machinelearning

7 ms·

151.
by machinelearning 7y ago
Isn’t that the goal of every AR startup/project? Care to explain?
152.
Show HN: NuronLabs – A Worldwide Neural Layer(nuronlabs.com)
2 points by machinelearning 7y ago|1 comments
153.
by machinelearning 7y ago
Nah, its probably a consolidation strategy
154.
by machinelearning 8y ago
https://arxiv.org/pdf/1709.06030.pdf
155.
by machinelearning 8y ago
It seems to be published here: https://github.com/manoloide/AllSketchs
156.
by machinelearning 8y ago
This is why the article says "wrong" abstraction. Replacing duplication with abstraction often leads to the wrong abstraction. That's the point of the article. That being said, good thoughts on solving duplication correctly.
157.
by machinelearning 8y ago
Yea this sounds extremely unlikely unless the other dataset has a fairly easy decision boundary. The kind of cross-domain transfer learning you seem to think deep neural networks have is nothing I've observed before in my formal studie
158.
by machinelearning 8y ago
Or in Thiel's own(?) words: "Competition is for losers"
159.
by machinelearning 8y ago
Yea this is spot on and the pattern of probing the opposite argument for possible pay off is a classic contrarian move.
160.
by machinelearning 9y ago
Doesn’t seem to match what’s written here: https://www.uscis.gov/tools/glossary/country
161.
by machinelearning 9y ago
It takes little imagination to see how methods used for designing neural networks can be applied to other parts of ML (e.g. optimization, feature selection etc.) AutoML is definitely not a subset of NASNet..
162.
by machinelearning 9y ago
corpus christi.
163.
Network to Network Compression via Policy Gradient Reinforcement Learning(arxiv.org)
1 points by machinelearning 9y ago|0 comments
164.
by machinelearning 10y ago
I find it amusing how you make an extraordinary claim as though its obviously true, and then criticize life extension as "desperate egocentrism". How exactly is death directly responsible for new strategies? And how is it a "
165.
by machinelearning 10y ago
I'd like to hear the rationale of the people who downvote this comment. This comment looks correct to me. The main characteristic of established protocols in the event of an emergency are that they are supposed to be reliable and not r
166.
by machinelearning 10y ago
both problems have constraints and you solve the puzzle by finding the right configuration of data that meets the constraints.
167.
by machinelearning 10y ago
Of course, imitation plays a part whenever you have some kind of group identity. However, there's a distinction between why the practice actually exists and why people partake in it. The question being answered is the former.
168.
by machinelearning 10y ago
Not a muslim myself, but its pretty apparent you haven't the faintest clue about the reasons behind the fasting. If you actually read even a little of the article, you will notice that there is no reference to "making one's f
169.
by machinelearning 10y ago
Some companies use a successful Kickstarter campaign as a positive signal to attract investment. If they set a low but reasonable goal and achieve it, it is ostensibly a positive signal that a market exists for whatever niche product they a
170.
by machinelearning 10y ago
www.gest.co A better designed device that does not require you to learn novel patterns and can do much more
171.
by machinelearning 10y ago
"I've met quite a few with more talent for abstract thought than some math professors." Such a statement can be made about almost any two fields. I suspect "quite a few" is either hyperbole to strengthen your point
172.
by machinelearning 10y ago
From what I can tell, this seems to reduce the cognitive load of a search session way more than using Google (for certain use cases). The advantages are: fewer clicks/taps, session state is saved on the page and suggestions reduce typi
173.
by machinelearning 10y ago
peter thiel is a contrarian. That implies that he follows the unpopular opinion. Since elections are based on popularity, if thiel is right, Hilary becomes president
174.
by machinelearning 10y ago
In this case yes, because they are influencers of the community and were not being responsible in ensuring the proof was rigorous enough
175.
by machinelearning 10y ago
Surprised this hasn't been posted yet. From Gavin's reddit account, "Craig signed a message that I chose ("Gavin's favorite number is eleven. CSW" if I recall correctly) using the private key from block number
176.
by machinelearning 10y ago
This is so false. Applying machine learning to a real world problem requires correct intuition and the ability to quantify tradeoffs mathematically. This is developed by understanding the math behind the model and what the tradeoffs are.
177.
by machinelearning 10y ago
2001 src: http://www.cs.cmu.edu/~NatProg/publications.html
178.
by machinelearning 10y ago
There is a huge asymmetry in the consequences of not pushing back. While the expert might be frequently frustrated by having to explain away the same cliches, the patient/customer has a lot to lose by not pushing back. Especially when
179.
by machinelearning 10y ago
Obviously it's not possible to pick a sample size large enough to eliminate all bias. However it's important to pick a sample that doesn't inherently bias the effect that is being measured . This has to apply to both the ori
180.
by machinelearning 10y ago
The original paper was proven to be more alarmist than true. Dan Gilbert's team put out an excellent analysis of the paper and found many poorly performed replications. That being said, it sheds light on the greater problem of p-value

More ›