Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Dive into Deep Learning

d2l.ai

Post reply on HN
Print view View original

81–90 of 94 posts

Re: Dive into Deep Learning

#81

Post by rsp1984 »

fxtentacle wrote: post #27
throwlaplace wrote: post #23

Earlier quoted context omitted.

>As a practical example, figuring out where a given pixel moves from one video frame to the next one, when working on real-world videos, the best known algorithms get about 50% of the pixels correct. With clever filtering, you can maybe bump that to 60 or 70%, but in any case you will be left with a 30%+ error rate. what do you mean by this? optical flow isn't really a learning problem? it's a classical problem with…

It used to be. Then the AI fanboys arrived and started treating it like a learning problem. https://arxiv.org/abs/1612.01925 https://arxiv.org/abs/1709.02371 https://arxiv.org/abs/1904.09117 BTW, also the classical algorithms deal very badly with noise and repetitive textures, e.g. a video of a forest in the afternoon.

Ever tried "DIS optical flow" in OpenCV? Works like a charm for me even in challenging conditions.

Re: Dive into Deep Learning

#82

Post by hnarayanan »

sanjose321 wrote: post #3
throwlaplace wrote: post #2

this looks pretty good. certainly much better than goodfellow's deep learning book. definitely much the diagrams and code are much appreciated but i'm curious why mxnet over pytorch?

I find this comment amusing, have you read the goodfellow's book? That book is amazing.

I suppose you and I have very different notions of the word 'amazing'.

Re: Dive into Deep Learning

#83

Post by michaelcampbell »

mrspeaker wrote: post #46
newfeatureok wrote: post #28

Earlier quoted context omitted.

I'm somewhat surprised at the responses for this. I believe your issue can be easily solved - have supervisors wear a distinctive color from a non-supervisor. For example let's say it's yellow. OK so now you have yellow wearing supervisors and everyone else. To resolve the issue you have described acquire a month or so of footage, with labels per minute describing how many yellow wearing supervisors and how many peop…

"Easily solved - just have them wear special clothes." Everything is easy if you can arbitrarily change the requirements!

Changing the context is one of the many well respected ways to solve a problem.

Re: Dive into Deep Learning

#84

Post by michaelcampbell »

mrspeaker wrote: post #73
newfeatureok wrote: post #68

Earlier quoted context omitted.

I really don't understand your comment. 1. Supervisors are already by definition "superior" than their subordinates. 2. Supervisors on factories already wear distinctive clothing - especially in fully automated factories. Finally, you have yet to propose a solution to the problem yourself that would be highly accurate and easy to train. You vastly underestimate the difficulty to create a bespoke solution from scratch…

I don't have a deep-learning solution to the problem (I know nothing about it, that's why I clicked on it!). Seems really hard to me. I'd certainly go with an obvious "clock in and out" or "rfid" approaches... but I've worked in factories - and if you make someone wear some special clothes (or do some special tasks) when they didn't have to before - you're asking for trouble. People really hate change. That's presuma…

"nerd", "computer guys bad at..." it feels like you have an irrational axe to grind here, when a simple solution presented causes this line of argument.

Re: Dive into Deep Learning

#85

Post by coredog64 »

deepGem wrote: post #78
newfeatureok wrote: post #28

Earlier quoted context omitted.

I'm somewhat surprised at the responses for this. I believe your issue can be easily solved - have supervisors wear a distinctive color from a non-supervisor. For example let's say it's yellow. OK so now you have yellow wearing supervisors and everyone else. To resolve the issue you have described acquire a month or so of footage, with labels per minute describing how many yellow wearing supervisors and how many peop…

This is a pragmatic and valid approach. No matter what anyone else says. Imagine, you told a 10 YO child to do this task. Even the child would ask the same question - how do I know who is a supervisor and who is not. Not only is face recognition hard, it is almost impossible to accomplish in a factory floor like setting. Not totally impossible but it is really really hard. Face detection is still possible but face re…

I remember hearing a similar story circa 2003-2008. Some BigCo was spending a bunch of money to automate their inter office mail and was hung up on OCR of handwritten stuff. Some consultants come in to look and one asks if they can just use different color envelopes/baskets. The answer was “yes”.

Re: Dive into Deep Learning

#87

Post by tel »

samcodes wrote: post #74
fareesh wrote: post #43

Earlier quoted context omitted.

Thanks - this validates many of the assumptions I had about this part of the process. It has been challenging communicating many of these realities to non-technical folks, who seem to be quite misguided about implementing these types of systems as opposed to "non-ML" systems where there is a less imperfect and more predictable idea of what's possible, how well it will work, and how much effort is required to pull it…

Yeah IME expectations with ML are just the worst. Somehow, non-ML-educated stakeholders expect it to be predictable, like they pretend traditional software engineering is... but also to be magical in scope.

Honestly, it's not surprising. ML is billed as a tool, one that in the last 5 or so years we've surprisingly "figured out". This is vast overselling, but it still creates the basic mental model for folks without further training: ML is a tool you can apply to certain situations to achieve outcomes that you used to need people for, especially in vision and NLP.

I personally believe this is false, but also false in a way that we're remarkably far away from that. Even more than software, predictive automation is a process. It often relies on particular customization to your own situation to be successful. It can demand vast resources. It's wildly difficult to debug.

So we should be working to retrain those around us. ML is a process.

Re: Dive into Deep Learning

#88

Post by asiachick »

newfeatureok wrote: post #28
fareesh wrote: post #8

As an engineer I find myself in this type of situation quite often - if anyone can point me to some good resources or has any advice, I'd be quite grateful: - Some non-technical stakeholder comes to me and says "can we solve this problem with Machine Learning?" usually it's something like "there need to be two supervisors on the factory floor at all times, and I want an email alert everytime there are less than 2 sup…

I'm somewhat surprised at the responses for this. I believe your issue can be easily solved - have supervisors wear a distinctive color from a non-supervisor. For example let's say it's yellow. OK so now you have yellow wearing supervisors and everyone else. To resolve the issue you have described acquire a month or so of footage, with labels per minute describing how many yellow wearing supervisors and how many peop…

Just passing on info but ANA (the airline company) has colored helmets in their maintenance to facility to distinguish supervisors (color 1) from non supervisors (color 2) and 1st year employees (colors 3) and guests (color 4). I don't know if they do any tracking.

Re: Dive into Deep Learning

#89

Post by lindbergh »

bor100003 wrote: post #86

Has anyone read this book ? It look very attractive but I want to hear some feedback before bookmarking another ML book.

Kinda did, but mostly the first chapters, actually up to CNN chapter (where real modern DL start). But so far, I really liked what I read. It has a very good blend of code and theory, with hands on applications throughout the whole book. Most importantly, all those applications could perfectly be copy pasted into your own environment. So it actually reminded me of a very thorough tutorial on a framework, more say than a regular textbook, although the authors don't compromise on mathematical arguments (but don't get lost in it either, they skimmed pretty fast on regularization theory imho). If you've had previous exposure to classical ML, I think it's a fantastic introduction to DL, enough to get started.

Re: Dive into Deep Learning

#90

Post by the-alchemist »

mpfundstein wrote: post #48
dragandj wrote: post #42

Earlier quoted context omitted.

Only a limited hand-crafted hardcover edition is planned. That being said, you can print a dead tree version from the PDF at your local printing shop (or at home) if you care about the text, and not that much about binding.

Yes, that would be an option.

Most local print shops will do something like that for And I just did a quick price check, and https://xpress.lulu.com/ will do it for $10 as well, with shipping in 2 or 3 days (US).
Post reply on HN