with No Comments

Post No.: 0400rewarding

 

Furrywisepuppy says:

 

Shall we talk a little about training pooches now?

 

Fluffystealthkitten says:

 

Let’s do it!

 

Furrywisepuppy says:

 

Dogs, and cats, respond well to conditioning – specifically positive reinforcement training if the rewards are delivered more-or-less immediately after each desired behaviour is performed, and if they’re delivered reasonably consistently. This means rewarding, in a predictable manner, an animal for demonstrating a desired behaviour. Even fear-related behaviours may be overcome with such training.

 

Fluffystealthkitten says:

 

A reward must be delivered in a timely manner just after a dog performs a desired behaviour otherwise the dog won’t associate the two events together. So you cannot say, “This reward is for that thing you did ten seconds ago” because the dog won’t understand your words and will think the reward is for something they did just a second or two ago. The reward signal must be clear and unambiguous (e.g. not a half-arsed pat or praise delivered with an indifferent tone). And a reward must be consistently applied after each occasion the dog demonstrates a desired behaviour otherwise again there’ll be no association made between the behaviour and the consequence, and the link will seem random and meaningless rather than causal.

 

Furrywisepuppy says:

 

Well perhaps a ‘rough consistency’ might be the absolute optimal rate because it’s currently unknown whether dogs, or cats, respond best of all to a slightly variable reinforcement schedule, like humans do, as exploited by gambling machines, which don’t reward players every single time they play? Dog and cat-specific research will need to be done on this. However, most people won’t be perfectly rewarding their pets literally every single time they do something desirable anyway (e.g. some good behaviour will have gone unnoticed) hence the general advice is to maintain consistency.

 

Fluffystealthkitten says:

 

In Post No.: 0376, I mentioned that if an animal can learn then it’ll increase the repetition of any behaviour that comes with a rewarding outcome, and decrease the repetition of any behaviour that comes with a punishing outcome.

 

But rewarding desired behaviours rather than punishing undesired behaviours teaches a dog or cat what she/he should do, which is more important than teaching a dog or cat what she/he shouldn’t do. If we’re doing a good thing then we’re logically not doing a bad thing, but if we’re not doing a bad thing then it still hasn’t necessarily clarified to us what we should be doing instead – we might do another bad thing? It’s analogous to telling people not to sell drugs to make a living – so they go steal cars instead. But if you can pay them well to do some cooking for you then they’ll seriously consider cooking for a living instead (i.e. provide decent and stable jobs if you want to reduce such crimes).

 

Furrywisepuppy says:

 

That’s right. Whether with humans, any other animal that can learn and even regarding machine learning – get into the mindset of considering ‘what do I want him/her/it to actually do?’ or ‘where would I rather him/her/it be?’ rather than thinking ‘don’t do that’ or ‘get away from there’. It’s lazy and ineffectual leadership to lead via ‘don’ts’ and to not proactively support the ‘dos’. The ‘dos’ tend to elicit rewards while the ‘don’ts’ tend to elicit punishments too.

 

Hence rewarding positive behaviours is what you should primarily do if you are training animals. Consistency and predictability, including routines, are important for the sense of security for both cats and dogs too, and without them there’s no learning between cause and effect because it’ll all seem random (e.g. ‘every time I sit on command, I get a treat, pat, stroke or praise, so I’m going to want to do it again’ rather than ‘I’m not sure if sitting on command or not makes a difference to anyone so I don’t care if I sit or not when I hear a ‘sit’ command’).

 

Plus with positive reinforcement – from the animal’s perspective it hopefully won’t be about ‘doing as the master says just because the master says so’ but analogous to enjoying playing a videogame where one collects coins or rings and can’t get enough of doing so, so much that it’s two-way – the animal will be excited to play and will often ask to be trained! Dogs end up happier and more cooperative with reward-based training compared to fearful and coerced with punishment-based training. The latter would also predispose a dog to more aggressive behaviours for fear is at the root of all aggression. Humans often misconstrue this aggression as ‘a struggle for dominance’ but it’s the complete opposite because it’s an expression of anxiety. It’s not nice for us. Woof.

 

Fluffystealthkitten says:

 

Now threats and punishments can work to get cats or dogs to do what you want but you need to question what sort of relationship you want with your pet – one ruled though fear like slaves are, or one through love like companions? And do you really want to punish them for feeling basically stressed or frightened (e.g. if a dog urinates in the house out of anxiety)? This’ll only make them even more anxious, thus it creates a vicious cycle. It’s all the same advice and reasons as with raising human children primarily with positive reinforcement.

 

Furrywisepuppy says:

 

Also, when attempting to use threats or punishments to dissuade a behaviour, such as getting a dog to not eat the cat’s food – such punishments would need to be more severe than the food is delicious to them too. So it’s better to redirect their attention to desired behaviours by rewarding desired behaviours, such as rewarding the dog when he/she doesn’t enter the cat’s feeding place when the cat’s food is out. It’s like it needs a great deal of punishment to discourage lucrative financial crimes, to make the risk-to-reward not worth any illegal behaviour. (Albeit here it’s hard to think of how to reward the desired behaviours even more because working in the financial sector is already potentially highly financially rewarding if one does one’s job in a legal manner. But for some, a lot is still not enough.)

