sarthakjain

23 ms·

61.
by sarthakjain 10y ago
Have any of the methods made radiologists more efficient? If you were to imagine a system that made a radiologist 10x more efficient what would it look like?
62.
by sarthakjain 10y ago
The approach was applied without any real knowledge of art, even though it has been applied to the domains you mentioned I don't see why not. [edit]: it is a lot harder to build a NN when there are very constraint rules. But it is also
63.
by sarthakjain 10y ago
Deep learning architectures built by machines (so we no longer have to design architecture to solve problems) https://arxiv.org/abs/1611.01578 Transfer Learning (so we need less data to build models) http://
64.
by sarthakjain 10y ago
Deep Learning is mostly just about automatically creating good features. As an example if you had the driving data and were too lazy to hand engineer features for what looked like loops looking for parking. You could use deep learning and t
65.
by sarthakjain 10y ago
Thanks for trying us out! Internally we have validation metrics which have a number that says how good a model is at class separation. One naive way to do this entropy (-plogp). We are planning on exposing this to users soon. So once you cr
66.
by sarthakjain 10y ago
Your right the article doesn't do justice to all of the transfer learning techniques. I cut out large portions of the original draft for brevity. Will write a longer post referencing all possible techniques, hopefully it won't tur
67.
by sarthakjain 10y ago
While I agree that you could build a demo in Deep Learning 101 that could work for some small set of examples, I disagree that this is 101 level material. Facebook just released: https://techcrunch.com/2017/02/02&#
68.
by sarthakjain 10y ago
Great feedback, I really appreciate you taking time to read the post. Though I would like to make a few notes regarding what we are trying to achieve and some improvements we've made: 1) Our target audience is someone who hasn't t
69.
by sarthakjain 10y ago
Yes transfer learning is a fairly umbrella term encompassing a lot of different approaches. We tried to give an example of the one most commonly used in NNs almost exclusively with regards to feature extraction. Do you have some resource th
70.
by sarthakjain 10y ago
Your right about excitement been generated from new techniques in the research community. Also if you look at Prisma, Google photos or any number of applications that have gained mass popularity. They are just implementations of existing ne
71.
by sarthakjain 10y ago
Spot on. If you take that argument a step further that means an average developer who is not a data scientist or ml researcher might not know about it. Which implies a super easy to use dead simple technique which is used by most researcher
72.
by sarthakjain 10y ago
This does seem to be holding back a lot of people from trying out DL in production. There are some advantages to using transfer learning in the cases you mentioned (eg models not generalizing for watermarks or differences in training and te
73.
by sarthakjain 10y ago
Totally agree, everybody in the industry knows about it. However if you look at https://www.google.com/trends/explore?date=2014-01-01%202017... nobody outside seems to know. I might be wrong but a lot of people outside
74.
by sarthakjain 10y ago
Machine Learning and AI seem to be in vogue but become tough to implement unless you have boatloads of data. We've personally had multiple frustrating experiences over the last ~7 years of trying to solve problems using ML. In almost a
75.
How to use Deep Learning when you have Limited Data(medium.com)
252 points by sarthakjain 10y ago|50 comments
76.
by sarthakjain 10y ago
Golang, React
77.
by sarthakjain 10y ago
Also a post to slack option
78.
by sarthakjain 10y ago
It would be super cool to add multiple browser engines and specify a site map/bulk list of subpages
79.
by sarthakjain 10y ago
We hired 50% of a 10 person startup team from Angel List, rest from referals, no luck hiring from LinkedIn. For startups AngelList makes a lot of sense and solves for problems like a wide variance in salary, non existent recruitment pipelin
80.
by sarthakjain 10y ago
Post some of the example logos generated using your app as a blog post and post it as show HN, ideally people should be able to try it out so you might want to give the first 100 users a free logo and display those as well
81.
by sarthakjain 10y ago
Listening to lectures is a good start. I found that listening to it at 2-3x speed multiple times works best for me. Skimmimg a variety of articles on the topic, this helps for understanding the jargon used in the field. Jargon intimatedes m
82.
by sarthakjain 10y ago
Hire for problems you need solved as opposed to generalists. Hire for strength as opposed to lack of weakness. Hire for skillset but understand that a large portion is going to be learning on the job. Hire someone people who care about your
83.
by sarthakjain 10y ago
If solutions were demonstrated, do you think customers would still care about black box vs understandable?
84.
Ask HN: Biggest obstacles to widespread Machine Learning adoption
5 points by sarthakjain 10y ago|2 comments
85.
by sarthakjain 11y ago
An algorithmic feed has the problem of forcing people to shout so they can be heard. Facebook has esentially became a shouting match of tye best photo or video. Which is why HN has the newest section where I found this post, likely hood of