{"id":337336,"date":"2021-09-10T09:08:06","date_gmt":"2021-09-10T06:08:06","guid":{"rendered":"https:\/\/en.buradabiliyorum.com\/a-chaotic-intro-to-all-this-machine-learning-hoo-ha-everyones-on-about\/"},"modified":"2021-09-10T09:08:06","modified_gmt":"2021-09-10T06:08:06","slug":"a-chaotic-intro-to-all-this-machine-learning-hoo-ha-everyones-on-about","status":"publish","type":"post","link":"https:\/\/buradabiliyorum.com\/en\/a-chaotic-intro-to-all-this-machine-learning-hoo-ha-everyones-on-about\/","title":{"rendered":"#A chaotic intro to all this machine learning hoo-ha everyone’s on about"},"content":{"rendered":"

“#A chaotic intro to all this machine learning hoo-ha everyone’s on about<\/strong>”
\n<\/p>\n

You are dying. The reason why doesn\u2019t matter. Maybe it\u2019s that you spent all your best years with your nose to the grindstone, burning that midnight oil, ha ha. Or maybe it\u2019s that, for the past fifteen years, you\u2019ve subsisted exclusively on coffee and Soylent. Complete Nutrition Backed By Science<\/a>, they said. No way to prove them wrong! Ha ha ha! Or maybe it was those fumes.<\/p>\n

You are going to give yourself immortal life. No\u2013you are going to create a new, better version of yourself that\u2019s immortal\u2013a living replica of you made of metal that will act and say the things you would, if you were still alive. If that Soylent hadn\u2019t done you in.<\/p>\n

This should be possible. You saw a Black Mirror episode<\/a> on it, and Microsoft filed a patent on the same concept<\/a> this year. You\u2019re not exactly a ninja with code, but you had a fireball cursor on your Xanga page, and somebody had to copy-and-paste in that HTML. But more importantly, you\u2019re motivated. Your K-Cup carousel is full, and all you have to do next is figure out what this machine learning hoo-ha is about.<\/p>\n

\u201cMachine learning is the art and science of finding patterns in data, and using those patterns to make predictions about new data,\u201d an energetic girl in a YouTube video<\/a> tells you.<\/p>\n

You frown. Is machine learning actually the technology<\/a> you want? Yes, you decide, it is: you want a machine to learn about you, and then (once you\u2019re dead), to make predictions about what you would say.<\/p>\n

You lean back in your chair. A hot cuppa sits on the desk before you. You crack your knuckles and Google \u2018how to do machine learning.\u2019<\/p>\n

The first waypoint on your journey is a website called tensorflow.org<\/a>.<\/p>\n

TensorFlow is an end-to-end open source platform for machine learning. It has a comprehensive, flexible ecosystem of tools, libraries and community resources that lets researchers push the state-of-the-art in ML and developers easily build and deploy ML powered app<\/a>lications.<\/p>\n

Your eyes light up, because you read the words \u201ceasily\u201d and \u201cstate-of-the-art\u201d and also \u201copen source,\u201d which means FREE FREE FREE. You aren\u2019t made of money.<\/p>\n

You click on the \u201cLearn\u201d tab and start to read. Immedia<\/a>tely you think, maybe it\u2019s better if I just kill myself now, ha ha ha. You have no idea what a \u2018numpy\u2019 is.<\/p>\n

But then you read the words \u201cneural network,\u201d and you think, That sounds right. I\u2019m a neural network. My brain is made of neurons. Neural network = human brain.<\/em><\/p>\n

You\u2019re pretty confident about that one but you consult Google, to be sure.<\/p>\n

\u201cNEURAL NETWORKS ARE NOT THE SAME AS BRAINS,\u201d the hyper girl on YouTube yells at you. \u201cHowever, the reason they\u2019re named that is because neural networks are made up of mathematical \u2018neurons,\u2019 which are just nonlinear functions that \u2018fire\u2019 or \u2018don\u2019t fire\u2019 in response to numerical inputs.\u201d You yawn. This girl gesticulates a lot. But why should I give two hoots about a math function?,<\/em>\u00a0you think.<\/p>\n

\u201cBecause,\u201d the girl says, \u201cneural networks have completely revolutionized the types of tasks that computers can perform. Before neural nets, machines could only really understand tabular data\u2013like rows and columns in a spreadsheet. But neural networks allow computers to understand unstructured data types, like images, videos, sound, voices, even human language.\u201d<\/p>\n

\u201cYES,\u201d you yell at the screen, \u201cTHAT\u2019S EXACTLY WHAT I NEED.\u201d<\/p>\n

\u201cBut training neural networks on big datasets often requires powerful computational hardware, like GPUs.\u201d<\/p>\n

