
2 votes
0 answers
35 views

How was the DQN trained to play many games?

Some people claim that DQN was used to play many Atari games. But what actually happened? Was DQN trained only once (with some data from all games) or was it trained separately for each game? What was ...
user avatar
  • 573
2 votes
0 answers
323 views

How does adding noise to the action in DDPG help in learning?

I can't understand how playing with the action generated by the actor network in DDPG by adding the noise term helps in exploration.
user avatar