Cats vs Dogs? Let's make an AI to settle this: Crash Course AI #19

[@BlackVoidsRevenge]

This show is great! Jabril is such a great host! I love his laid back approach!

[@CairoQuinn]

What a great way to teach about ai and data!

[@zaneearldufour]

I Love that you guys put up colab notebooks for these videos!

[@enaszaq3051]

Honestly, both pets are great but the type of person you are is the main thing in the decision. If you are an extrovert and pretty energetic then a dog is best but if you are an introvert and a really calm or laidback person then a cat is best. Though some cats will act like dogs and vise versa, so there's that but you get the point.

[@deathstroke2697]

Dog-lovers base their whole case on these commonplace, servile, and plebeian qualities, and amusingly judge the intelligence of a pet by its degree of conformity to their own wishes. The dog appeals to cheap and facile emotions; the cat to the deepest founts of imagination and cosmic perception in the human mind.

[@akemi2929]

Wow... I never thought of it this way... 😯

[@Arahathan]

I’d still prefer a pet with more “degrees of conformity” since it’s, you know, a pet ;) If I want more than that, I’d turn to people, not animals.

[@AverytheCubanAmerican]

I love dogs, my favorite breeds are the Havanese and Shiba Inus

[@makeracistsafraidagain]

Avery the Cuban-American Bichon only love. They do nothing else.

[@vennymeow]

I have a havanese , he is next to me rn.

[@thecontroversialgamer7079]

>insert "Why Not Both Taco Meme"

[@DarthSivius]

Getting a representative sample is statistics 101 and has nothing to do with AI. This really should be video 1 and not the last one.

[@HornetCucumber]

Some researchers say that we find cats so attractive and cute because they resembles our human babies. *That is where that hairy babies joke comes from.*

[@SupremeLeaderKimJong-un]

I love dogs, my people love them too

[@smokeybrown8752]

What about cats

[@alfajorcin]

They dont taste the same, cats. But i like rats they spicy with peppers

[@chattingesque372]

You racist bastard. Cats are a CHINESE delicacy.

[@wildchicken679]

Chattingesque Mmm China.

[@lilimarlene7813]

Jabril needs to include a poop variable in addition to energetic, cuddly, etc. “Do you enjoy taking walks several times each day regardless of weather conditions?” “Are you willing to scoop clumps of pee and poop from boxes distributed around your residence on a daily basis?”

[@danieldanielson2650]

Statistically speaking dog owners enjoy better health and longer life because they are forced to go for a walk several times a day.

[@wildchicken679]

Dogs pros: -Usually more friendly and playful which can make it very easy to boost your happy hormones (endorphins) -Walks and playing make good exercise for both owner and dog so you both feel great after a play session. Cons: A lot more maintenance; grooming, walking, bathing, etc Needy, many dogs get depressed when left alone for a while and constantly want attention. Cats Pros: -Self maintained, they self groom and clean themselves. Litter boxes make it so they can go to the bathroom on their own and they can fair well alone. -They are not as noisy for the most part and you can play indoors due to their size (yes there are dogs that are small but theres never going to be a domestic cat the size of a medium to big dog) Cons: -Their introverted nature often means they’re not available 24/7 so if you catch them on a bad day they might want to bite and scratch -Hard to get em to go on walks with you. They really like their independence. Summary: With dogs you have a clingy needy friend. With cats you ARE the clingy needy friend.

[@mami-rh6zv]

c a n i g e t a d a m n h u g ? U i have a big backyard so my dog can run around also i rarely have to bathe my dog soo

[@wildchicken679]

mami Correct. I also wouldn’t need to groom if I had a bald dog. I guess its a generalization.

[@drsingingeagle]

Dude! Just replace “energetic” with “playful.” Would that work?

[@TenguWolf]

That would depend on if you are changing the definition of the feature along with changing the label from "energetic" to "playful", like how would you define "playful" differently from "energetic"? Of course if you change the definition, you would have to collect new data using this new definition. If you don't change the definition of the feature, then it won't change anything at all :)

[@drsingingeagle]

@Tengux - Splendiferous explanation. Thanks! .^__^.

[@Moddie1989]

Jabril, just take the Pikacat, that solves the AI and any power problems you have.