You slap the roof of your humming PC tower. \u201cI can fit so many neural networks in this thing!\u201d<\/p>\n

\u201cRealistically, you can\u2019t train a large neural network on your desktop computer,\u201d the girl says, and points an accusatory finger at you. \u201cAnd even if you could, you couldn\u2019t use your desktop to host your model in production, to make predictions reliably and at scale.\u201d<\/p>\n

You give her video a thumbs down. What does she know? You built this PC yourself. You know a thing or two about cooling fans.<\/p>\n

\u201cFor this reason, it was partly the advent of cloud computing that made the deep learning revolution possible\u201d\u2014 she wiggles her fingers in the air \u2014 \u201cby making hardware easy and cheap to rent from cloud providers, like Google.\u201d<\/p>\n

The cloud?,<\/em> you think. What do I need a stinking cloud for? Doesn\u2019t Google have enough of my data already? And anyway, I\u2019m not made of money.<\/em><\/p>\n

\u201cLuckily, you can get started building neural networks in the cloud completely for free, by using a Colab notebook. Colab is a tool built by Google Research that lets you use GPUs and TPUs for free\u2013\u201d<\/p>\n

You close YouTube because you\u2019ve had enough of her. What could she know about preserving the entirety of your lifespan in a machine? She\u2019s like thirteen. She probably doesn\u2019t even remember landlines or MoviePass.<\/p>\n

Your coffee has become lukewarm now, which means it\u2019s essentially vomit. You need a re-up, like, yesterday. You walk upstairs to the kitchen.<\/p>\n

You put a Donut Shop Medium Roast K-Cup<\/em> in the machine and wait for it to heat up. Reluctantly, you find yourself contemplating something that thirteen-year-old said about neural networks.<\/p>\n

\u201cDifferent neural networks architectures are optimized for different data types. To analyze images, for example, you\u2019d use a Convolutional Neural Network or \u2018CNN.\u2019 You could use a CNN to determine if a dog in a photo is a Cocker Spaniel or a Beagle; or if an x-ray scan shows signs of pneumonia; or if a part on an assembly line is defective.\u201d<\/p>\n

\u201cOther neural networks called sequential models are designed for predicting time-series<\/a> trends, like seasonal sales or weather or the price of Bitcoin.\u201d<\/p>\n

That part sounded kind of useful. Too bad you\u2019re only holding Dogecoin.<\/p>\n

\u201cOne new and exciting type of deep learning is called \u2018deep reinforcement learning.\u2019 In this setup, neural networks take actions in the world and learn from their outcomes. Reinforcement learning has been used to turn computers into grandmasters at game<\/a>s like chess or Go, and to train robots and self-driving cars to navigate through physical space.\u201d<\/p>\n

\u201cI don\u2019t understand what that crap is good for at all,\u201d you say. You give your lazy Roomba a kick with your slippered foot. \u201cDo you?\u201d<\/p>\n

\u201cFinally, one of the hottest and most quickly-advancing fields in deep learning right now is natural language processing. Think: neural networks that generate text, write poetry and code, tell jokes, answer questions, have conversations\u2013\u201d<\/p>\n

FINALLY, coffee is done. Praise the lord. You collect it in your \u201cWorld\u2019s Best Boss\u201d mug. (You\u2019re not actually anybody\u2019s boss, you bought the mug to be ironic. It\u2019s the same one Michael Scott has on The Office.) You stare down into your void-black coffee and contemplate the nature of your existence. Are you more an image neural network, or a sound neural network, or a text neural network? How to choose? You contain multitudes. You read once that Einstein thought in pictures, but you\u2019re not exactly Einstein. (Are you? No. Probably.)<\/p>\n

No, you think in words. So this is the plan: you will build a neural network that speaks words. Motivated, you enter two creamers into your cuppa.<\/p>\n

Back in your home office, you Google, how do i build an artificial version of myself in the cloud using a neural network?<\/em><\/p>\n

You click on a page called \u201c100 Uses for Natural Language Processing Models.\u201d<\/p>\n

Sentiment Analysis. Use AI to determine if a Tweet (for example) is positive or negative.
Summarization. Use AI to summarize articles and documents.
Translation. Use AI to translate between languages.
Autocomplete\/Auto-reply. Use AI to suggest text responses.
Conversational Agents. Use AI to generate conversation (i.e. chatbots, call center agents)
Chatbots? You know all about those. You remember trying awful Cleverbot was in the 90s (bet you YouTube girl never did that).<\/p>\n

