I am trying to train a neural network to navigate a physical robot through a maze. I have no training data and have to use reinforcement learning to train it. I am using a deep q network. However I am running into problems when trying to generate the training data for the experience replay. As far as I understand it at the beginning of deep q learning the q network has to predict q values for all possible actions. In the simulation I am using for training the robot can rotate 360 degrees and the move forward as much as it wants every time it takes an action. The number of possible actions it can take is too big to be able to reasonably compute all of them. Also because the q network has to predict the q values of every possible action at every step I can't just do this calculation once to populate the experience replay and be done. I have heard this a neural network can sometimes make this problem easier but that you need an output neuron for every possible action. With the number of possible actions this doesn't seem doable either. Is there any way to train a deep q network without calculating q values for every possible action. Thank you in advance and sorry for long post.
1
There are 1 answers
Related Questions in REINFORCEMENT-LEARNING
- pygame window is not shutting down with env.close()
- Recommended way to use Gymnasium with neural networks to avoid overheads in model.fit and model.predict
- Bellman equation for MRP?
- when I run the code "env = gym.make('LunarLander-v2')" in stable_baselines3 zoo
- Why the reward becomes smaller and smaller, thanks
- `multiprocessing.pool.starmap()` works wrong when I want to write my custom vector env for DRL
- mat1 and mat2 must have the same dtype, but got Byte and Float
- Stable-Baslines3 Type Error in _predict w. custom environment & policy
- is there any way to use RL for decoder only models
- How do I make sure I'm updating the Q-values correctly?
- Handling batch_size in a TorchRL environment
- Application of Welford algorithm to PPO agent training
- Finite horizon SARSA Lambda
- Custom Reinforcement Learning Environment with Neural Network
- Restored Policy gives action that is out of bound with RLlib
Related Questions in Q-LEARNING
- How do I make sure I'm updating the Q-values correctly?
- Which Q-value do I select as the action from the output of my Deep Q-Network?
- Does anyone have an Atari breakout Deep Q-learning implementation that works?
- Pytorch deepQL code that fail when using tensor.flatten() instead a One Hot Encoding function
- Does initializing a Q-table with zeros introduce bias towards the first action in reinforcement learning?
- Game-like model in Q-learning
- Using non-negative derivative to enforce same results between functions
- How to resolve the issue of Input layer expects different dimensions
- Qtable index out of bounds
- Tabular Q-Learning: Is a variable for "action_history" needed for backpropagating the q-value for all previous actions?
- Tabular Q-Learning for TicTacToe - Only the last state/action pair is stored in the Q-Table Dictionary with a value other than 0
- flappy bird linear q leanring approximation don't learn
- Q Learning agent taking too many steps to reach goal
- Python Gymnasium Render being forced
- How to do Q learning delayed reward
Popular Questions
- How do I undo the most recent local commits in Git?
- How can I remove a specific item from an array in JavaScript?
- How do I delete a Git branch locally and remotely?
- Find all files containing a specific text (string) on Linux?
- How do I revert a Git repository to a previous commit?
- How do I create an HTML button that acts like a link?
- How do I check out a remote Git branch?
- How do I force "git pull" to overwrite local files?
- How do I list all files of a directory?
- How to check whether a string contains a substring in JavaScript?
- How do I redirect to another webpage?
- How can I iterate over rows in a Pandas DataFrame?
- How do I convert a String to an int in Java?
- Does Python have a string 'contains' substring method?
- How do I check if a string contains a specific word?
Trending Questions
- UIImageView Frame Doesn't Reflect Constraints
- Is it possible to use adb commands to click on a view by finding its ID?
- How to create a new web character symbol recognizable by html/javascript?
- Why isn't my CSS3 animation smooth in Google Chrome (but very smooth on other browsers)?
- Heap Gives Page Fault
- Connect ffmpeg to Visual Studio 2008
- Both Object- and ValueAnimator jumps when Duration is set above API LvL 24
- How to avoid default initialization of objects in std::vector?
- second argument of the command line arguments in a format other than char** argv or char* argv[]
- How to improve efficiency of algorithm which generates next lexicographic permutation?
- Navigating to the another actvity app getting crash in android
- How to read the particular message format in android and store in sqlite database?
- Resetting inventory status after order is cancelled
- Efficiently compute powers of X in SSE/AVX
- Insert into an external database using ajax and php : POST 500 (Internal Server Error)
Well, you would want to calculate the q-values for every possible action otherwise you do not know which action yields the highest expected reward. So, you would want to limit your number of actions. You could do that by stating not allowing all 360 rotations to be involved, but rather the movement as output, such that 2 actions: turn left 1 degree and turn right 1 degree, will make up all 360 rotations of the robot, limiting the action space in this section from 360 to 2. Of course, you could also program this a little smarter, stating that you would want to rotate by 45 degrees or 90 degrees, depending on your task.
For sure do you want to use a DQN when the observation space is large. Depending on your task, you might also want to take a look into other models, such as PPO. Check https://stable-baselines3.readthedocs.io/en/master/ for easy application and training of reinforcement learning models.