[@ianrbuck]

If you want to learn more about how AIs feel about cats, I highly recommend the novel Catfishing on CatNet!

[@hanneschalfork7306]

Was that a Django reference at the beginning?

[@torimallory9499]

I would get a dog🐶

[@jhonnyrodriguezolguin3760]

Merry Christmas wonderful teachers

[@C-Jenkins]

Did you get a cat or dog?

[@torimallory9499]

Can you make more of ai videos

[@voidpuppy5145]

Whats up with those shoes man ?😂

[@6500s1]

dogs are annoying, cats are chill, end of topic. I won't even start talking about unprovoked lethal attack statistics in home environment.

[@ew__2]

what shoes is he wearing?

[@robert4445]

Does PCA prior to feeding into an MLP help prevent feature correlation?

[@TenguWolf]

Sadly, no. The PCA will simply hook on to the exact same correlation as the model in the video example did. Like: "Oh! If I just look at this one feature, I can possibly just guess straight 'No' with a super high success-rate. Ez-game!". Although, you could potentially catch the problem before it got to training the model, because I would imagine that doing a PCA on his dataset would show "Energetic" as the first and most significant component of the dataset by a huuuge margin (like much larger than you would normally expect), at which point your alarm bells should start ringing.

[@robert4445]

@@TenguWolf It just seems like this isn't a problem which requires a "human look". Like how regularization prevents overfitting by penalizing higher order models. It's hard to think there isn't an algorithm which penalizes the fallacy we see with the "energetic" feature. And if there isn't an algorithm like that out there I dont see why there can't be.

[@TenguWolf]

@@robert4445 To my knowledge there isn't a particular algorithm for this. I think the closest thing would be to further describe the categories involved in your model and then have a function to tell you that the distribution is off by a great amount. I doubt something like regularization would work against this, as regularization is done to kinda 'smooth' out the line between previously seen cases and future cases. If you penalized the behaviour of relying heavily on a single feature for predictions, you would just be handicapping the model, based on all the available information the machine has, this one feature will largerly give it the correct answer as shown by the example in the video, why wouldn't you want the model to take advantage of that? The solution would more likely be to simply provide more accurate and uniformly distributed information/data to the machine.

Jesus i felt like I was back at c++ class. How dare you?!

[@arif7479]

What do u use to write code?

[@arif7479]

@@Corgi_fax I mean what IDE or text editor or jupytor etc

[@TenguWolf]

Because the code is available on Google Colab, the obvious answer would be that he is writing the code using Python in an Ipython Jupyter Notebook, as the format is exactly the same as what you see in Google Colab and is a popular choice for coding when working with machine learning. Simply look up Jupyter Notebook on Google if you are interested.

[@arif7479]

@@TenguWolf thanks. Jupiter is confusing for me but I will get it.

[@Undiscovered1st]

I will save you the work get a cat.

[@wildchicken679]

Compton Alleyne Cats all the way.

[@Undiscovered1st]

c a n i g e t a d a m n h u g ? Lol a toy poodle will cover both a cat and a dog.

[@wildchicken679]

Compton Alleyne Im assuming you’re either an eccentric electric engineer Or you like playing with hex bug toys

[@carlnikolov]

Flip a coin, heads get a dog, tails get a dog

[@masaakunokouchi]

Cool video.

[@martisleister9197]

Get a Maine Coon. That's a cat who actually likes you. (If you need to have that from a pet.

[@alankent]

Interesting video. The hat is silly.

[@Thomahawk1234]

Next time put jabrils in the thumbnail! I almost didn't click this video.

[@kikinatrone]

Bruh you want a Dog. Get a dog. Don't change the features you wanted in the beginning.

[@gotto8386]

Furry Baby!... So is Baymax related to John green bot

[@harleyquinn5774]

Birds! 🐔🐧🐦🐤🐣🦅🦆🐥

[@hopefuldavid3320]

Ur data is wrong. Cats are softer, and are equally energetic and cuddly. And more quiet because dogs bark to much

[@smokeybrown8752]

But they pee on your stuff

[@hopefuldavid3320]

@@smokeybrown8752 Mine always pees in the litterbox.

