User Neil Slater

45 votes

Accepted

What is the relation between Q-learning and policy gradients methods?

answered Apr 28, 2018 at 7:47

45 votes

What's the difference between model-free and model-based reinforcement learning?

answered Jun 13, 2018 at 12:47

36 votes

Accepted

What are "bottlenecks" in neural networks?

answered Dec 31, 2017 at 9:41

22 votes

Accepted

How to define states in reinforcement learning?

answered Aug 31, 2018 at 16:02

20 votes

Accepted

Why does the policy network in AlphaZero work?

answered Sep 14, 2018 at 21:48

17 votes

Accepted

How does LSTM in deep reinforcement learning differ from experience replay?

answered Aug 27, 2018 at 9:28

17 votes

What is the relation between online (or offline) learning and on-policy (or off-policy) algorithms?

answered Feb 10, 2019 at 14:49

17 votes

What is the difference between an observation and a state in reinforcement learning?

answered Apr 9, 2018 at 21:36

17 votes

How do neural networks play chess?

answered Mar 18, 2022 at 16:44

15 votes

In a CNN, does each new filter have different weights for each input channel, or are the same weights of each filter used across input channels?

answered Mar 22, 2018 at 8:24

14 votes

Accepted

Why should the number of neurons in a hidden layer be a power of 2?

answered Feb 22, 2018 at 18:09

13 votes

Accepted

Why does DQN require two different networks?

answered Jul 2, 2018 at 9:12

12 votes

Accepted

What is a "trajectory" in reinforcement learning?

answered Jul 31, 2018 at 15:13

12 votes

Accepted

How is it that AI can become biased, and what are the proposals to mitigate this?

answered Feb 17, 2019 at 11:28

12 votes

Accepted

How can an AI freely make decisions?

answered Dec 11, 2019 at 0:02

11 votes

Why exactly do neural networks require i.i.d. data?

answered Feb 23, 2019 at 20:35

11 votes

Accepted

Is neural networks training done one-by-one?

answered May 25, 2019 at 6:01

10 votes

Can Q-learning be used for continuous (state or action) spaces?

answered May 11, 2019 at 11:32

10 votes

Is the optimal policy always stochastic if the environment is also stochastic?

answered Feb 15, 2019 at 13:47

10 votes

Can the mean squared error be negative?

answered Nov 17, 2018 at 15:56

10 votes

Accepted

Is LSTM a subcategory of RNN?

answered Sep 13, 2021 at 15:57

9 votes

Accepted

Why does the discount rate in the REINFORCE algorithm appear twice?

answered Aug 22, 2018 at 18:58

9 votes

Accepted

What does "stationary" mean in the context of reinforcement learning?

answered Aug 21, 2018 at 8:36

9 votes

Accepted

What is the formula for the momentum and Adam optimisers?

answered Jan 13, 2020 at 9:30

8 votes

Can neural networks be used to prove conjectures?

answered Aug 4, 2018 at 15:03

8 votes

Accepted

Do we know what the units of neural networks will do before we train them?

answered Jul 5, 2018 at 8:02

8 votes

Accepted

How could I use reinforcement learning to solve a chess-like board game?

answered Nov 10, 2017 at 18:23

8 votes

Accepted

Can non-differentiable layer be used in a neural network, if it's not learned?

answered Aug 24, 2018 at 18:58

8 votes

Accepted

Is Experience Replay like dreaming?

answered Sep 10, 2018 at 6:20

8 votes

Accepted

Deep Learning with Best-so-far instead of Where-you-are

answered Jul 13, 2022 at 17:26

Stack Exchange Network

687 Answers

What is the relation between Q-learning and policy gradients methods?

What's the difference between model-free and model-based reinforcement learning?

What are "bottlenecks" in neural networks?

How to define states in reinforcement learning?

Why does the policy network in AlphaZero work?

How does LSTM in deep reinforcement learning differ from experience replay?

What is the relation between online (or offline) learning and on-policy (or off-policy) algorithms?

What is the difference between an observation and a state in reinforcement learning?

How do neural networks play chess?

In a CNN, does each new filter have different weights for each input channel, or are the same weights of each filter used across input channels?

Why should the number of neurons in a hidden layer be a power of 2?

Why does DQN require two different networks?

What is a "trajectory" in reinforcement learning?

How is it that AI can become biased, and what are the proposals to mitigate this?

How can an AI freely make decisions?

Why exactly do neural networks require i.i.d. data?

Is neural networks training done one-by-one?

Can Q-learning be used for continuous (state or action) spaces?

Is the optimal policy always stochastic if the environment is also stochastic?

Can the mean squared error be negative?

Is LSTM a subcategory of RNN?

Why does the discount rate in the REINFORCE algorithm appear twice?

What does "stationary" mean in the context of reinforcement learning?

What is the formula for the momentum and Adam optimisers?

Can neural networks be used to prove conjectures?

Do we know what the units of neural networks will do before we train them?

How could I use reinforcement learning to solve a chess-like board game?

Can non-differentiable layer be used in a neural network, if it's not learned?

Is Experience Replay like dreaming?

Deep Learning with Best-so-far instead of Where-you-are