Clicker Training - Positive Reinforcement Training

There's a new book out called Cat Fancy's Naughty No More which teaches people to use clicker training, management, and other positive methods to resolve many cat behavior issues.

I'm curious to see what positive methods she uses for preventing and curing furniture scratching and counter surfing.

I am already hatching plans to build one of these in my own yard in the spring... ;)

Have you ever used your clicker training knowledge to change your own behavior? If so, tell me about it! What did you try? Did it work?

My partner and I have just set up a bunch of contingencies to change our own behavior. We're having a hard time finding good reinforcers, though! Does anyone have any good ideas for things we can use as reinforcers? So far we've got: buying certain special grocery items, going out for breakfast together, and accumulating money/points towards a bigger-ticket item, like a massage or going hot tubbing.

Any other ideas are welcome! :)

TED talk video here.

Pretty neat stuff! Rats turn out to be much more efficient than machines (for detecting TB), and can be trained at 1/5 the cost of dogs (for detecting land mines).

It puts effective, sustainable solutions for problems into the hands of local people, instead of leading them to rely on outside aid (as most technologies do).

This is an absolutely amazing video of a guy (Jay Sisler) and his trained performing dogs. Apparently he trained them using three 15 minute sessions per day!



I found the video thanks to Positive Petzine, Aiden Bindoff's blog/site.

Check out this amazing father-daughter team! The cats are clearly eager performers, and if you look carefully, you can see that the trainers are feeding the cats after each completed behavior. (When they do the "ta da" hand, check out what the other hand is doing.)

Wolfgang Kohler did experiments in the first half of the 20th century in which apes had to figure out how to get to food that was out of their reach by using "ingenuity"; that is, by putting together previously-learned behaviors in new ways to solve the problem at hand.

Skinner argued that this ability to use "ingenuity" wasn't unique to apes, and proved it by teaching a pigeon three separate behaviors: peck at a banana, stand on a box, and peck the box to move it around.

Here's a video of his results:



I love watching the pigeon find a creative solution to the problem!

I had some free time last night for the first time in weeks, so I decided to do a training session with the dog. It's been forever since we worked on anything new (although we practice his known behaviors regularly, in the course of day-to-day life). We had been working on Level 2 of Sue Ailsby's Training Levels, so I figured I'd work on the one behavior we hadn't tackled at all: going around an obstacle and coming back to me.

I got out a clicker and chopped a piece of apple into tiny chunks. I put our round step stool in the middle of the kitchen floor. I stood behind the stool and called the dog to me, holding out one hand as a target. Albee happily trotted over, tail wagging, and touched the target hand, earning a click and a treat. I then tossed a second treat to reset him, then repeated the process.

Over the course of the session, I gradually moved my body around the stool so that he had to choose to go out of his way to go around the stool instead of coming straight to me. I tossed the treats so that he was positioned to come straight to me without going around the stool. I started clicking and then tossing the treat, instead of giving him one treat for going around and one treat to reset.

Albee went the short way around a couple of times, but on the whole he did really well, learning almost errorlessly. Since he's a fairly soft dog, it's important that I set him up to succeed; he hasn't built up much endurance for failure yet.

But the best part of the whole experience was how happy he was. His tail was up and wagging, his eyes were bright and focused, his mouth was open in a doggy grin. He is the kind of dog who likes to be right, so it still takes some coaxing to get him to try out new behaviors. Once he gets started, though, he loves every second of the game. He even remembered and offered the one other behavior that I'd worked on involving this stool (which we did for one training session of about 2 minutes, weeks ago): putting his front paws up on the stool.

This is why I love clicker training: because the learning is not only painless, it's joyful. My learners are not just willing to do what I want, but often eager. It's a fun game that we play together, not a dictatorial system where they have to do what I want "or else." I like having creative, trusting animals who are willing to try new things. I'd rather have a companion than a doormat any day.

I know you all have had similar experiences with your animals, seeing their eyes light up during training. Tell me about them!