[@aidalove9411]

@@smokeybrown8752 I've seen dogs pee inside the house more often than cats, primarily when they get excited. I have no pets but many of my friends have both

[@wildchicken679]

Smokey Brown Dogs pee on your stuff.

[@mami-rh6zv]

HopefulDavid my dog never barks

[@cierrabird6463]

Nice

[@charley2556]

Cat people we Stan

[@awesomeproductreviews9962]

All animals have different personalities. There are some cats that behave just like a friendly dog. You should be asking what pet you could give a better home to. Do you have room for a dog, time to walk it, money to care for it, etc. You can leave a cat alone a couple days if you need to leave town by giving it enough food and water, but you cannot do that with a dog. Do you have someone to care for the pet if something happens? Animals have a personality all their own and it also has a lot to do with the person. Do you show him/her a lot of affection or ignore him/her, etc. Dude, you didn't even give us the answer LOL. Bummer. Thumbs down :(

[@tonytins]

Haiirrry... *hic* BABY!

[@ricjackson2841]

So how does this work with election polls? 🤔😂🤷🏾‍♂️👍🏾

[@V01DG0D]

get a dog because in a disaster scenario it’s more meat to eat

[@network7689]

Wondering savior

[@magister343]

Birds are better than turtles and turtles are better than cats or dogs.

[@lilimarlene7813]

You’ll never get a group of cats to pull a stupid sled across the snowy Alaskan tundra in below zero weather. 🐈 + 🛷 = 🤮

[@vixxcelacea2778]

You might be able to get trained bob cats to do so actually. Thing is, wolves wouldn't do that job either. Wild animals don't listen to humans. Cats are still wild comparatively to dogs, they hold and act on more instincts. Dogs are also pack animals, house cats are colony animals. That said, both animals absolutely have their uses and can be wonderful companions. For people with need for a support animal, a dog is a lot better because they can be trained to help with anxiety attacks, walking the blind etc, cats not so much. But cats can be great emotional support animals, they often help autistic children in particular. Horses, rabbits etc also good for emotional and mental health for people. it really depends on the person. I think it's truly apples to oranges and depends on a lot of factors on what animal is a good fit. If nothing else, why not have both?

[@roywastaken]

Dogs express their emotions better than Cats. My cat always looks indifferent about everything!

[@danieldanielson2650]

Probably because it IS indifferent about everything (but food) 😅

[@malia8693]

Roy Tries Tech no!

[@IceMetalPunk]

The answer to "should I get a cat or a dog?" is just always "dog". Puppers 4 lyfe!

[@IceMetalPunk]

@@Corgi_fax Unfortunately, I don't have a dog, because I work full time and would feel bad leaving a pup alone 8 hours a day. But why would you feel sorry for my dog just because I like dogs more than cats?

[@malia8693]

No!

[@appletree8441]

Dogs taste better

[@iclaimthisname9767]

Cats

[@drsingingeagle]

Anybody else feel strangely dumber after watching this? Even with two PhDs, I can barely keep up. >__

[@DarthSivius]

Honestly, makes me wonder what your phds are in and where you got them...

[@drsingingeagle]

@ DarthSivius - Truly worthless fields in today's "market," anthropology and English. UC Santa Barbara (go, Gauchos!). Soon I'll go back to study astrophysics.

[@DarthSivius]

@@drsingingeagle Good luck with that.

[@drsingingeagle]

@ DarthSivius - Thank you, and merry Christmas!

[@annagarza8649]

Yee haw

[@attq3980]

Dogs pee and poo on everything and then smell and eat it.

[@grusbekse]

Cats pee and poo on everything and then let it lay around for you to find and smell.

[@dangerx69]

@@grusbekse 🤣🤣🤣

[@jaykatz9785]

Cats give you toxoplasmosis

[@makeracistsafraidagain]

attq You are not good with pets.

[@oddbigking9069]

Cats are cute

[@Kalaphant]

the way you're talking is scary...

[@rightjb]

Umm. That's not what AI is. Using a simple predictive program from a data set to completely eliminate one of your original criteria (changing your entire success standard) is called, "HUMAN LEARNING." So congratulations on showing how the human brain and non-artificial intelligence works.

[@rightjb]

@@Corgi_fax then say that. Don't mislead newbies as you are here.