Hacker News Reader

ericjangon SEC Rejects Rule Change for Bitcoin Exchange-Traded Fund [pdf]

I almost read that as Wink-evil. https://en.wikipedia.org/wiki/Priming_(psychology)

ericjangon Neural Machine Translation and Sequence-To-sequence Models: A Tutorial

Stephen Merity of Metamind has a nice visual tutorial here as well: http://smerity.com/articles/2016/google_nmt_arch.html

ericjangon Uncertainty in Deep Learning (2016)

Suppose you train a neural net on cat pictures to classify the breed of cat. We desire the property that if we were to feed in a picture of a horse instead of a cat, we could somehow measure how good the network's parameters are for classifying this particular image. This is uncertainty estimation, and Yarin's blog post + thesis provides an elegant way to compute this, which get nearly for free from the existing model.

Concretely, if you are trying to train a neural net to forecast stock prices or drive a car safely, not only do you want to have predictions, but you want to estimate some measure of how confident your model is of that prediction. This is eminently useful for models that lean towards the "black-box" spectrum, such as deep neural nets.

Note that parameter uncertainty and risk estimation are quite different, which are addressed in this preliminary work http://bayesiandeeplearning.org/papers/BDL_4.pdf

ericjangon 12,000 startups are being created every day in China [video]

Fascinating documentary, thanks for sharing! If I understand the business model correctly, isn't the 融道网 (Roadoor.com) 9.00% guaranteed principal + interest mentioned in the documentary considered a Ponzi scheme in the US? How much of the return on capital is from RoI from investment into companies (exits, dividends, organic growth) vs. just increased AuM from (老百姓) investors?

Edit: just realized that businesses requesting loans to 融道网 are screened prior to being made available to public investors, so the deal flow appears to be capped by available business opportunities.

ericjangon The power of role models

Thank you for your civil reply, and I agree that the question has not been put to rest, as evident by the ongoing debate regarding how diversity initiatives should be implemented, if at all. I also agree that all this should be taken with utmost seriousness and consideration of opposing points of view.

Since I do not have firsthand experience with being a woman, my information comes from talking to women computer scientists and engineers from my school CS program and my workplace, and listening to their firsthand experiences dealing with sexism in the industry. Their opinions are diverse, but generally are in support for more empowerment of women/minorities in the workplace, so I generally support their agenda. Personally, I enjoy working in a more diverse workplace, anyway (long explanation for another day).

In Soviet Russia during the Cold War, I believe boys and girls were both encouraged to study math and sciences, which resulted in a fairly even gender representation in the career, and neither gender was better than the other on average. Food for thought.

ericjangon The power of role models

> And what of the girls who learn that their role models have been given a handicap, pushed upwards beyond their skill by well-meaning but naive men?

Proponents of affirmative-action policies argue that the handicap offsets disabilities from the glass ceiling, discrimination / stereotype threat in the workplace/education pipeline. The calculus of what "balances out inequality" is imperfect, but arguably better than just having the disability.

ericjangon Co-founder Conflict

Thanks for sharing this, Garry. In hindsight, would you have preferred to run your startup as a solo founder, or do you believe that a solo founder startup presents a greater operational/execution risk than co-founder conflict?

ericjangon Show HN: Julia Observer – Package Browser for the Julia Language

Is anyone else concerned that despite being an open-source project with permissive licensing, the formation of the Julia Computing group has made Julia's development a bit of a walled garden? I don't mean this in the sense that Apple or Matlab are "walled gardens", but the for-profit company, of which many core developers are founders/employees, develops a number of tools and services that are not open-source AFAIK.

I believe the group also implicitly controls what packages are allowed/disallowed from the packages list. At the same time, I also recognize the need to support the core developers and the business model makes a lot of sense. Thoughts?

ericjangon Neural Architecture Search with Reinforcement Learning

> is it viable for me to even think about this kind of thing - doing such a residency? Would I even have a chance of being considered, vs someone younger? Furthermore - how does one do a residency of this nature, while still paying bills (I live in Phoenix, AZ - I own a house, plus have other bills)?

Good on you for pursuing continued education.

Yes it's viable - never lose hope in yourself for anything. The Brain Residency has no age requirement; in fact, the program looks for diverse academic/career experiences. You'd be surprised how close you are to your dream job if you are passionate, hardworking, and lucky.

I too, was intimidated by the math & magic of ML when I first took Andrew Ng's Intro to Machine Learning Course 4 years ago (my first exposure). I took regular CS courses in college and started focusing on ML 1.5 years ago. Beyond taking courses, I highly recommend building your own projects - it exercises your independent research ability and creativity.

