A blog about dog training and dog breeding…and other sundry matters

Posts tagged ‘operant conditioning’

Effective Training: The Reinforcer

Remember when training a dog, whether adult or puppy, that the reinforcer must be something that the dog finds rewarding (and not necessarily what the trainer thinks that the dog should like).  Some dogs are not “foodies” but prefer a game as a reinforcer; others prefer to be petted, or hugged, while some dogs cringe at being hugged, and do not find pettings to be particularly reinforcing.

If the dog is a “foodie,” then find the treats that the dog really enjoys and use those only when training.  This ensures that the treats do not lose their value to the dog, but will be highly effective as a reinforcer.

Advertisements

Doggie Adolescence and Obedience Training: A Contradiction.

Okay, so I didn’t train Gretchen yesterday.  I had a big project due for a class and really didn’t have time to work with her.  One wouldn’t think she would have forgotten everything in the space of one day, however.

Well, I took her out in the yard after work for some training, and it was most frustrating.  She wouldn’t focus, but spent the whole time with her nose to the ground.  When I called her name in my “working” voice,  she slowly gave me her attention.  When I said “sit” she looked at me as though she had no clue what I was saying.

Then when I asked for a “down” she looked at me as though I had an arm growing out of my forehead.  Luring didn’t work; she just tried to pry the food out of my hand.  I resorted to the old fashioned pull-‘n-push method to get the sit, and another trick that I learned from a friend to get a down from a dog that won’t down.

I rarely use those physical methods because I feel that there are other methods that are effective; and there are some dogs that are touch-sensitive to begin with, and forcing a sit or a down from such a dog could result in a bite (remember, I train Rottweilers and German Shepherds).   Anyway, I broke my own rule because Gretchen knows these cues, and I’m not sure why she chose not to do them.  She’s 10 months old, so it’s possible that the virus called doggie adolescence has invaded her brain and wiped it clean.

One thing I learned today while training:  it’s important to have many approaches, techniques, and tricks in one’s arsenal of training because it gives you a choice of things to try when something doesn’t work.  Yes, I’m a die-hard operant-based clicker trainer, but sometimes luring, and capturing doesn’t work.

One must use common sense when training.  Capturing and luring have their place, and that’s my first choice when training a puppy or a dog that has little or no experience with training.  Once you have a dog with fluent behaviors that are on cue, it becomes tricky.  It’s up to the handler/trainer to figure out why the dog won’t perform a specific behavior.  Sometimes there’s a good reason, but at some point the dog needs to understand that performing a given cue is not an option.

Teaching the Down/Stay and Sit/Stay using Clicker Training

It isn’t hard to teach stays to a dog once it understands the difference between “sit” and “down”.  Last week Gretchen had difficulty differentiating the two, but I took her back to basics, and this week she improved.  It was time to incorporate the stays.

I teach sit and down stays concurrently, and have been successful because I concentrate on getting duration first BEFORE I ever try to add distance to the cue.  And, I start with very short durations, sometimes as short as half a second!  This is particularly important when training a young puppy that has a hard time being still.

Gretchen isn’t exactly a young pup:  she’s eight months, an adolescent dog.  Her impulse control has improved significantly, but she is still an enthusiastic learner, and as such has a hard time sitting still for long.  I start off with a half second stay then increase the durations by a second after that with a 1:1 ratio of reinforcement; that is, one click and a treat for every second that she remains in the sit and the down.

Once I had Gretchen successfully and consistently doing four-second stays, I put the behavior on a variable schedule of reinforcement.  That is, I varied when I clicked and treated instead of the usual 1:1 ratio.  So I might click after a two-second stay; next I’d click after a four-second stay and give a jackpot of treats; then I’d click for a half-second stay.  The dog would never know when she’d get her treats, or whether she’d get a jackpot of those yummy steak bits, so she’d hold her stays in eager anticipation.

It’s important when training stays that the trainer doesn’t ask the dog to hold the stay longer than it’s capable.  If I went from asking for four-second stays to asking for a thirty-second stays before I clicked and treated, Gretchen would be confused that I had not clicked sooner and think that she did something wrong. Because clicker training promotes thinking on the part of the learner, Gretchen would break the stay and offer me some other behavior thinking that’s what I really wanted.  Dogs are smart.

Teaching the “Stand” using Operant-Based Clicker Training

Gretchen is making great strides now that her training is a bit more consistent.  Her sits and downs are fluent, and she’s much more focused than the last few days.  Today we started, or rather, restarted learning the stand from a sit or down.

We started with the very basics,  and used the touch cue, which is nothing more than having the dog follow my hand.  I do not use a verbal cue to touch because by simply holding my hand out, palms flat and fingers straight, Gretchen knows that she’s to touch my hand with her nose.  She also knows to follow my hand wherever it goes.

So, to train the stand, here’s the approach that I took:

1)  Placed Gretchen in a sit at my left side.

2) Using my right hand with palms facing the dog, I moved my hand straight out from her face, but did not move too far.

3)  As soon as she got up, I clicked and delivered the treat.  The timing of the click is critical at this stage.  Ideally, I want Gretchen to stand up without moving forward.  Only her back end should move out, if she’s getting up from a sit.

For now, however, I keep my criteria very low to ensure her success at this exercise and to cut down on the likelihood that she’ll get frustrated and shut down.  A dog that has shut down, like a human, will not learn, and will dread future learning situations.