I'm reading Karen Pryor's book On Behavior, and I found this passage in one of her essays on dolphins:

"In traditional force-training of domestic animals, the subject typically is not given choices (other than the implied choice, "Obey, or else!"). For example if one asks a horse pulling a wagon to turn left, one wants only that behavior, and no other: not a right turn or an increase in speed, and certainly not some self-initiated behavior such as standing on the hind legs or jumping in the air. However the training of dolphins by positive reinforcement techniques often gives the animal freedom to demonstrate whatever capabilities it may have. During the training process the animal is at liberty to initiate its own behavior, as well as interactions with the trainer, in a way that is almost impossible in the restrictive circumstances of traditional training of domestic animals (or, for that matter, in the "whip-and-chair" aversive training traditionally used with circus animals). To a certain extent it is this circumstance, rather than some intrinsic characteristic of the dolphin, which has given the public and scientists alike such respect for the animal's cognitive abilities: We get more chances to observe cognitive processes in these animals than in most others."

(Emphasis added by me.)

My two caique parrots have been wearing their Aviator harnesses for about three years. They don't love the process of having the harnesses put on them, but they don't actively fight it either. For a while now I've been wanting to change that, to make it a happier experience for them. Lately I actually set about making a training plan and putting it into action.

In the past I'd started this process by encouraging them to approach and beak-target the harness voluntarily. This was to combat their behavior of flying away upon seeing the harnesses.

The next step of training was to get them very comfortable with the head loop. To do this, I started out by making a large loop out of the harness material and shaping the behavior of putting their heads through the loop.

Once they were comfortable with that, I gradually made the loop smaller and smaller until I could transition to using the actual head loop of the harness, which is very small (especially for a big, bulky caique head!).

Within five days, I had both birds (who had previously flown away from the harness, and occasionally bitten my partner when he tried to harness them) voluntarily seeking out the harness, finding the head loop, and putting their own heads through it! I was shocked at the rate of progress we made.

To illustrate the process, I made a video of some of my training steps:



(x-posted to parrots101, caiques, and clickertraining)

I've been working with Milton on "leave it" - that is, don't eat the piece of tasty food when I tell you not to; instead, make eye contact with me. We're still in the beginning phases of this (we've moved up to me tossing pieces of dried chicken around him, but he's still not 100%, and doesn't yet understand the cue - he just knows he's not supposed to eat the first piece of chicken I throw at him.)

Here's a video of our progress so far:


Today I had to go pick up my stand mixer, which was being repaired at a place about a half hour away. I decided to take Milton and start building some positive associations with being in the car, since usually the only time the cats go in the car is to go to the vet once a year. Since a) it's an infrequent occurrence, and b) they have to go to a strange place and be poked and prodded, the car has some bad associations for them.

A couple of years ago I discovered, purely by chance, that when I let the cats out of their carriers so they can explore the car and calm down a bit they are much more relaxed during their vet appointment.

I decided to work with Milton, who is the more eager trainee. I brought him to the car and spent a few minutes petting him and feeding him slivers of freeze-dried salmon. He shed a little stress dandruff, but ate the treats and settled onto my lap fairly quickly. He was a bit nervous when I started up the car and it started moving, but I used a lot of tactile reinforcement to help him feel safe.

Milton did fabulously on the car ride! He spent a bit of time on my lap, but I fairly quickly transitioned him to hanging out in the passenger seat. He lay down of his own accord, and even rolled over and solicited some belly-rubs. He was all right while I was in the shop, and he settled right down in the passenger seat for the ride home. He wasn't totally comfortable, and wouldn't always take treats, but he lay calmly and did some purring. He only meowed a few times, and didn't pant or pace nervously at all.

1. Last Thursday I decided to grab a few strips of dried chicken (the treats I use when training my cats).  As soon as I opened the drawer where the treats are kept, Milton (the cat) came running over, expecting a training session. I took pity on the poor guy and broke off a big chunk of chicken and tossed it to him. He looked at it, caught himself, and sat up and looked at me - exactly what we've been working on during our "leave it" exercises! What a smartie. 