 

We dogs can be trained to respond to vocal commands. We can also understand it when a human points to something or somewhere. People can build effective communication strategies by learning about the individual personality, body language and vocalisations of their dog, and through consistent and positive training that builds trust with them. If people invest in ensuring proper socialisation and training, it’ll make for well-mannered, as well as happy, dogs; which will make for happy humans too! Like with children and parenting, we largely reap what we sow, or what we reward in this case (e.g. rewarding calmness with attention and maybe a treat, rather than ignoring it and only paying attention to misbehaviours – which is an error that most people instinctively make because it’s more natural to leave alone what’s going right and to only attend to what’s going wrong because these pose possible hazards).

 

Fluffystealthkitten says:

 

You could be inadvertently rewarding undesirable behaviours, such as interacting with your dog every time she/he barks at a stranger. You may be saying, “Stop” with stern eye contact but the dog perceives it as ‘whenever I bark at strangers, I get rewarded with my owner’s attention, so I’ve been taught to keep doing it – although it can be quite confusing because I also sometimes get forcefully yanked away from these strangers’.

 

Furrywisepuppy says:

 

If you don’t want your dog jumping up on people with over-excitement all of the time then turn away to the side, hands tucked in with no words or eye contact whenever he/she jumps on you, and importantly reward him/her with attention and interaction whenever he/she is on all fours and is calm.

 

Fluffystealthkitten says:

 

In general, ignore what you don’t want or provide a distraction, and reward what you do want – you may need to be very persistent and patient but the dog will eventually learn the pattern. It’ll take longer the longer you’ve been rewarding and reinforcing their undesired behaviours, like humans trying to reverse a bad habit that’s been ingrained for years.

 

Furrywisepuppy says:

 

Yep. And like with children – lead by example. In this case, if you’re always agitated and loud then your dog will read your body language and be constantly anxious about why you’re seemingly constantly anxious. Be consistent – predictability and good routines reduce stress and the fear of the unknown. And build up a good relationship with your dog so that if he/she gets anxious then he/she will know that you are his/her safe place for comfort and support that he/she can always return to, rather than becomes a dog who attempts to tackle every perceived threat on his/her own or runs away.

 

Fluffystealthkitten says:

 

The appropriate training is thus important for the dog, as well as for building the human-dog relationship. Do note though that dogs, and cats, are constantly learning even when you’re not formally training them (e.g. like with children, people tend to never praise a dog for just being quiet and calm but instead consistently reward their boisterous behaviours with attention, eye and physical contact and vocal responses, which only reinforces those boisterous behaviours!)

 

Furrywisepuppy says:

 

Training and learning can only effectively happen when an animal is calm and relaxed too. Like with humans during a stressful ‘fight or flight’ state, a dog or cat will be essentially tunnel-visioned on dealing with the present threat or pain rather than being open-minded for learning. We must logically be open-minded to effectively learn new things. This is another reason why learning is far less effective via threat or punishment-based training. You can interact with, play with, treat or use tactile contact to reward a dog for doing something desirable.

 

Relax the tension in a game of tug-of-war if you want a dog to release an object from its jaws – most people intuitively try to pull harder to get a dog to release an object but that only treats it even more as an enjoyable game for them! So if you want a dog to loosen his/her bite on an object – yield a little tension yourself first.

 

Fluffystealthkitten says:

 

Always bear in mind that if an animal doesn’t seem to be behaving then it’s always the fault of the teacher – for using inappropriate techniques, for having unrealistic expectations and/or for not having the animal checked to see if she/he has any health problems that might be affecting her/his learning.

 

Furrywisepuppy says:

 

Some people believe that raising a child is, or should be, nothing like raising a puppy, or vice-versa, for it’s the politically correct thing to think and say. But, although there are differences (e.g. there’s no need to take your child outside to pee and poo in the garden!) – there are some very key similarities, such as reinforcing or extinguishing behaviours via conditioning (particularly by rewarding desired behaviours rather than punishing undesired behaviours), the importance of routines, consistent treatment, not over or under-feeding them, and of course the love and affection.

 

Maybe some people are thinking about the harsh punishments dealt to dogs when they misbehave ‘to show them who’s boss’ – but I hope that everyone knows now that this isn’t the best way to treat dogs either. Others may think that spoiling a dog is fine but not a child – but again a dog shouldn’t really be spoilt either.

 

Extrinsic motivations, such as incentivising children to do something only because they’re going to receive a treat for doing it could become a problem, but hopefully they’ll be gradually weaned off the treats and still be left with the feeling of joy being associated with doing a particular desired activity.

 

A pet is legally considered more like property, unlike a child, but many people believe a pet should be treated as simply another member of the family too.

 

Fluffystealthkitten says:

 

And that’s an agreeable notion I’ll say. Meow!

 

Furrywisepuppy says:

 

Ditto that. Woof!

 

Comment on this post by replying to this tweet:

 

Furrywisepuppy25

 

Share this post