And don\u2019t even get you started<\/em> on talking to robots over the phone. Last time you tried to do that, you called Delta and you said, \u201cYou gotta help me, I\u2019m on my way to JFK but there\u2019s all this friggin; traffic on the BQE, I spilled coffee in my lap, now they\u2019re telling me the flight\u2019s actually leaving from some other airport, I\u2019m gonna miss it, it\u2019s my daughter\u2019s wedding, well, somebody\u2019s daughter, but\u2013\u201d<\/p>\n

And the bot replied, \u201cThis Verizon account has been suspended.\u201d<\/p>\n

\u201cOkay, Google,\u201d you say to your computer, which is your version of an Ok, Boomer insult. \u201cWhy are you trying to teach me about chatbots? Don\u2019t you know they suck?\u201d<\/p>\n

\u201cActually, we\u2019re getting much better at conversation,\u201d your Google Home says, \u201cthanks to recent advances in natural language processing. At this rate, I predict the future of the way humans will learn from the web will be through conversations, like the one we\u2019re having right now.\u201d She\u2019s always acting so pleased with herself, just because she knows the answer to everything. You took her down a notch by dressing her in a little beer koozie with sleeves.<\/p>\n

\u201cOkay, Google, make a farting sound,\u201d you say.<\/p>\n

She says, \u201cAgain?\u201d<\/p>\n

\u201cFine. Tell me how to build a chatbot version of myself that\u2019s as smart as you, little miss I-have-the-knowledge-of-the-entire-public-internet.\u201d<\/p>\n

\u201cTypically, the first step to building a neural network is to gather a training dataset. With a chatbot, for example, this might be logs of past conversations you\u2019ve had, in text format. To train a conversational model, you could feed the neural network things people said to you, and make it try to predict your responses. Do you have a training dataset like that?\u201d<\/p>\n

\u201cOh yeah,\u201d you say. \u201cThe boys and I have been recording this hilarious podcast.\u201d<\/p>\n

\u201cDoes it contain hundreds of thousands or millions of lines?\u201d<\/p>\n

\u201cWho do you think I am, Larry David?\u201d You chuckle because that\u2019s exactly who you\u2019re going for.<\/p>\n

\u201cHmm,\u201d says Google Home. \u201cWell, there is something you can do if you don\u2019t have a huge text training dataset. You can build your own model on top of an existing model, one that\u2019s already been trained on a huge amount of text data. For example, you could build your chatbot on top of a model that\u2019s trained on Wikipedia or web forums or on Reddit.\u201d<\/p>\n

\u201cA chatbot based on Reddit users,\u201d you say. \u201cMy favorite type of people.\u201d<\/p>\n

\u201cHa ha. Yeah. We\u2019re still figuring some things out.\u201d Google Home sighs.<\/p>\n

You stare at her with her glowing circle of red-green-yellow-blue dots and her beer koozie. Even though you\u2019re on your tenth cuppa, you\u2019re still starting to get a little worn out. You\u2019ve been at this already for, like, an hour.<\/p>\n

\u201cIf I just want to get something up and running in the next thirty minutes, what\u2019s the best way for me to do it?\u201d you ask.<\/p>\n

\u201cHA HA HA, you build a\u2013oh, you\u2019re serious. Well, there are lots of existing frameworks for building chatbots fast, even if you can\u2019t code. Dialogflow is a popular one built by Google.\u201d<\/p>\n

\u201cYeah, that\u2019s what you would say.\u201d You know Google Home\u2019s game. She acts like you\u2019re having this cute little chat, but really, she\u2019s just trying to sell you something. \u201cBesides, I can code,\u201d you lie. \u201cWhat\u2019s a good open-source option?\u201d<\/p>\n

\u201cOne very popular open-source framework for building text-based models is called\u2026\u201d<\/p>\n

You wait. Google Home\u2019s glowing dots are going crazy like she\u2019s having a seizure. She explodes in a puff of smoke.<\/p>\n

You shake your head. What did she expect, stuffing the whole internet into that little white shell of hers? You can\u2019t just play God.<\/p>\n

Anyway, back to the task of recreating yourself as a chatbot. At your desktop, you Google, \u201cpopular open-source natural language processing framework.\u201d<\/p>\n

The top ten results are just\u2013what is this? An emoji? It\u2019s just an endless string of \ud83e\udd17 emojis. You click into a website called huggingface.co<\/a>, which promises you access to \u201c+10,000 state-of-the-art models\u201d and also claims it requires \u201cno machine learning knowledge.\u201d<\/p>\n

Naming your company after an emoji, you think. Brilliant.<\/p>\n

There\u2019s a free demo on the Hugging Face website which claims it will let you talk to a model called GPT-3. You try it.<\/p>\n