2. While I was busy doing some sort of household chore this weekend, transversely  decided to have some parrot bonding time, aka reinforce the parrots' "good" behaviors with lots of treats so that they will actually listen to him later on. (They love training, but they definitely have a "what's in it for me" attitude - which I think is healthy.)  He decided to teach Icarus a new behavior: hold a small spoon in his foot. He put the small spoon on the coffee table (our usual training base) and slowly shaped him to approach the spoon (new objects are usually Scary Things to Icarus), beak the spoon, and then grab it with his foot. Go transversely ! All that time spent listening to me blab about behavior has evidently not been ill-spent.

3. Last night during "evening family hang-out time," Icarus started offering spins, the behavior we're currently working on. He offers them frequently, so I'm in the process of adding the cue (the word "spin" and my index finger, pointing downwards, moving in a small circle). We've done a couple sessions of "when I cue you and you do the behavior, you get a treat; when you offer the behavior uncued, you get nothing."  This session was going in much the same way as the previous ones: he seemed to not really understand that uncued offering of the behavior got no treat, and just kept offering and offering it.  So this time I tried a different strategy: I cued him to do a couple of other behaviors, then immediately cued the spin. He responded to the cue every time!  Asking for the other behaviors was designed to prevent him from offering lots of uncued spins. Verdict: he definitely knows the cue, so hopefully the uncued spins will extinguish themselves.

Anyone have any tips on how to more quickly bring a behavior under cue control? Should I just keep reinforcing the cued spins and not reinforcing the uncued spins? 

(x-posted to clickertraining  and rumorofrain )

I've been apprenticing at a local dog training school, and it's inspired me to start training my own animals a little bit again. The behavior I decided to work on is "leave it," i.e. ignore a piece of delicious food on the floor on cue.

I started working on this with Milton, my cat.  I use Dogswell dried chicken or duck strips to train my cats, cut into incredibly tiny pieces. I started by slowly lowering a giant piece of dried chicken to the floor a few feet away from Milton.  If he got up from a sit, I lifted the treat out of reach again. We've done similar self-control exercises before (e.g. "take treats gently from my hand instead of doing shark-mouth"), so he caught on quickly. Soon I was able to put the treat all the way on the floor.

Once we'd progressed to that level, Milton would hunker down and stare intently at the meat with LASER EYES. I made all kinds of goofy noises to get his attention, and when he gave me even an ear-flick I clicked and treated. With a little practice, he understood that if he looked at me, he got a treat, while if he stared at the hunk of chicken on the floor he got nothing. (I stayed close enough to the chicken on the floor to cover it with my hand if he made a move. Surprisingly, he never did!)

Once he understood that treats came from me, I started adding the verbal cue "leave it." He would stare at the floor-chicken, I would say "leave it!" in a cheerful voice, and he would turn his head and look up at me for a click and treat.

After a couple of sessions of this, I started tossing his treats to him so he would have to get up out of his sit to get them. Still, amazingly, he never went for the treat on the floor! I worked it closer and closer to him, but so far he's been great about staring but not touching, and about giving me eye contact when I say "leave it." What a smarty!


Today I've got my two parrots at work with me, so I decided to do a 2-minute training session with Icarus on "leave it," a behavior we'd never worked on before. I used a similar method to the one I used with Milton: I put a chunk of walnut on the table and covered it with my hand. When Icarus stopped beaking at my hand to try to get it, I clicked and treated. He almost immediately understood that backing off = click, and soon I was able to leave the walnut chunk on the tabletop uncovered, perhaps 6" away from him. Even within the first training session, I started adding the verbal cue. Like Milton, Icarus would laser-eye the walnut, but when I said "leave it!" would turn to look at me for a click & treat.

I thought this behavior would be awfully hard to start training! I've been pleasantly surprised in both cases. Way to prove me wrong, critters!

