Questions tagged [reinforcement-learning]

For questions related to learning controlled by external positive reinforcement or negative feedback signal or both, where learning and use of what has been thus far learned occur concurrently.

Filter by
Sorted by
Tagged with
4
votes
1answer
45 views

How does the repetition of features across states at different time steps affect learning?

Let's say you are training a neural network in an RL setting, where the state (i.e. features/input data) can be the same for multiple successive steps (~typically around 8 steps) of an episode. For ...
1
vote
1answer
31 views

How do you manage negative rewards in policy gradient reinforcement learning?

The same basic question here, but 3 years old and no definitive answer: Negative reward (penalty) in policy gradient reinforcement learning The question is, if I'm doing policy gradient in keras, ...
3
votes
1answer
59 views

What is the difference between on-policy and off-policy for continuous environments?

I'm trying to understand RL applied to time series (so with infinite horizon) which have a continous state space and a discrete action space. First, some preliminary questions: in this case, what is ...
2
votes
0answers
18 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 ...
1
vote
1answer
28 views

How to best make use of learning rate scheduling in reinforcement learning?

How to best make use of learning rate scheduling in reinforcement learning? To me, a low learning rate towards the end to fine-tune what you've learned with subtle updates makes sense. But I don't ...
2
votes
0answers
69 views

Is this the correct gradient for log of softmax? [duplicate]

I am currently implementing the very basic version (REINFORCE) of the Monte Carlo policy gradient algorithm. I was wondering if this is the correct gradient for the log of softmax. \begin{align} \...
3
votes
1answer
36 views

How is the incremental update rule derived from the weighted importance sampling in off-policy Monte Carlo control?

Here's the approximated value using weighted importance sampling $$ V_{n} \doteq \frac{\sum_{k=1}^{n-1} W_{k} G_{k}}{\sum_{k=1}^{n-1} W_{k}}, \quad n \geq 2 $$ Here's the incremental update rule for ...
0
votes
1answer
17 views

Why aren’t heuristics for Connect Four Monte Carlo tree search improving the agent?

I’ve created an agent using MCTS to play Connect Four. It wins against humans pretty well, but I’d like to improve upon it. I decided to add domain knowledge to the MCTS rollout stage. My evaluation ...
0
votes
0answers
16 views

DQN not showing the agent is learning in a snake grid environment game

I've been trying to train a snake for the snake game in DQN. Which the snake can essentially just move up, down, left and right. I'm having a hard time getting the snake to stay alive longer. So my ...
2
votes
1answer
198 views

Why isn't my implementation of A2C for the the atari pong game converging?

I have two different implementations with PyTorch of the Atari Pong game using A2C algorithm. Both implementations are similar, but some portion are different. https://colab.research.google.com/drive/...
2
votes
1answer
45 views

How to evaluate a Deep Q-Network

Good day, it's a pleasure having joined this Stack. In my master thesis I have to expand a Deep Reinforcement Learning Network, to be precise a Deep Q-Network, which is used to control machines in an ...
2
votes
1answer
50 views

In what RL algorithm category is MiniMax?

Q-learning is a temporal-difference method and Monte Carlo tree search is a Monte Carlo method. In what category is MiniMax?
4
votes
2answers
156 views

What is the difference between 'prediction' and 'control' problem in the context of Reinforcement Learning?

What is the difference between the term 'prediction'/value estimation in RL as compared to the 'control' problem? Are there scenarios in RL where the problem cannot be distinctly categorised into ...
2
votes
0answers
20 views

Is there an online RL algorithm that receives as input a camera frame and produces an action as output?

I want to build a reinforcement learning model, which takes a camera picture as input, that learns online (in terms of machine learning). Based on the position of an object on the camera, I want the ...
3
votes
1answer
27 views

Effect of the order of the reward function

I have implemented a simple Q-learning algorithm to minimize a cost function by setting the reward to the inverse of the cost of the action taken by the agent. The algorithm converges nicely but there ...
1
vote
0answers
17 views

Benchmarking SAC on Pybullet

So far I have seen TD3 and DDPG benchmarks on Pybullet environments, but I am looking for SAC benchmarks on Pybullet too, anyone can help?
1
vote
1answer
29 views

Is it possible to prove that the target policy is better than the behavioural policy based on learned Q values?