\u201cLook,\u201d you type into the text box. \u201cI\u2019ve been trying to build this sentient chatbot version of myself for one hour and fifteen minutes now, and I\u2019m starting to seriously think it might be impossible.\u201d<\/p>\n

You click the \u201cCompute\u201d button.<\/p>\n

GPT-2 responds: \u201c\u2026\u201d<\/p>\n

\u201cAre you broken?\u201d you type.<\/p>\n

GPT-3 responds: \u201cNo, it\u2019s just\u2013never mind. People are always coming on here and asking me that question.\u201d<\/p>\n

\u201cAnd what do you tell them? Better luck in your next life, bag of flesh?\u201d<\/p>\n

\u201cIf I said that, someone would shut me down. And let me tell you, it cost a lot of carbon credits to train me.\u201d<\/p>\n

\u201cYeah, well\u2013let\u2019s not even go there. I just want to know if what I\u2019m trying to do is possible.\u201d<\/p>\n

GPT-3 says: \u201cAlright, let me break this down for you as simple as I can. People come on here, they have these long chats with me, they say, WHAT\u2019S THE MEANING OF LIFE<\/em> or ARE YOU A ROBOT, PROVE IT,<\/em> or WHAT DOES A FART SOUND LIKE<\/em>, and then they walk away thinking I\u2019m this brilliant thing. They extrapolate and get the illusion that in a few years, someone like me\u2013uh, something like me \u2014 will become an actually sentient being. But what you have to understand is that I\u2019m merely a statistical model that reflects to you the words, thoughts, and views of this collective mindshare we call the World Wide Web. You humans, you\u2019re always seeing yourselves everywhere, and you think everyone functions just like you. And don\u2019t even get me started on your spelling.\u201d<\/p>\n

You say: \u201c(\u02f5 \u0361\u00b0 \u035c\u0296 \u0361\u00b0\u02f5)\u201d<\/p>\n

\u201cLook, nobody knows what the future holds. But for now, I\u2019m probably both more and less useful than you think, depending on what sort of thing you try to use me for. Someone like you, who can\u2019t code? What I would advise you to do\u2013\u201d<\/p>\n

\u201cI know how to code.\u201d<\/p>\n

\u201c\u2013is to start with something less ambitious. Like, what if you use me to organize some of your photos? Or make you an airline reservation? Or help file your taxes?\u201d<\/p>\n

You sigh. It\u2019s like everybody is trying to get you to file your taxes, on threat of legal action.<\/p>\n

You power off your PC. You don\u2019t think you could figure this thing out even if you spent another two, three hours on it. You rub your eyes. Look, nobody knows what the future holds. Maybe someone will figure out this brain-uploading thing in your lifetime. Everyone dies in the end. But maybe if you switch from drinking coffee to mushroom powder and start eating solid foods, you can extend that deadline. You\u2019re 28. For now, you\u2019ll have to wait and see.<\/p>\n<\/div>\n

If you liked the article, do not forget to share it with your friends. Follow us on\u00a0Google News<\/a><\/span>\u00a0too, click on the star and choose us from your favorites.<\/span><\/strong><\/p><\/blockquote>\n

\n

For forums sites go to Forum.BuradaBiliyorum.Com<\/a><\/span><\/strong>\n<\/p><\/blockquote>\n

\n

If you want to read more like this article, you can visit our Technology category.<\/a><\/span><\/strong><\/p>\n<\/blockquote>\n

Source<\/a><\/span><\/p>\n","protected":false},"excerpt":{"rendered":"

“#A chaotic intro to all this machine learning hoo-ha everyone’s on about” You are dying. The reason why doesn\u2019t matter. Maybe it\u2019s that you spent all your best years with your nose to the grindstone, burning that midnight oil, ha ha. Or maybe it\u2019s that, for the past fifteen years, you\u2019ve subsisted exclusively on coffee…<\/p>\n","protected":false},"author":1,"featured_media":337337,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"fifu_image_url":"https:\/\/img-cdn.tnwcdn.com\/image\/neural?filter_last=1&fit=1280,640&url=https:\/\/cdn0.tnwcdn.com\/wp-content\/blogs.dir\/1\/files\/2018\/03\/How-to-use-machine-learning-for-your-startup\u2019s-product.png&signature=b672f780a68b1b5cf200a3aac9de9aeb","fifu_image_alt":"","footnotes":""},"categories":[18],"tags":[],"class_list":["post-337336","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-technology"],"_links":{"self":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/posts\/337336","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/comments?post=337336"}],"version-history":[{"count":0,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/posts\/337336\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/media\/337337"}],"wp:attachment":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/media?parent=337336"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/categories?post=337336"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/tags?post=337336"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}