Acrocats

This was shared with me from my mother-in-law:

https://www.youtube.com/watch?v=PFsNGOVj71I

The Acrocats!

Modern Principles of Shaping - 10. Quit while you're ahead!

I'm writing a 10 part blog post where I discuss the Modern Principles of Shaping by Karen Pryor, to help solidify them in my mind.

10.   Quit while you’re ahead. End each session with something the learner finds reinforcing. If possible, end a session on a strong behavioral response, but, at any rate, try to end with your learner still eager to go on.

1. Stop the session on a postive note.  Specifically, a situation where the learner gets a C/T.  
2. When ending, make sure to either stop with a C/T where the learner does something good OR manufacture one by going back a step or two, so the learner gets a reward to remember for next time.  
3. Ending with the learner eager to continue is hard.  If the learner is eager to continue, stopping could be interpreted as a punisher.  But it also makes the more likely to come back next time with renewed vigor.  
4. A jackpot doesn't need to come at the end, but it could help if occasionally the jackpots are at the end and distract the learner from the fact that the session is over. 

I'm not really in agreement with this wording.  It is really focused on the goals of the teacher, which make sense because the Principles of Shaping are geared toward putting together a shaping plan with a learner.  But, it is the learner who you need to keep engaged and if engagement drops off you as the teacher have to realize it and stop.  

Modern Principles of Shaping - 9. Stay ahead of your learner!

I'm writing a 10 part blog post where I discuss the Modern Principles of Shaping by Karen Pryor, to help solidify them in my mind.


9.       Stay ahead of your learner. Be prepared to “skip ahead” in your shaping plan if your learner makes a sudden leap.

1. Stay ahead both when progressing on the shaping, but also when reading your learner and knowing when to move in and out of the shaping process.  Both need to be calculated so you can end on a positive note.   
2. Skipping ahead is good, but I'm not confident that I can identify it when it is happening.  You can gauge the learners grasp of something by attempting to solicit a behavior 5 times, and if the learner meets your criteria 4 of the 5 times, you can move on.  But if the learner exhibits something much farther down the path perfectly, be prepared to capture it.  
3. Don't just be prepared for a leap in what you're shaping, but also for things that are not being shaped right now, but will need to be in the future.  At least this is what I'm doing now, and don't know otherwise.  I need to do a little more reading on it. 
4. This will build confidence that you know what you're doing.  This is more important for a human learner, I think, but any animal can lose confidence that you are reliable, and give up the attempt. 

Modern Principles of Shaping - 8. Keep your attention on your learner!

I'm writing a 10 part blog post where I discuss the Modern Principles of Shaping by Karen Pryor, to help solidify them in my mind. 

8.       Keep your attention on your learner. Interrupting a training session gratuitously by taking a phone call, chatting, or doing something else that can wait often cause learners to lose momentum and get frustrated by the lack of information. If you need to take a break, give the animal a “goodbye present,” such as a small handful of treats.

1. I'm spending a lot of time making this one happen.  I do get distracted sometimes, but most of the time I feel like I'm watching the learner more than anything else. 
   
2. I don't interrupt a training session.  Instead I end it.  Not sure if this is better, but having a short training session has always been a better choice than having one that in some way is inferior. 
3. The learner can get frustrated in a lot of ways, but not having me pay attention isn't one of them.  I'm glad I'm following this already!
4. Should the goodbye present be a C/T - a jackpot?  That seems like a better idea than just giving out a bunch of treat. 

Modern Principles of Shaping - 7. Go back to kindergarten!

I'm writing a 10 part blog post where I discuss the Modern Principles of Shaping by Karen Pryor, to help solidify them in my mind.


7. Go back to kindergarten, if necessary. If a behavior deteriorates, quickly revisit the last successful approximation or two so that the animal can easily earn reinforcers.

1.  First, this suggests that you don't need to continually re-teach something.  You should teach it and maintain it by occasionally rewarding, so the behavior keeps occuring.  
2. Behavior can deteriorate because the behavior isn't being rewarded over a long period of time, the learner is finally understanding what the teacher wants and has forgotten what they're doing or the behavior has become sloppy over time.  
3. The trip to kindergarden is usually very quick, maybe 1-3 reinforcers per approximation.  Also, the approximations can skip over levels if the learner is understanding what you're doing.  
4. The more reinforcers you can give to the learner, the less likely they are to become frustrated.  Keeping the request within the normal boundaries of variation means they are doing it sometimes, and you can ratchet the request down further. 