I have retrospective data for a sort of "behaviour policy" which I will use to train a deep q network to learn a target greedy policy. After learning the Q values for this target policy, can we make ...
3
votes
1answer
40 views

Would you categorize policy iteration as an actor-critic reinforcement learning approach?

One way of understanding the difference between value function approaches, policy approaches and actor-critic approaches in reinforcement learning is the following: A critic explicitly models a value ...
2
votes
1answer
67 views

On-policy preventing us from using the replay buffer with the PG?

One of the approaches to improving the stability of the Policy Gradient family of methods is to use multiple environments in parallel. The reason behind this is the fundamental problem we ...
3
votes
0answers
18 views

How do reinforcement learning and collaborative learning overlap?

How do reinforcement learning and collaborative learning overlap? What are the differences and similarities between these fields? I feel like the results I get via google do not make the distinction ...
1
vote
2answers
58 views

How does the gradient increase the probabilities of the path with a positive reward in policy gradient?

Pieter Abbeel in his deep rl bootcamp policy gradient lecture derived the gradient of the utility function with respect to $\theta$ as $\nabla U(\theta) \approx \hat{g} = 1/m\sum_{i=1}^m \nabla_\...
0
votes
0answers
13 views

How's the action represented in MuZero for Atari?

MuZero seems to use two different methods to encode actions into planes for Atari games: For the input action to the representation function, MuZero encodes historical actions as simple bias planes, ...
2
votes
0answers
50 views

Can we use a Gaussian process to approximate the belief distribution at every instant in a POMDP?

Suppose $x_{t+1} \sim \mathbb{P}(\cdot | x_t, a_t)$ denotes the state transition dynamics in a reinforcement learning (RL) problem. Let $y_{t+1} = \mathbb{P}(\cdot | x_{t+1})$ denote the noisy ...
1
vote
1answer
24 views

Can I do state space quantization using a KMeans-like algorithm instead of range buckets?

Are there any reference papers where it is used a KMeans-like algorithm in state space quantization in Reinforcement Learning instead of range buckets?
0
votes
1answer
39 views

Is it required that taking an action updates the state?

For some environments taking an action may not update the environment state. For example, a trading RL agent may take an action to buy shares s. The state at time t which is the time of investing is ...
1
vote
1answer
22 views

Do smaller loss values during DQN training produce better policies?

During the training of DQN, I noticed that the model with prioritized experience replay (PER) had a smaller loss in general compared to a DQN without PER. The mean squared loss was an order of ...
1
vote
0answers
29 views

What are the procedures to get RL paper results? [closed]

I finished working on a new algorithm in Reinforcement Learning, I need to compare it to some well-known algorithms. That's why I need to know the step-by-step procedures that RL researchers usually ...
0
votes
1answer
100 views

Once the environments are vectorized, how do I have to gather immediate experiences for the agent?

My main purpose right now is to train an agent using the A2C algorithm to solve the Atari Breakout game. So far I have succeeded to create that code with a single agent and environment. To break the ...
0
votes
0answers
10 views

How is the parameterised server updated in distributed DQN?

In this paper about Massively Parallel Methods for Deep Reinforcement Learning, the parallelisation of DQN is done via separating the actors and learners. Multiple actors carry out the $\epsilon$ ...
0
votes
0answers
18 views

How to observe or measure convergence of Monte Carlo Tree Search?

As above: how does one observe/measure a Monte Carlo Tree Search to be able to update the algorithm and compare results?
1
vote
0answers
21 views

When to do quantization to decrease the state/action space in RL?

When to do quantization to decrease the state/action space in RL? Can you give me some references that such a technique is used?
1
vote
0answers
19 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. ...
3
votes
1answer
53 views

Why do RL implementations converge on one action?

I have seen this happening in implementations of state-of-the-art RL algorithms where the model converges to a single action over time after multiple training iterations. Are there some general ...
1
vote
0answers
50 views

Is the reward following after time step $t+1$ collected based on current policy?

