Mercurial > cortex
view winston-intro.txt @ 468:258078f78b33
insert argument about senses.
author | Robert McIntyre <rlm@mit.edu> |
---|---|
date | Fri, 28 Mar 2014 16:25:31 -0400 |
parents | 9ba3959b261b |
children |
line wrap: on
line source
1 Dear Professor Winston,3 I've finished the first part of my project, building a4 framework for virtual sensate creatures; I would like your5 help evaluating what I've done so far, and deciding what to6 do next.8 For the work I've done so far, I compiled the results into9 short articles that explain how I implemented each sense,10 with videos that show each sense in action. Please look11 through the articles, in particular the video showcase, and12 tell me what you think.14 Video Showcase : http://aurellem.org/cortex/org/cover.html16 Introduction:17 http://aurellem.org/cortex/html/intro.html18 http://aurellem.org/cortex/html/sense.html20 Physical Bodies : http://aurellem.org/cortex/html/body.html21 Vision : http://aurellem.org/cortex/html/vision.html22 Hearing : http://aurellem.org/cortex/html/hearing.html23 Touch : http://aurellem.org/cortex/html/touch.html24 Proprioception : http://aurellem.org/cortex/html/proprioception.html25 Muscles : http://aurellem.org/cortex/html/movement.html26 Full Demonstration : http://aurellem.org/cortex/html/integration.html28 I think this work could be a fruitful foundation for a29 Master's thesis, so in particular, I'd like critiques,30 suggestions, and project ideas. For example, here are some31 projects I think would be worthy, in increasing order of32 complexity:34 * Create a self-powered joint that can determine its35 range of motion and joint type (hinge, cone,36 point-to-point, etc.) by making exploratory muscle37 movements and observing their effect.39 * Develop an agent that writes and debugs low-level40 motor control programs to achieve simple goals like41 "look at the light" or "extend all of your42 fingers". These simple "calisthenic" programs could43 then be combined to form more elaborate procedures of44 motion, which in turn could be the basic instinctive45 reflexes in the "spinal cord" of some more advanced46 creature. (like Sussman's HACKER program but in a47 richer world)49 * Program a group of creatures that cooperate with each50 other. Because the creatures would be simulated, I51 could investigate computationally complex rules of52 behavior which still, from the group's point of view,53 would happen in "real time". Interactions could be as54 simple as cellular organisms communicating via55 flashing lights, or as complex as humanoids completing56 social tasks, etc.58 * Simulated Imagination -- this would involve a creature59 with an effector which creates an entire new60 sub-simulation where the creature has direct control61 over placement/creation of objects via simulated62 telekinesis. The creature observes this sub-world63 through it's normal senses and uses its observations64 to make predictions about its top level world.66 * Integrate the simulated world with Genesis, so that67 Genesis could use the simulated world to answer68 questions about a proposed physical scenario. For69 example "You stack two blocks together, then hit the70 bottom block with your hand. Does the top block71 move?". This project is complicated and very large in72 scope, but it could be narrowed to focus on a single73 key aspect. For example, one key aspect of turning a74 scenario into a simulation is knowing when you're75 constructing "typical" or "atypical" examples of the76 scenario. So, a narrower project might simply learn77 about the edge cases of different scenarios (e.g. "A78 block stacked on top of another block is usually79 stable, provided the bottom block is large enough, and80 is not moving, and is level, etc."). With this81 knowledge, this kind of program could aid Genesis not82 only in answering common-sense questions, but in83 refining them: "A block is stacked on top of another84 block. Is it stable?"; "Usually, but do you know if85 the bottom block is slanted?", etc.87 These are some ideas, but I think you can come up with88 better ones. I can't wait to hear your critiques and89 suggestions.91 Finally, regarding next year at MIT, can I be considered for92 the position of TA for 6.034 or 6.xxx? Also, do you want me93 to return to MIT at the beginning of Fall or at the94 beginning of Summer?96 Sincerely,97 --Robert McIntyre