1. Project number and Title
 Proposal 4
 Learned Shape Affordances through Simple Tool Use
2. Should this project be considered for the Best Project award?
No
3. Should this project be considered for the top 3 project awards?
Yes
4. On a scale of 1 to 10, how would you rate the overall organization/clarity of the project?
9
5. On a scale of 1 to 10, how would you rate the overall project idea?
8
6. On a scale of 1 to 10, how would you rate the overall research contribution of the project
idea, methodology and/or results?
8
Feedback
Overall, is the project report clear, concise, and well-organized?
Overall, you have written a very detailed and well laid out report—there were a few typos and
missed proposal holdovers (like a missing figure) but overall a very nice report. The table of contents
demonstrates your structure and organization. I appreciate the inclusion of a section dedicated to
identifying your audience—given their description, I feel you have done a good job of writing to the
necessary level of detail. You have also done a great job of clarifying the limitations of current robot
machinery and why robots with recognition of object affordances may be beneficial. You all have really
added a significant amount of background information and content overall to your final report. This
addition clearly demonstrates your understanding of affordances and the possible advantages of their
implementation in robotic systems. The previous and related work specifically is very well written—the
linkages between paragraphs is very well thought out and the incorporation of the other authors’ ideas
is well done. The implementation section is solid—you have clearly wrapped your heads around the
simulation environment and how to implement machine learning in it. The risk section seems
unnecessary considering the project is over or perhaps It could have been added as a conclusion section.
The figures and charts are very helpful in illustrating your findings and give the reader a great appreciate
for the amount of work that went in to setting up the simulation environment. Given the difficulties you
did encounter, you do well to explain these issues, how you addressed them, and ways you hope to
possibly rectify them in the future. Your overall conclusion and findings seems relatively brief in
comparison to the lessons learned and future work, but you have clearly demonstrated a simulated
behavior-grounded tool affordance by a computer. Very well done all! This seemed to take a significant
Page 6

Page | 5
amount of work so I commend you on producing interesting results and hopefully a useable platform for
future developmental robotics students.
How does the project idea and methodology fit within the framework of Developmental Robotics?
The project idea and methodology fit perfectly within the framework of developmental
robotics—especially considering you are building from Alex’s pervious research! The methodology is
comparable to other papers we have read throughout the course. You clearly demonstrate your
knowledge of affordances as they occur in intelligent beings and how they might be learned by a
machine. In the end, you have developed a machine that learns from its experiences with a tool and
various shapes. Through its interactions, as you state, you were able to show that the exploratory
behaviors of the machine were sufficient for determining the affordance of a few basic shapes. It is very
interesting to see the computer attempting to interact with the environment in an intelligent manner.
Describe what you like BEST about the project.
I really appreciated your approach to developing a simulated robotic environment. While of
course real world applications are always more desired, this may become a powerful tool for individuals
who do not have access to a robotic platform—at the very least, a great starting point. You all appeared
to be very knowledgeable on setting up the simulated environment which was not an easy task given
the difficulty and working with an unfamiliar open source SDK like Bullet. I would hope that your
frustrations serve as a great starting point for future groups.
Describe what you like LEAST about the project.
I really feel your conclusions section is really brief. You spent a lot of time in the background
section defining affordances and then in the end only give one brief paragraph covering how your
implementation met these expectations. I think it would have been nice to go full circle and
demonstrate explicitly how you taught a machine to recognize the affordance of the tool and shapes
using the same terminology from the background section. Instead, you spend a lot of time talking about
the issues you had, what you learned from the issues, and how they might be improved in the future;
this seems to take away from the actual great work you did do!
Do the methods, results and contributions of the final project correspond to what was presented in
the initial project proposal?
The methods, results, and contributions of the final project correspond to what was presented
in the initial project proposal. I really feel you made a great improvement from you proposal in
demonstrating you knowledge on the topic of affordances and your overall software implementation of
them being learned by a computer. In the end, you met the objectives of your proposal.
Are there any major details left out with regards to the methods, algorithms, or experimental design
described in the report?
Page 7

Page | 6
I really have to say you packed a lot of information into your report. It is hard to say that
anything is missing. There was one apparent missing figure in the “Applications to Robotics” section, but
other than that, I believe your approach was sound.
Do the experimental results reported in the paper demonstrate success?
I would have to say that the results in the report paper demonstrate success but it almost seems
discounted given that the discussion on the success is very brief. I believe that discussing more on how
you achieved success would clarify this discrepancy.
Do you have any suggestions for improvement and future work?
I would suggest looking into the Robot Operating System (ROS) developed by Willow Garage and
see if your software can be built in to the RVIZ package. The ROS RVIZ environment is a 3D visualization
environment for robots that appears to be fairly good at importing existing robotic models. It is meant
more for physical interaction but I see how it may be useful for your application. You have done well to
identify other future works—dynamic learning, learning with other tools and other shapes. It would be
nice to learn how your efforts can be built upon for future work.
How close is the final project report to being publishable as a conference or journal paper (consider
the research papers that were part of the course reading)? What would it take to get there?
The final project is close to being publishable as a conference or journal paper. There were a few
typos and sentences that seem to be holdovers from the project proposal that need to be updated. I
think added some more significance to your conclusion it would greatly enhance your paper.