1. Proposal number: 12 2. Should this proposal be considered for the Best Proposal prize? (yes/no) no 3. On a scale of 1 to 10, how would you rate the overall organization/clarity of the proposal? (1-10) 7 4. On a scale of 1 to 10, how would you rate the overall project idea? (1-10) 8 Then write 2-3 pages of helpful feedback to the proposal's author(s). The following questions should help you organize your feedback: * Overall, is the proposal clear, concise, and well-organized? * Does the proposal meet the posted proposal guidelines? * How does the project idea fit within the framework of Developmental Robotics? * Describe what you like BEST about the project idea. * Describe what you like LEAST about the project idea. * Do you have any concerns about the project? * Does it seem doable in the remaining time? * Does it seem too difficult? * Are there any major details left out? * Does the idea rely upon technologies that are not currently available? * Do you have any suggestions for improvement? * Do you have any suggestions for related work that should be cited? * Any other comments or suggestions? OVERALL: This purpose of this proposal is to study the application of Intelligent Adaptive Curiosity to the simulated snake robot from the snake video game. The proposed project is different from related work because the body form of the snake changes over time. The goal of the project is to show that the behaviors of the snake grow more complex over time, which sounds like they may be different from existing work given the added complexity of the variable body size. It would be interesting to see what kind of body size the snake stabilizes to with IAC. It might be the case that there are several locally optimal body sizes for the snake. It would be worthwhile to run several experiments to identify which body forms the snake converges to, using different values for: 1) the initial snake size; 2) the size that the snake grows per food item; 3) the number of food items on the field; and 4) the size of the field. To establish a baseline for comparison, you could also identify how the behaviors of the snake evolve over time when the snake doesn't change size. The results could include a comparison of the rate of behavioral change and the types of behaviors learned for both the static-sized snake and the dynamic-sized snake. So, instead of having his part optional, think about making it the main way to evaluate the project. Another goal of this project is to evaluate different objective functions that are used to drive the IAC algorithm. The reasoning is that some algorithm may lead to more complex behaviors than others, but little related work has addressed the best algorithms. It's nice to see that IAC is a relatively recent approach to learning (2007). It seems like the evaluation for the complexity of the behaviors is a bit subjective. Is there a good, objective way to say that the robot's behaviors are actually meaningful, given that it only moves around in 4 directions? You may not have enough time to do the work you proposed. You could take out the user interface and don't worry about doing anything with the robot. Instead focus on finding a learning algorithm that you can use to implement IAC. IMPLEMENTATION DETAILS/QUESTIONS: --It's huge that you don't yet know what learning algorithm you will use. --What are the limitations of IAC that may work against you in this project? SPECIFIC COMMENTS ABOUT EACH SECTION: TITLE: The title is too general. It doesn't tell me what your project did. INTRODUCTION: This is a bit too long. A more effective intro could be written in 5 paragraphs. Think of it like an argument essay. Holes in existing research would be identified, and it would be concluded by stating what your paper did to address the holes in existing research (like what you wrote in your second-to-last paragraph). Also, in a concise intro, there is no need for a summary of it. Specific points: --The second and third paragraphs are more like related work and add little to the intro. --It's nice to know words like 'burgeoning', but using big words can subtract from the flow of the paper, which is the case here. --Paragraphs 4-6 could be condensed. PRIOR WORK: It's nice to see that the citations are up to date. Have you tried searching for citations of simulated robot learning in games that are like snake? APPROACH.Game Specifications: The learning aspect of this project is the most important part, but is not finalized. This is huge. There is probably only enough remaining time for coding, making the poster, and writing the paper. If you don't figure out this part quickly you may have trouble getting a good project. When will you have an idea of how the robot will learn? What you need is a continuous learning algorithm. You might be able to find an existing implementation. Try Weka utils in Java. It has several different types of learning algorithms. However, I don't know if any of them offer real-time learning. What algorithm do they use in [8]? Can you use the same one for your project? APPROACH.Algorithms: APPROACH.User Interface: Why do you need a user interface for this project? You could always write a function to randomly place the apple somewhere on the field rather than having an experimenter manually place it. Your results would be more meaningful if the apple was placed at a random location on the field. You could always specify the parameters in the code itself. How do you know what learning parameters you will have in the 'SETTINGS' window, since you don't know what learning algorithm you will use? Also, what is a knowledge adjustment weight? The idea to run the learning algorithm on the robot should be moved to the future work section. You don't have time for this. TIMELINE: Although you won't get this back for a while, you probably don't need to spend more time on related work. The implementation of this project is the challenging part, which will consume most of the remaining time. EVALUATION.Goal: Take out the citations and focus on how you will evaluate your work. It looks like you did this in the next section, so remove everything in the goal section. (See the overall section of my review above for an additional idea of how to evaluate this project). REFERENCES: Start reference numbering with [1], not [0] Also, you don't need the url in the references to publications, just give the complete citation (like what you gave for [8]).