Modern Principles of Shaping - 6. Keep Training Continuous!

I'm writing a 10 part blog post where I discuss the Modern Principles of Shaping by Karen Pryor, to help solidify them in my mind.

6. Keep training sessions continuous. The animal should be continuously engaged in the learning process throughout the session. He should be working the entire time, except for the moment he’s consuming/enjoying his reinforcer. This also means keeping a high rate of reinforcement.

1. Continuous engagement can me either that the learner is always doing something or getting CT.  OR it could also mean that the learner is always eager and paying attention, even if they aren't actually doing anything.  The latter situations is what I strive for, and sometime hope to accomplish so I can capture some behavior  
2. Working the entire time means that the animal should respond as best they can to what they think you want them to do.  However, if something is on queue, and they choose to do something different, they're still engaged.  
3. Training time is valuable to both the teacher and learner, so be sure to set the expectation that the reward source is available now.  
4. A high rate of reinforcement could be as many as 20 treats per minute.  That seems tough to do when the learner is enjoying their reinforcer.  But maybe that means my treats are too substantial....
5. "High rate" is subjective, so some subjects will be able to do more with a lower rate of reinforcement.  
6. Reinforcement (and the information conveyed by the click) is the key to successful training.  So the more C/T you have/minute, the more likely you are to see change and success. 

Outstanding Questions: How many times should you repeat something?  Should the same criteria be used for multiple rewards in a single session?  Or is switching between criteria every C/T OK? 

Modern Principles of Shaping - 5. If one door closes, find another!

I'm writing a 10 part blog post where I discuss the Modern Principles of Shaping by Karen Pryor, to help solidify them in my mind.

5. If one door closes, find another. If a particular shaping procedure is not progressing, try another way.

 

1. Depending on if this happens toward the end of within a shaping session, you may need to do this in real time.  I was trying this yesterday with Jazz with getting her to move her right paw so she can get toward a shake.  Very tough to separate actions, and easier if you can go the smoothest route.  
2. Alternatives for how to get the beginnings of a behavior: Capture, Targeting, Luring.  I tend to use capture, but it can slow down the process if the cat is just looking at me.  
3. Preparation for this part would be considering alternatives for how to get certain behaviors.  If you want the cat to raise its paw, maybe you move your hand in and the cat paws at it, or you lean forward, and the cat leans back and raises one paw in the process. 
4. This one is probably the hardest for me.  I tend to have a lot of patience, but I don't have as much as a cat.  

Outstanding Questions: How many times should you repeat something?  Should the same criteria be used for multiple rewards in a single session?  Or is switching between criteria every C/T OK? 

Modern Principles of Shaping - 4. Relax criteria when something changes!

I'm writing a 10 part blog post where I discuss the Modern Principles of Shaping by Karen Pryor, to help solidify them in my mind.

4. Relax criteria when something changes. When introducing a new criterion or aspect of the skill, temporarily relax the old criteria for previously mastered skills.

 

1. Introducing a criteria could be, for a human, spelling out what you want that person to do.  In this case, use only 5 words, use a positive statement, and make it memorable/measurable. 
2.  For non-verbal (differently abled, animals), begin rewarding a different criteria, but be lenient on the other.  This can be a hard experience, because the subject is now being rewarded for something different than before.  
3. When you relax the old criteria, use your judgement about how much to relax it.  For the first few C/T, relax it completely, to the point that you're not looking for it at all.  
4. When the new criteria is trained up, you can begin adding the first criteria back in.  The learner needs to be able to do both of the criteria before you ask for both.  
5. Don't lose heart that the learner seems to have lost all the things he learned.  You relax the criteria because the learner WILL be more variable in what they know until they know the new criteria. 

Outstanding Questions: How many times should you repeat something?  Should the same criteria be used for multiple rewards in a single session?  Or is switching between criteria every C/T OK?