This year we will be having Thanksgiving at our house. Which inevitably means that my brother-in-law Ryan will be over and my dog (an 8 month old lab/shepherd mix) does not like him. Ryan is a nice guy but kind of a rowdy fellow and rather clueless about dog behavior. He is most excellent at having threatening body language and responding to a fear aggressive dog by being threatening, although unintentionally. So, needless to say, anytime Ryan comes over, my dog barks at him and gets increasingly uncomfortable with him around.

So, my current plan of action is to have Ryan be as un-threatening as possible (don't make eye contact, slow movements, quiet voice, stand with arms crossed, face away from him, etc). Then let the dog come and investigate on him own terms and treat him profusely if he is calm about it.

Has anyone else had similar experiences? What has worked for you?

Here's news of an interesting study on the chemical/neurological reasons why spacing out training sessions is better than doing fewer but longer sessions.

Have you all noticed this in your own personal training?

One of the first behaviors I taught my parrots was to wave. Because this grew out of the "shake hands" behavior, they only waved with their right feet.

Recently I decided to teach them to wave with both feet, distinguished by different cues. We're still solidifying the behavior, but they're both getting quite good at "right wave" and "left wave."

I'd like to teach other "right" and "left" behaviors, and eventually even generalize the cues for "right" and "left."

Has anyone else taught their animals "right" and "left" behaviors? What were they? Any tips on how to cue it for future generalization?

I wanted to report back to say that Operation: Don't Go There was a success! Icarus and Daedalus are now consistently flying to the appropriate perch in the kitchen, even though they have to bank sharply to the right to land on it when they fly in from the living room.

My training plan was:

1. Do lots of active training sessions where I cued them to fly from my hand to the perch in question. Gradually increased the distance so they were flying around a sharp corner to get there. Eventually cue them to fly from their living room perch to the kitchen perch. Treat, treat, treat!

2. Whenever they fly onto inappropriate places (cabinets, fridge, pantry shelves, hanging baskets), immediately cue them to fly to the appropriate perch. Treat!

This took a lot of patience at first; it's really annoying to have to get up and call them down off of the cabinets every couple of minutes while I'm eating dinner or reading on the couch. If they persisted in the bad behavior and I got too frustrated, I'd either put them back in their room until I was feeling patient enough to do good training again, or I'd do an active training session to reinforce the appropriate place to land.

I also tried to make the appropriate perch an interesting place to be, so I hung a couple of foraging toys from it and fed them on the tray underneath it.

It only took a week or so for them to really get it. Daedalus is obsessed with flying to one particular spot on top of the cabinets; my guess is that she is ogling the tempting fruit in the hanging baskets below. She is very good about flying back to the perch when cued, though, and she's flying to that spot less and less as time passes.

Check out this letter from Karen Pryor about the R+ training being done at the St. Louis and Denver zoos. The best part is that it includes a video of a rhino doing its cued behaviors! (Mouse over the rhino picture for the video controls.)

My two (3.5-year-old) caique parrots are really good about knowing the places they're supposed to hang out in the house and the places they aren't. (With occasional slip-ups, or possibly boundary-testing, but hey, nobody's perfect all the time.)

Those "acceptable places" have included, for the past few months, the top of the fridge and the top of the kitchen cabinets. However, since they know that the places they are allowed to hang out are also the places it's okay for them to poop, I've decided this needs to change. They usually don't poop there, but even occasional fridge-pooping is too much fridge-pooping for me.

Last night Operation Don't Go There commenced. They already know the "Go!" cue, which means "fly to an appropriate perch," sometimes combined with pointing to said perch to indicate where I want them to go. Their remaining appropriate perch in the kitchen is a hanging toy in the deep window well over the kitchen sink. I spent a large chunk of last night cueing them to fly to me, then fly to the toy to get a reinforcer. If they flew to the cabinet instead, I immediately pointed and used the "Go" cue to redirect them to the right place.

