I’ve seen a lot of posts here that dive into the details of various modules and testing environments for Monty, but they all leave me feeling rather lost and overwhelmed. Can someone provide a high-level overview of the current and near term landscapes?
Like, what sorts of components (eg, modules) are being tested and developed, what testbeds are they being used with, and what are the major benefits and constraints on each? Finally, what is the current (and planned) state of affairs for connecting the major components together? ELI5…
have you looked at this part of the documentation yet? Implementation Overview The “How Monty Works” section gives an overview of the different components in the Monty architecture and how they are connected. Each subsection that corresponds to one of the main elements from this diagram below, contains a table with the custom versions that are implemented in Monty and a description of them.
For instance, if you want to read more about the different types of learning modules we experimented with, you can have a read here: Learning Module or for the different environments we have integrated in Monty, you can have a look at the table at the bottom of this page: Environment & Agent