Right now, I’m only interested in her getting up.  I don’t mind if she moves around. The click ends the behavior and serves as a “release” signal anyway, so once I’ve clicked,  it doesn’t really matter what she does next.  Once she understands this concept, then I’ll raise the criteria slightly.

One Step Forward, Two Steps Back: An Inconsistent Approach to Dog Training

So tonight I took Gretchen into the yard to train her for the first time in many weeks.   Since my work schedule went crazy in early April, my poor dogs’ training suffered.  Now that my schedule has regained a bit of sanity, it’s now frightfully hot.  My training is still restricted, but this time to indoor work at night. That’s the joy of living in the tropics.

Although Gretchen has retained quite a bit of training, I’m appalled at how much she has left to learn!  Yikes.  She hasn’t learned the formal recall (her informal one is quite good, at least in my yard), and she hasn’t learned the heel, either.  She’s seven months old.

Gretchen does her sits and downs pretty well in the house.  She’ll sit out in the garden on cue, but is VERY reluctant to go into a down.

It is not uncommon for dogs to perform behaviors perfectly on cue in one location but not seem to understand the same cue in another.  An extreme example would be the dog that will give you a sit very quickly and enthusiastically at home, but once he’s at the vets or in some other new environment, he suddenly has no idea what sit means.

The dog is not being stubborn or “hard ears.”  The real reason for this annoying and sometimes embarrassing breakdown in training is simple:  dogs do not generalize.  “Sit” in your living room doesn’t necessarily mean “sit” in your kitchen, in your backyard, front yard, the vets, etc.

So, with Gretchen I took a deep breath, swallowed my pride, admitted that I am inconsistent with my work, and went back to basics.  When she was a babe of about five weeks, I taught her the down by capturing the behavior.  Now that she’s older and had already learned the down, I decided to root around in my training arsenal and try a different approach.  I tried luring.

Now I’m not a huge fan of luring because I feel that it adds more steps to the learning process, at least the way I use luring; and if it’s not done correctly, the dog could become dependent upon the food lure in your hand.  In order for luring to be effective in the long run, it is essential that the food lure be faded as quickly as possible within the training session.

As it turned out, I had to lure only twice before Gretchen caught on.  After two attempts I took away the food but continued using the downward motion of my hand (as though I had food in my hand), and added the “down” cue.  She performed beautifully and consistently.

Alas, it was late in the afternoon and the mosquitoes descended upon us, threatening to exsanguinate me.  I ended the session and returned to the house with my dog.

Operant Conditioning allows German Shepherds Retain Skills

The three dogs–Gretchen, Tuvok, and Athena–continue to make progress in their training, although I am so inconsistent in my work with them.  That’s the beauty with clicker training:  what you teach sticks!

This is the first week, since I’ve returned to work in January, that I’ve trained three days, and the week’s not yet over.  That’s huge progress.  The most I’ve ever trained is two days in a week.  Last year this time, however, I didn’t train any of my dogs between January and May.  I did a bit of work with Athena during the summer (when it wasn’t too hot), but she still managed to win the Novice trial in November, with High in Trial (99/100 points).

Athena now has a reliable 30-second stand-stay, and is doing the high and long jumps quite well.  The jump heights are under the required height for her class, but that’s okay for now.  She did two lovely recalls for me today, too.

Gretchen has a semi-reliable 30-second sit-stay.  She does this on leash.  She has an awesome down-stay, too.  I have started doing rear-end awareness exercises with her and her brother, Tuvok.  Tonight I had them walk through the rungs of a ladder that was laid out flat on the grass.

Gretchen’s loose-leash walking is weak, and I realize that I really need to focus on that skill now, while she’s five months old and still controllable.  Tuvok’s isn’t much better.  My problem is that I train late at night when going into the yard after dusk guarantees a feast for the vampires, I mean MOSQUITOS.  I train in the house at night and in the yard during the day (problem is I’m never home during the day).

I’ve started getting up early to train at least one dog before work.  Some mornings it just doesn’t happen.  Yet, I press on and hope for the best.

Taking an Intermediate Dog Back to Beginners

How many times have you trained in your living room, and your dog performed beautifully, but failed to obey when you took him out in public? 

There are several reasons for that.  Perhaps the dog is over-aroused, distracted, doesn’t find the cue convenient to execute at that precise moment (you are probably trying to get him to down near poop, or something equally disgusting), or he just may not know the cue properly.

Yesterday Tuvok and Gretchen went to SAR training then to agility.  Tuvok is my white German Shepherd puppy, whom I’m hoping will become one of Jamaica’s first SAR dogs.  He performed his sits and downs admirably.  Today when I worked him in the yard, which I rarely do, he just would not down. 

I took him inside and lured the down once more for about three times before fading the lure.  I am not a big fan of luring, but there are times when it is appropriate.  Tuvok knows his downs in the house, but for some reason didn’t want to do them in my yard.  

Once he got the downs fluently without the lure in the house, I took him outside once more.  He would not down.  I stayed outside, but lured the down once more for about three times, then I faded the lure and asked for the down with the hand signal.  He performed this time.

Sometimes when a dog doesn’t do a behavior on cue in a strange location, you have to go back to baby steps.  Although the back yard isn’t strange to Tuvok, he finds lying on dirt and grass strange because he’s accustomed to lying on tiles and wood floors in doors.  He is an indoor dog.  As he will be a working dog, it is vital that he complies with cues immediately when they are given to him.

Tag Cloud

%d bloggers like this: