Proposal number:  2
Should this proposal be considered for the Best Proposal prize?  Yes.
On a scale of 1 to 10, how would you rate the overall organization/clarity of the proposal? 10
On a scale of 1 to 10, how would you rate the overall project idea? 9.5

Feedback:

	This proposal is very well written, and it’s clear that the authors have put a lot of time and effort into making this proposal clear and concise.  The proposal is well
organized and easy to follow.  From the point of view of a reader, the content of this proposal is excellent because the general flow of information moves in a comfortable way. 
This seems like a very complicated project, but the authors have supplied more than enough background information and previous work to give the audience a good understanding of
your idea before the project idea is even discussed.  Furthermore, the introductory material not only sets up the framework for the project, it recognizes some flaws of these
previous projects, and most importantly; it establishes a sense of necessity for the project in the viewpoint of the reader.  Many people can explain the background of a project,
but it takes rather eloquent writing to persuade the reader to recognize this necessity.  You have shown exactly where the project will fit in the scope of developmental robotics,
and this is important when considering many other aspects of your proposal such as initial clarity, professionalism, and comprehension down the line.  Had this been a real
proposal, I am confident that this proposal would be selected for funding due to the overall professionalism, conciseness, and completeness.  Excellent job!

So let’s talk about the idea:

	My favorite part of the project idea is the explanation of autonomously differentiating objects with less user input.  It is explained in the introduction, that a few previous
approaches to this project assume that the groups of objects are differentiable and it is explicitly made clear in the experiment what object is being manipulated in the
exploratory behaviors.  This idea of giving the robot a group of items and letting it decide whether it has been previously explored or not is a fundamental question in object
recognition and there are few projects that approach this problem.  If a general method of approach could be developed to solve this problem, our current level of robotic
application and usefulness will skyrocket.  This brings me to one concern, that for the ability to re-recognize a familiar object, one must have a pretty robust system.  The
success rate of this process will increase with robustness.  However, in my opinion, there may reach a point where an overgeneralization occurs or the system becomes too task
specific to be improved on in the future.
	My least favorite part is how task specific the procedure is.  This paper looks to be somewhat of an expansion of another couple of papers discussed in class, where the robot
was given a group of objects and a group of exploratory behaviors, and was to find ways to characterize these objects in terms of their properties.  To avoid criticism here, I
will add that you have introduced a new level of generality beyond the aforementioned projects, and I do not mean to put a label on your project, because it is obviously more
complicated than I am making it sound in this paragraph.  It would have been nice to read a little more about how you may plan to use the results of these previous approaches in
the interest of your own project.

Here are a few things that may help you improve on your paper and idea:

	The approach lacked a clear explanation of a target audience.  Furthermore, the points of evaluation in the proposal did not seem to support how a model such as yours can be
employed in robotics and serve towards the advancement of developmental robotics.  Over the course of a year or so, I have been reading a lot of publications, mostly from students
in graduate positions.  I have come across many, even a few in this class, where little explanation is given on how the project and results can be used by future experiments or
projects.  No one intends to do this, but it is almost like “ok, I published a paper, now I can have my masters’ degree.”  They are merely finding results to report on.  Now, I know
that this is not the main point of your paper, and it would have been nice to read about, for instance: how this model could be used in a robot to be successful in daily
activities, or how a robotic system could benefit from such a system.  On that point, it would have also been nice to read about an intended target audience, to know exactly how
the above points could be implemented.  This project is a fundamental building block in developmental robotics, so it makes sense that there is no specific target audience, but it
would be nice to know how such a system could be applied to be useful to the general public.
	I had one interesting idea that may help make your project even more general.  Many papers thus far regarding your approach use simple exploratory behavior hard wired to the
robots motor controls by use of specific routines.  It would be interesting to have a system within the robot where, as it begins to recognize its objects, it also knows what
variations of its exploratory behavior yield the most information about the object that is being manipulated.  For instance, if a robot shakes a box of nails, it will get more
auditory input if it shakes the box really fast, but with a half filled water bottle, it may be more beneficial to shake it slow.  This could, in essence, simulate a stage of
development with the exploratory behaviors used in your project.  These modifications to exploratory behaviors could even serve as feedback to reinforce the robots prediction of
object identity.

To sum up.

	This project seems very complicated, but with the information of skills and experience possessed by the team members, seems sufficient to complete the task in the time frame
with a high level of devotion.  The most complicated aspect to me seems to be the scripting and tools used to characterize the objects, however, there are no methods or
assumptions made that make the level of this project unreasonable.  Overall, this proposal is very strong.  I look forward to learning more about your project as time goes on.