Milton's Brain
Milton's Brain is an Alexa Skill aimed at teaching kids (and adults) about the brain. It was started as part of the Amazon Alexa Skills Challenge: Kids. Here is the demo video that I submitted.
miltonsBrainDemo from Milton Huang on Vimeo.
Getting it
If you have an Amazon device like the Echo or the Dot, you can try it out by going to the skills store, and typing in 'Milton's brain'. If you don't have a device, you can still access it through a web browser using the Echo simulator. The Echo simulator is a little different than a real device since you have to push a button to talk (either the button on the screen in your web browser or the spacebar if you have a keyboard). It also doesn't time-out like a real device - the Echo prompts you if you don't say anything for 8 seconds, and if you don't reply in 8 more seconds, it leaves the Skill you are in and goes back to listening mode.
Note that all these options require that you have an Amazon account. Kids will need to have their parents go to their Amazon account and provide permission to use this skill as Amazon is careful about COPPA (Children's Online Privacy Protection Act), which places limits around design of Internet experiences for those under 13.
Structure
The game is structured as a list of locations which you can directly jump to. You can also navigate through a sequence of locations by following paths or stories by saying "next". Paths also have choices that allow you to move to different locations, based on what we know of the actual connectivity of the brain.
In addition to locations, there are also facts. These are grouped as a series of statements under a topic like "memory" or "alpha waves". Once you are listening to a topic, the "next" instruction gets you the next fact in that topic till you hit the end when it asks if you want to review. The "more info" command will also cycle you through facts or more information about locations. There are currently 22 topics including:'Glia', 'Alpha waves', 'Beta waves', 'Brainwaves', 'The Cell', 'Cortex', 'Kim Peek', 'Neuron', 'Brain structure', 'Lobes of the Brain', 'Brain fat', 'The knee jerk', 'Brain function', 'Memory', 'Procedural Memory', 'Henry Molaison', 'Lateralization', 'Seizures', and 'Color blindness'.
At any time, you can ask "how am I doing?" to find out how many of the locations and facts you have covered. There is also a "where am I?" command if you get lost. The "help" command should tell you some instructions to try, and also will list a location and a fact you haven't visited yet. The easiest way to explore a new fact or location is to say to Alexa, "you choose." She likes that.
Learning
The program remembers every visited location and fact. The Location descriptions you hear will change, depending on whether you have visited the proper sequence of locations and facts. The program does not collect any personal information, but rather stores visit information in an AWS DynamoDB database that is linked with a specific identifier that Amazon provides. (Amazon always knows who uses a skill, but the developers don't unless they ask.) You can reset all of your progress to the beginning if you want with a "reset" command. Alexa will ask you to confirm with the phrase "full reset" to make sure you want to do that.
Feedback
Because the skill doesn't collect personal information, I can see what Alexa says to people, but I don't know who they are. Still, I would like feedback about things you like and dislike about the program, and especially program flaws. There are areas of the location map where Alexa will tell you that this part of the brain is 'broken', because I haven't had time to add all the places I would like to add. Those are things I'm aware of, but as with any game, I haven't play tested everything so I am particularly interested in the following:
- Bad pronunciation. Alexa reads words so used to say 'conduct' (as behavior) instead of 'conduct' as conducting electricity.
- Bad understanding. Alexa often has trouble understanding words. If you tried to say something clearly and slowly (but not too slow!), I would like to hear it as I can improve the model.
- Program failure. Messages like 'there was a problem with the requested skills response', means my program died. I definitely want to hear about these and if possible the sequence of commands that were used when leading up to it. I also like to know if Alexa doesn't follow the description of behavior I outlined above.
- Poor directions. At the end of each location description or fact (or anything that Alexa says) there should a suggestion about what to do next. I want to know if any of these suggestions are missing or badly worded, or send you somewhere that isn't really appropriate.
- Suggestions. I would like to hear suggestions. Some possibilities:
- topics you want me to cover
- other intro phrases
- other exit phrases
- things that would make it more fun!
Development
If you are interested in what went into developing this software, read my description on DevPost