[docs] Add POCA to major features

[ervteng]

Proposed change(s)

Added POCA to major features on front readme page.

Types of change(s)

• Bug fix
• New feature
• Code refactor
• Breaking change
• Documentation update
• Other (please describe)

Checklist

• Added tests that prove my fix is effective or that my feature works
• Updated the changelog (if applicable)
• Updated the documentation (if applicable)
• Updated the migration guide (if applicable)

Other comments

[mmattar]

I'm cool w/ this change. However, taking a step back, I wonder if having 4 bullets around different trainers is just too much detail. One idea (don't love it tbh) would be to have one bullet that says something like:

• Multiplate advanced training algorithms to support DRL and IL for single-agent and cooperative or competitive multi-agent environments.

This would replace these 4:

• Training using two deep reinforcement learning algorithms, Proximal Policy
  Optimization (PPO) and Soft Actor-Critic (SAC)
• Multi-agent cooperative training using a novel deep reinforcement learning algorithm,
  Multi-Agent POsthumous Credit Assignment (MA-POCA)
• Built-in support for Imitation Learning through Behavioral Cloning (BC) or
  Generative Adversarial Imitation Learning (GAIL)
• Self-play mechanism for training agents in adversarial scenarios

cc: @surfnerd for thoughts.

[ervteng]

@mmattar what if we split it into two lines, one for RL and one for IL? e.g.

• Support for training single-agent, multi-agent cooperative, and multi-agent competitive scenarios via several deep reinforcement learning algorithms (PPO, SAC, MA-POCA).
• Support for learning from demonstrations through two imitation learning algorithms (BC and GAIL).

I think the callout of the algorithms is more for researchers TBH, they would want to know which algorithms the toolkit support at a glance. We might not need to spell out the entire acronym for this audience (though POCA isn't a known/Googleable algorithm).

[mmattar]

@ervteng - I dig it!