jklontz

8 ms·

1.
by jklontz 6y ago
> For example, NN-512 can exceed 48 effective FMADDs per cycle (on the 27 peak FMADD machine) with Winograd-Cook-Toom-Lavin, if the tensor is deep enough (enough channels) Roughly how many channels do you need for this approach to be wor
2.
by jklontz 6y ago
Would you share this no-knead recipe?
3.
by jklontz 7y ago
The AWS Rekognition study I believe you are referring to was fundamentally flawed in two ways. First, the threshold used was the default suggested for one-to-one matching (lower false reject rate, higher false accept rate). It doesn't
4.
Jim Keller: Moore's Law Is Not Dead(hexus.net)
1 points by jklontz 7y ago|0 comments
5.
How Automated Face Recognition Works(blog.rankone.io)
2 points by jklontz 8y ago|0 comments
6.
by jklontz 10y ago
I've seen this argument made a lot, but have never found it convincing. Is there any empirical evidence to back it up? A quick Google search suggests it is a myth.
7.
by jklontz 10y ago
Rank One Computing | Denver, CO | Onsite | Full-Time Rank One Computing is an employee-owned small business specializing in computer vision and facial recognition algorithm development. Our technology is licensed by numerous organizations a
8.
Ask HN: How was your experience hiring a white hat hacker?
23 points by jklontz 10y ago|5 comments
9.
by jklontz 10y ago
Learned something new, fixed thanks!
10.
by jklontz 10y ago
Also in Denver! Would be thrilled to hang out with any fellow HN folks. E-mail in profile :)
11.
by jklontz 10y ago
True, but unlike other commodities it's easy to take a long position on gold with ETFs like GLD which actually store the physical gold. For other commodities (like oil with ticker USO) it's not so simple to take a long position. T
12.
by jklontz 10y ago
Targeting FPGAs is something we've been keeping our eye on as well. Most of the pieces are in place. I think the last non-trivial part is removing any malloc()/free() calls. This should be possible as the size and lifetime of memo
13.
by jklontz 10y ago
Going to take this opportunity to plug my related project [Likely](www.liblikely.org), a DSL for lowering machine learning inference algorithms. One of the projects we've built on top of it is a static compiler for Caffe model files. T
14.
by jklontz 10y ago
Agreed! Instead we use a Slack channel called #todayiwill and people just enter a few bullet points when they get in.
15.
by jklontz 10y ago
Very exciting! I wonder how the short-term nature of the pilot program might have some inherent biases compared to the long-term study? Presumably there would be less dramatic changes in behavior / lifestyle?
16.
by jklontz 10y ago
Seems like the perfect opportunity to insert a plug for my project: http://liblikely.org/
17.
by jklontz 10y ago
Yes! And I would also add that: > LLVM IL is more cluttered with type annotations and casts. With LLVM moving to typeless pointers[1], this will be less true in the future than it is now. [1] https://groups.google.com/for
18.
Ask HN: Small business bank account – who do you use?
4 points by jklontz 10y ago|0 comments
19.
by jklontz 10y ago
We published a paper on this a few years ago! Short answer: algorithms exhibit different accuracies on different races, genders, and ages. B. F. Klare, M. J. Burge, J. C. Klontz, R. W. Vorder Bruegge and A. K. Jain, "Face Recognition P
20.
by jklontz 11y ago
The people I know have been notified.
21.
by jklontz 11y ago
Replacing my workstation with a high-end laptop removed all but the power cord for me ;)
22.
by jklontz 11y ago
You might be interested in [Likely]( http://liblikely.org/ ), the programming language I've been working on. One of its core tenants is that "training == compilation". I haven't given much thought to depar
23.
by jklontz 11y ago
This is _not_ a post about auto-vectorization in Rust (which would have been a lot more interesting!). The provided Mandelbrot Set example is algorithmically similar to loop unrolling with a 4x unroll factor. The strategy works well in this
24.
by jklontz 11y ago
WebAssembly Backend RFC on the LLVM mailing list: http://llvm.1065342.n5.nabble.com/RFC-WebAssembly-Backend-td...
25.
by jklontz 11y ago
GitHub: I've emailed their support desk a half dozen times over the years, and each time received a response from an engineer within minutes containing an accurate and precise answer to my question.
26.
by jklontz 12y ago
For reference, current DDR4 RAM to CPU is ~0.5 Tb/s [0]. [0] http://ark.intel.com/products/81059/Intel-Xeon-Processor-E5-... (see "Max Memory Bandwidth")
27.
by jklontz 12y ago
Yes! For those interested in a DIY project, I would add that "Permaculture: A Designer's Manual" [0] is arguably the reference for these techniques. [0] http://www.amazon.com/Permaculture-Designers-Manual-Bi
28.
by jklontz 12y ago
Your weekly is best part of my Monday morning, glad to hear it will be around in 2015 :)
29.
by jklontz 12y ago
OpenBR does gender & age (probably pose in 2015). YMMV though, current algorithm is very dependent on an "ID card" quality photo.
30.
by jklontz 12y ago
OpenBR author here, thanks for the plug!

More ›