Jun Tian

Questions on Contributing Python Demos

4

Last year I had studied the book [Machine Learning: a Probabilistic Perspective](http://www.cs.ubc.ca/~murphyk/MLbook/). However, with only a little knowledge of Matlab I found it hard to understand the source code. Then...

Add link to the Julia binding

3

This package is awesome! The APIs are really concise, which makes it very easy to provide a Julia binding with the C API.

Disable formatting multi-line comment inline

2

``` julia> format_text("f(a;#= b=3 =#, c=4)") "f(a, c = 4)#= b=3 =#" ``` Could we disable formatting here?

Next Release Plan (v0.11)

135

# Goal Improve the interactions between ReinforcementLearning.jl and other ecosystems in Julia. ## Why is it important? In the early days of developing this package, the main goal is to...

Currently, in each `policy` or `learner`, we allocate a temp memory to record intermediate data. And to record these data, we need to add an extra hook. There're at least...

Model based reinforcement learning

1

After taking a look into [facebookresearch/mbrl-lib](https://github.com/facebookresearch/mbrl-lib) . I think our current design is flexible enough to implement most algorithms in it. It's just that we need to standardize some interfaces....

Ref: https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/

Combine transformers and RL

1

This seems like an interesting intersection between NLP and RL. I'll give it a try when I have time. - [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) - [Reinforcement Learning...

Rename some functions to help beginners navigate source code

3

I only realize this problem very recently. Multiple dispatch seems to be overused here in this package. For example, the `update!` function. I thought it was quite straightforward. When we...

A2C and PPO can be improved further to support mutiple discrete action space