I am currently learning policy gradient methods from the Deep RL boot camp by Pieter Abbeel in which he explains the actor-critic algorithm derivation. At around minute 39, he explains that the sum ...
7
votes
2answers
135 views

Are Q-learning and SARSA the same when action selection is greedy?

I'm currently studying reinforcement learning and I'm having difficulties with question 6.12 in Sutton and Barto's book. Suppose action selection is greedy. Is Q-learning then exactly the same ...
1
vote
0answers
14 views

Trying to Train Cards Game with RL

I am trying to train a card game called Callbreak I tried inputs like all the opponents discarded cards all hands everything a human can see and calculate with "common sense" I fed it to the Agent but ...
1
vote
1answer
46 views

Why can't DQN be used for self-driving cars?

Why can't DQN be used for self-driving cars? Why can't DQN and similar RL algorithms be used for self-driving cars? The reason why I am curious is that it successfully plays go and other multistate ...
1
vote
1answer
125 views

How can blackjack be formulated as a Markov decision process?

I am reading sutton barton's reinforcement learning textbook and have come across the finite Markov decision process (MDP) example of the blackjack game (Example 5.1). Isn't the environment ...
2
votes
1answer
53 views

Should Monte Carlo tree search be able to consistently beat me in the connect four game?

I've implemented the Monte Carlo tree search (MCTS) algorithm for a connect four game I've built. The MCTS agent beats a random choice agent 90-100% of the time, but I’m still able to beat it pretty ...
4
votes
1answer
47 views

Is there 1-dimensional reinforcement learning?

From what I can find, reinforcement algorithms work on a grid or 2-dimensional environment. How would I set up the problem for an approximate solution when I have a 1-dimensional signal from a light ...
3
votes
1answer
41 views

Simplification of expected reward under the limit in continuous tasks

I was reading the average reward setting for continuous tasks from rich sutton's book (page 202, 2nd edition). There he perform a simplification over the expected reward under the limit approaching to ...
2
votes
1answer
88 views

How long should the state-dependent baseline for policy gradient methods be trained at each iteration?

How long should the state-dependent baseline be trained at each iteration? Or what baseline loss should we target at each iteration for use with policy gradient methods? I'm using this equation to ...
1
vote
1answer
33 views

Can we combine Off-Policy with On-Policy Algorithms?

On-Policy Algorithms like PPO directly maximize the performance objective or an approximation of it. They tend to be quite stable and reliable but are often sample inefficient. Off-Policy Algorithms ...
1
vote
0answers
37 views

How should I deal with variable batch size in A3C?

I am fairly new to reinforcement learning (RL) and deep RL. I have been trying to create my first agent (using A3C) that selects an optimal path with the reward being some associated completion time (...
1
vote
1answer
46 views

How does the AlphaGo Zero policy decide what move to execute?

I was going through the AlphaGo Zero paper and I was trying to understand everything, but I just can't figure out this one formula: $$ \pi(a \mid s_0) = \frac{N(s_0, a)^{\frac{1}{\tau}}}{\sum_b N(s_0,...
2
votes
1answer
79 views

Are these two definitions of the state-action value function equivalent?

I have been reading the Sutton and Barto textbook and going through David Silvers UCL lecture videos on YouTube and have a question on the equivalence of two forms of the state-action value function ...
2
votes
1answer
51 views

Why can't pure KG embedding methods discover multi-hop relations paths?

According to Reinforcement Knowledge Graph Reasoning for Explainable Recommendation pure KG embedding methods lack the ability to discover multi-hop relational paths. Why is it so?
1
vote
1answer
44 views

What does it mean to parameterise a policy in policy gradient methods?

Can you explain policy gradient methods and what it means for the policy to be parameterised? I am reading Sutton and Barto book on reinforcement learning and didn't understand well what it is, can ...
1
vote
1answer
37 views

Q table not converging for an arbitrary experiment

This is an experiment in order to understand the working of Q table and Q learning. I have the states as states = [0,1,2,3] I have an arbitrary value for each ...
1
vote
0answers
24 views

State-of-the-art algorithms not working on a custom RL environment

I'm trying to train a RL agent on a custom, highly stochastic environment (MDP). In order to do so I'm using existing implementations of state-of-the-art RL algorithms as provided by Stable Baselines. ...

1
2
3 4 5
23