Deep Learning: anybody working in that field?

Any OGers working on the whole Deep Learning field? 

Where? What are you currently working on? What's your background, aka what did you do before that? 

 

 

ttt

No but intrested. 

I work in digital pathology, whilst I don't personally use it at the moment deep learning is going to be big. 

Can someone explain a little on what this is? 

Dropkick Joe -

Can someone explain a little on what this is? 

This

Sub

I can barely learn normally let alone deeply.

4 later

Deep Learning is a part of Machine Learning. 

Machine Learning is a part of  Artificial Intelligence. 

 

Short in OG Terms: It's the reason that in the future your wives and girlfriends are not being banged by BBC while you are not at home, but by caring, understanding Robots. 

 

Deep Learning is what made possible, that you upload a pic and it'll tell you how old you are. What gender, etc. 

Language processing, credit card fraud dection, chat bots, self driving cars etc there is so much application for it out there and it's just crazy. This is going to be one of the most important fields in IT in the very short future. 

Deep Mind, Google's AI, beat the best GO players (most complex boardgame in existance) a while ago. Because you have about 200 options everytime you move, you can't "think ahead 8+ moves" like in chess. Chess btw is child's play compared to GO. So the best human players, use intuition. 

Well, Deep Mind made sweet sweet love to that Go Player's anus. But the thing is: They did NOT program how DeepMind should play. All they did was code the basic rules, then leave it alone. AlphaGo (the Go Software from DeepMind) played against itself and everytime it won, it changed itself and upgraded it's logic. 

DeepMind takes around 4 hours to learn chess (from absolute scratch) until it can beat the best humans on the planet. 

 

Short and sweet: 

 

 

More Details, but still "dumb enough"

 

 

Average,

Are you looking to get into this field? 

I work in insurance industry and I am hoping to switch to machine learning/data science in the future. Deep learning is not that used in this industry.

Big tech companies - Google, Amazon Nvidia are the big users. 

Good stuff guys. 

I'm working IT in large financial firm. We recently had an introduction to this (brown bag style during lunch). It was highly interested and I asked alot of questions. 2 Weeks later we got the opportunity for 2 people to go on a intensive 3 day camp about machine learning, deep learning etc... kind of a uber nerd retreat for 3 days, learning and then toying around with these techs. And I got nominated for one of these spots, think the boss 2 levels above noticed me during the Q&A. 

So to prepare, I've been putting more time into this making sure I can get "as wise as possible" before that retreat. 

@fanat: Insurances are going to be some of the biggest consumers of this in the near future. You're doing the right thing. 

I'm a data scientist at a fortune 500 company. I use deep learning along with tons of other techniques. Dont know what they'll be able to teach you in 3 days tbh. Just in deep learning you need to understand optimizations, loss functions, transformations, network architecture and the proper apis to put it together. Not to mention data prep and use case applications. Not trying to be discouraging because I love my job but be prepared for a ton of extra learning on your own

dabigchet - i have responsibility for data science at my company (which is also insurance, fanat).

one of the things that annoys me is different terms and definitions that are used. its almost like "deep learning" took off as a term because it is cool sounding. as i understand it, it really is just specific techniques at the edge of machine learning and is only applicable/useful for certain use cases. we do have some projects that are technically deep learning projects but its kind of the least interesting/valuable stuff that we do.

data science is where it is at though. it seems like its hard to get taken seriously if you don't have the educational background.

How hard is it to get into this field?

If I have masters and good math background would that be enough? I am getting better with R and Python every day.

Would having experience in the industry help? I am hoping to make some sort of lateral move in the future where I am better off financially also. I'm not sure that's also easy.

Weiderloo - 

I'm a data scientist at a fortune 500 company. I use deep learning along with tons of other techniques. Dont know what they'll be able to teach you in 3 days tbh. Just in deep learning you need to understand optimizations, loss functions, transformations, network architecture and the proper apis to put it together. Not to mention data prep and use case applications. Not trying to be discouraging because I love my job but be prepared for a ton of extra learning on your own


Weiderloo,

What are the most important things to know for the interview?
Do they test you on coding? Math concepts?

Average - 

Any OGers working on the whole Deep Learning field? 

Where? What are you currently working on? What's your background, aka what did you do before that? 

 

 


I own a small consultancy that is focused on AI/ML/NLP ....we are doing a re-architecture of Chatbots for a major banking institution right now

This shit scares me. The pace at which we've made advancements in A.I. is downright breathtaking (sorry that sounded gay, but I'm trying to drive home a point). Think of where all of this will be 5 years from now? If it scares the piss out of a guy like Elon Musk then it should be scaring every other sane person in the world. The fact that he's fatalistic in his approach to A.I. now because nobody would listen to him seriously enough to be proactive with safeguards and such just shows you how wreckless we are as a species. The robots are taking over. It's inevitable. Let's just hope that theory where we're all holograms and those that didn't help contribute to its development will be punished with eternal torture isn't true. The best we can hope for is that the lights just shut off for good when the end comes.

dabigchet - 
fanat - 
dabigchet - i have responsibility for data science at my company (which is also insurance, fanat).

one of the things that annoys me is different terms and definitions that are used. its almost like "deep learning" took off as a term because it is cool sounding. as i understand it, it really is just specific techniques at the edge of machine learning and is only applicable/useful for certain use cases. we do have some projects that are technically deep learning projects but its kind of the least interesting/valuable stuff that we do.

data science is where it is at though. it seems like its hard to get taken seriously if you don't have the educational background.

How hard is it to get into this field?

If I have masters and good math background would that be enough? I am getting better with R and Python every day.

Would having experience in the industry help? I am hoping to make some sort of lateral move in the future where I am better off financially also. I'm not sure that's also easy.

masters in mathematics or related?

i would suggest finding ways to apply R and python to the actual work that you do or that is done in your area. there is limitless possibilities for this. promote the work you are doing among your own leaders. if you have the right group/team in your organization that can further your career, make sure they become aware of the good work you are doing.

an industry trend that is occurring now is that the traditional data science role is being segmented. before, it was thought of as this "unicorn" mindset. they had to know data, have subject matter expertise, have a strong math background, computer science, had to be able to present and explain results. now, there are data engineers who specialize there, consultant types who specialize there, letting the computational people focus more. if you have industry knowledge and strong data skills and an interest in data science, that can play to your advantage.

Thanks for lots of valuable advise here.

I have MBA in finance with math undergrad. I work as a risk analyst with some opportunities to use Python and R at work so all of will hopefully help!

I have to work more on presenting myself well and my communication skills. That's always been my challenge a little bit.

trobinson21 - 
fanat -
Weiderloo - 

I'm a data scientist at a fortune 500 company. I use deep learning along with tons of other techniques. Dont know what they'll be able to teach you in 3 days tbh. Just in deep learning you need to understand optimizations, loss functions, transformations, network architecture and the proper apis to put it together. Not to mention data prep and use case applications. Not trying to be discouraging because I love my job but be prepared for a ton of extra learning on your own


Weiderloo,

What are the most important things to know for the interview?
Do they test you on coding? Math concepts?

https://www.springboard.com/blog/machine-learning-interview-questions/


Thanks. That's a great list. Seems like I know lots of concepts there but I will to brush up on most of them and learn a lot more.