I started just across the kitchen from the perch, so it was within view. As the training session progressed, I slowly moved around the corner into the living room (the living room and kitchen are one big room, and where they and we usually spend our time) out of sight of the perch. I want them to learn that when they fly from the living room to the kitchen, they need to land on their toy (which requires a sharper turn than flying to the fridge or cabinets).

By the end of the training session, they were choosing the toy over the cabinet about 19 times out of 20. Not bad for an evening's work!

(x-posted to caiques)

Here's a National Geographic video clip (with transcript below) of dogs which are clicker-trained to sniff out cancer and hypoglycemia. Very cool!

OK, so there's a local talent cotest going on here just now called "Fife's Got Talent", yes they stole the idea.

Anyway, my boyfriend entered last week and got through to the final. A good friend entered this week and didn't. They both suggested I enter Sparky and do some "Doggy Dancing".

I agreed to give it a go...

So I need:

(a) A song idea. Something up-beat but not too long (roughly 2 minutes). The cheesier the better! I was kind of thinking "Friend Like Me" from Aladdin, but I don't know... Ideas!

(b) Some relatively easy "dance" moves I could teach Sparky. He currently knows how to:

sit
down
roll over
stay/wait
hide (put paw over nose)
jump over my arms
give paw
wave
beg
spin
take a bow
look sad (lie down with chin on floor)
weave through legs
walk between my legs with his paws on my feet
And we're working on crawl
He also knows how to heel and does rudimentary heeling whilst I turn and walk backwards. We can work on those.

And any choreography idea are amazing too!

This still isn't a definite, but I'm still looking for ideas!

Every night when we give the cats dinner, we feed Chaucer first because, being a delicate lady, she eats slowly and daintily. While she's eating, we distract Milton by making him stand on his hind legs and lick the spoon. He's remarkably good at it! My goal is to eventually put standing on his hind legs on cue. Also, check out transversely telling him to sit at the end.



(x-posted to rumorofrain)

shellazure posted a great story about desensitizing hir dog to thunder over in dogtrainers.

My birds are so easy-going that I sometimes get lazy about training them for husbandry behaviors like beak and nail trims. These things need to happen so infrequently that I usually forget to work on training the behaviors until it's too late.

Because of this, my birds have developed some fear of the nail clippers.

Icarus has been getting chips in his lower beak lately, which leave "fangs" which really hurt when he uses his beak to climb onto us or when he does his affectionate caique "surfing" on our hands. When he gets these, I hold him in my hand, put my thumb in his beak to hold it open, and snip or file off the offending fang. This is very minor; any more extensive beak trimming would require the vet. Still, the process isn't very pleasant for him (although he never bites and barely struggles), and when he sees me get out the nail clippers he tends to shy away from my hands. I always give him treats afterwards, and he gets over his shyness as soon as I put the clippers away, but I'd really prefer that it be a more pleasant experience for him.

Tonight was a fang-trimming night. We went through the usual routine: I cupped him in my hand, opened his beak, trimmed down the fang, and gave him a treat afterward.

Once I was done, I decided to try to do a little desensitization around the nail clippers. I got him to take treats from my hand while the clippers were nearby (not in my hand). Once he got over his initial nervousness about the proximity of the Evil Clippers, he really got into the game and quickly moved on to taking treats directly off of the clippers themselves. At this point Daedalus flew over to see what was afoot, and (after her own initial bout of nervousness) started actually targeting the clippers, touching them with her beak! Once she did it, Icarus started targeting them too.

This was amazing! For birds who shied away at the very sight of the clippers, voluntarily touching them with their beaks was a huge step. They would even walk a few inches to target the clippers. The whole training session took less than five minutes.

I stopped the game while everyone was still having fun on the playground. Now I just have to remember to keep up the training even when they don't need any trimming.

(x-posted to rumorofrain, clickertraining, and parrots101)

My Mum has just pointed me in the direction of this. It's a competition for "Britain's Most Talented Pet."

I shall be entering a video of Sparky and probably one of my bunny rabbits too. I thought some other peole here might be interested in entering as well.