The Brain residency pays a good wage for the SF Bay Area (comparable to new grad salary at average tech firms), so you should be able to pay your bills. I've heard rumors that many other companies highly invested in ML are following in Google Brain's example and starting similar residency programs. Your ML career path need not be at Google, though we're a pretty solid choice ;)

ericjangon Neural Architecture Search with Reinforcement Learning

The primary author is a Google Brain Resident (https://research.google.com/teams/brain/residency/). The Brain Residency is a really great program for starting a career in Deep Learning and ML research, and I'm really impressed by how quickly these new researchers churn out great work like this.

disclosure: I work at Google Brain

ericjangon AI Pioneer Wants to Build the Renaissance Machine of the Future

Here's some additional context from a Reddit discussion: https://www.reddit.com/r/MachineLearning/comments/5go4sa/n_w...

ericjangon Alphabet’s Waymo Cuts Cost of Key Self-Driving Sensor by 90%

Having spun/graduated out of X (also now an Alphabet subsidiary), Waymo is now a subsidiary of Alphabet and so by technicality it's more accurate to refer to it as Alphabet's Waymo, rather than "Google's Self-Driving Car"

ericjangon Finding an Alternative to Mac OS X

Another path you can take: build your own desktop and install a Hackintosh OS!

Here's my build (http://pcpartpicker.com/list/pYWQMp), which runs into some driver issues but can be fixed with Kext utility. This setup allows me to do CUDA programming + Deep Learning stuff natively in OSX, while having access to pretty Mac software. It's great, I highly recommend it!

ericjangon Summary of the Neural Information Processing Systems Conference 2016

(I've edited my blog post to clarify this)

ericjangon Learning to Learn in TensorFlow

another one: Hyper-Networks https://arxiv.org/abs/1609.09106

and a blog post to go with it: http://blog.otoro.net/2016/09/28/hyper-networks/

ericjangon Summary of the Neural Information Processing Systems Conference 2016

Sorry, that was what I meant to say. My wording was confusing

ericjangon Why do traders in investment banks feel their jobs are immune from AI, etc?

Due diligence on the financials of a company (what investment bankers are supposed to do) is actually really hard to get right with the algorithms we have today. Much of the data and insight compiled by an I-banker today does not exist in an easily parse-able form for automated algorithms, and a substantial amount of the computation relies on common sense knowledge.

ericjangon People's underestimation of AI

agreed. my worry is that even if people decide to ethically develop AI, there will be those who disagree with those ethics and/or willingly ignore their conscience. I don't see a way out short of a political revolution / bloodshed.

ericjangon People's underestimation of AI

I'll be grinding my magic lvl in the meantime.

ericjangon People's underestimation of AI

I like your insight. Is that the main message conveyed in [0]? (I haven't read it)

What do you think is going to happen? Can society do something to prevent mass concentration of wealth into the 0.00001% of people via the march of technology and innovation, or is this an inevitable outcome?

ericjangon When A.I. Matures, It May Call Jürgen Schmidhuber ‘Dad’

Hey Jurgen, I'm a big fan of your work!

ericjangon New MacBook Pro Is Not a Laptop for Developers Anymore

I dislike this post intensely. The opinion comes across as uncharitable, as the author probably has not used this Macbook Pro yet. How can you (alexeysemeney) be so sure?

I consider myself a developer, and I almost never use the Fn and ESC keys on my Mac. Everybody uses the computer differently, but I'm pretty sure this is not a deal-breaker for most. One could also argue that the touch bar might lead to innovative developer tools, such as timeline interfaces for Replay Debugging. There are already a lot of applications in the content creation (Photoshop, Illustrator, Maya) space, which developers frequently use to create assets.

ericjangon Google buys Eyefluence eye-tracking startup

This operates on the assumption that eye tracking and brain waves signals are positively or negatively correlated (for which the setup you propose can be thought of as an estimator with control variates). Is this indeed the case?

ericjangon Why the Film Industry Hasn't Been Disrupted Yet, Part 5

How much of this applies if the medium is pure CG/traditional animation?

ericjangon Peter Thiel, YC, and hard decisions

Wow, this is the third time I've seen this link climbing up HN front page in the past hour [1][2]. I'm interested in this topic, why does it keep getting flagged?

[1] https://news.ycombinator.com/item?id=12727973 [2] https://news.ycombinator.com/item?id=12726970

ericjangon [dead]

The previous discussion was at the top of HN 1 hour ago https://news.ycombinator.com/item?id=12726970

ericjangon Intel will add deep-learning instructions to its processors

Not exactly late. There's no physical law that says a CPU can't beat a GPU for a certain family of operations, and in fact in practice, small toy networks with non-dense operations (for instance, if you are doing research on optimization algorithms for DNNs or Bayesian inference algorithms) tend to run faster on the CPU than GPU.

ericjangon Differentiable Neural Computers

here you go http://rdcu.be/kXdj

ericjangon Differentiable neural computers

Link to full PDF: http://rdcu.be/kXdj

ericjangon Daydream – Google VR

"you're a wizard"

Say no more. Take my money.