Project:  Building Elements of Compassionately Efficient Cooking Bots aka CBots

 

Prologue: It is the dream of many to have bots cook dinners after a long day working remotely at home J.  You goal is to help these cooking bots do a better job by converting recipes in human text form so that these highly efficient but reading impaired (for now)  Cbots work with a graph representation of recipes. Yes, these Cbots are pretty good graph readers.  Thus if you build this tool well, you can be assured of  good meal in the coming future. Even better given that AI and robots will take over the world, you will apply the salve of visual analytics and make the bots compassionately efficient. They can cook what is asked asked of them but do it well with a smile like a doting parent J.  Parsers, graph drawing are all needed. What is missing is the fact that algorithms are brittle. Parsers, natural language processing tools produce not-so-perfect output which cannot be reduced to clean, well-laid-out graphs.  This is what your Cbot Recipe Process Visualizer will come in handy. You will get the tokens, make a simple graph and them stitch it into a useful graph that will allow for the cooking of delicious pasta dish J.

 

Resources: Please find

1.     Class slides here.

 

The Data:

Any recipe site is good. Our only recommendation is that you use a repository that uses clean and crisp language. For that reason using http://cooking.nytimes.com/ is a swell idea. Using other sources is good as long as they are easy to parse.

 

Deliverables: It is required you to create visualizations which allow the easy creation and The visualization has to be interactive. No way around it. Sorry.  It needs to deployed on at least three non-trivial recipes. You stick with pastas or another popular dish or dishes drawing from any of the globes’ many culinary traditions.  

 

And most importantly, a user evaluation of your peers should be included. Come up with a list of properties of the recipe graph. These could include “Do you find the text and the graph to be similar?”, “Does the layout of the graph allow for the easy understanding of cooking steps ?” and so on.  Let your peers (about 10) score them using a binary scale. Essentially, you need to demonstrate that a process graph visualization tool is better than just poring over recipes and.

 

In decades to come, a human will use your tool and program the Cbot so that he/she is well fed and can fend off bots who will let AI rule the world.

 

Finally, yes you need to submit code, a report (with results of the user study) and any material that is required to demonstrate the project. If you cannot show in person make sure you have slide-decks/videos, links to web sites, etc. And yes you make your deposit into a Carmen dropbox.