Community Profile

photo

Ari Biswas

MathWorks

Last seen: 2 days ago Active since 2020

Statistics

  • Knowledgeable Level 3
  • Revival Level 1
  • First Answer

View badges

Content Feed

View by

Answered
Logging needed Information while training a Reinforcement learning agent.
Unfortunately there is no straightforward way to do this currently but we may have a solution in the upcoming releases (stay tun...

27 days ago | 1

| accepted

Answered
Training agent in reinforcement learning: reproducibility of the code
This could also be as a result of slight variations in floating point numbers across the different computer architectures. These...

1 month ago | 1

| accepted

Answered
Missing savedAgentResultStruct | How do I get the elapsed time from saved agent?
We have recently improved the design of saving agents with relevant training information. In the new design (available from R202...

1 year ago | 0

| accepted

Answered
What's the difference between getAction and predict in RL and why does it change with agent and actor?
The PPO agent with continuous action space has a stochastic policy. The network has two outputs: mean and standard deviation. C...

1 year ago | 0

| accepted

Answered
Reinforcement Learning Agents generating zero episode
There is an issue with the way you specified the reset function. Your function resetRobots should return a Simulink.SimulationIn...

1 year ago | 0

| accepted

Answered
ExperienceBufferLength in Reinforcement Learning Toolbox
The agent will train until at least one minibatch can be sampled from the buffer. If your mini batch size is 64, then the first ...

2 years ago | 0

| accepted

Answered
Saving simulation data during training process of RL agents
Elaborating on Emmanouil's suggestion: There are two ways to log and visualize data during training. Option 1 is to use the t...

3 years ago | 1

Answered
Reinforcement Learning Zero Reward
In your Simulink model workspace you have several agent objects saved with the same variable names as referenced in the RL Agent...

3 years ago | 0

| accepted

Answered
load trained reinforcement learning multi-Agents to sim
It could mean that the agents have converged to suboptimal policies. You can train the agents for longer to see if there is an i...

3 years ago | 0

Answered
Computation Time Reinforcement Learning Toolbox
Training the SAC agent in the ball balance example could take as long as a day, generally speaking. We are working on performanc...

3 years ago | 1

| accepted

Answered
multi-agent deep reinforcement learning
Assuming you are training multiple agents in Simulink using the Reinforcement Learning Toolbox in R2020b: The rewards are calcu...

3 years ago | 1

| accepted

Answered
The reward gets stuck on a single value during training or randomly fluctuates (Reinforcement Learning)
It could mean that the training is experiencing a local minima. You can try out a few things: 1. Change the OU noise options ...

4 years ago | 0

| accepted

Answered
Custom environment in Deep reinforcement learning
One way to solve this is by introducing a property to keep track of elapsed time in your custom MATLAB environment. You can use ...

4 years ago | 0

Answered
Is it practicable to train multiple agents simutaneously using RL Toolbox?
Multi-agent training is currently not supported, however, it will be soon in a future release.

4 years ago | 0

| accepted

Answered
Reinforcement Learning Toolbox train two agent
Training or simulating a Simulink model with multiple RL Agent blocks is not supported at the moment. However it will soon be su...

4 years ago | 0

| accepted