1. Proposal 4

2. yes

3. 9 very clear and well organized

4. 7 this project, while a natural extension of Alex's work, doesn't seem like a very big improvement over current work in developmental robotics

This is a very well-written proposal. It is concise and organized, though at some points it may be a little too detailed. I especially enjoyed the sentence "thereby freeing human
beings to concentrate on finding the answer to life, the universe and everything, one assumes." This project proposal meets the proposal guidelines almost to a T.

Since this project is a direct extension of Alex's work, it fits very well into Developmental Robotics. Indeed this is what I liked best about the idea: not only is it strongly
grounded in previous work, but the author of the previous work is acting as the mentor for this work, making it that much stronger. But the part I like the least about this
project is that it doesn't extend very far form Alex's work, limiting the possible contributions to the robotics community that it could make.

I have two major concerns with this project: The experience of the researchers and the simulator. From the descriptions in the proposal, it appears that none of the researchers
have a strong background in machine learning, robotics, or computer vision (though they do have some experience from taking Computational Perception). Given a project that is so
heavily reliant on all three, this could hinder the project quite a bit. As well the proposal indicates the the team will be creating a robot simulator. This seems like a large
distraction from the main focus of the problem. I'm just curious, but are there no robot simulators already made that you could use?

Given that there is only about a month left to do the entire project, if the team has not started already, then I think it will be very difficult to finish in time. The major
factors that will cause delays in the project are the lack of experience on the researchers' part, and the creation of a robot simulator. The project by itself doesn't seem too
difficult, but given these issues it may get delayed.

For the proposal itself, I think the researchers did a good job articulating their project. The timeline and list of responsibilities are very well layed out and are very well
defined. The test cases are also well defined and easy to understand. I only see a couple issues with the proposal. The related work section had a lot of good related work, but
didn't adequately relate that work back to this project. Also, the babbling behaviour and learning algorithms were not well defined.

To fix these problems, there are a couple things the team can do: First I would suggest checking to see if there are any robot simulators that you can aquire that will fit your
needs (if you already have disregard that); and second I would suggest formally defining the algorithms you will be using so that they are easy to understand. This will give your
project a better theoretical backing. Don't be discouraged by the critical tone of this review, the majority of the proposal is very good and has potential. Good luck.