kentran.net - Practical Machine Learning

Success! Domain kentran.net was analyzed on Wednesday 30. November 2016!

DomainsData.org: Practical Machine Learning

Title:

Practical Machine Learning
Age:

7 years old
Alexa Rank:

4,861,541
Total Sites Linking In (Alexa):

9
Domain's IP Country:

United States
Status Code:

OK
IP Address:

173.194.79.121
Description:

Ken Tran on machine learning, data technology, and beyond
Keywords:

kentran.net Whois Information:

1.

Domain Name:

kentran.net
2.

Domain Age:

7 years old
3.

Name Server 1:

ns1.dreamhost.com
4.

Name Server 2:

ns2.dreamhost.com
5.

Created:

Monday 16. February 2009
6.

Expires:

Thursday 16. February 2017
7.

Domain Registrar:

Dreamhost

Website Important Html Tags:

TAG
TEXT
b
Side notes
b
Side notes
b
Sigmoid
b
Softmax
b
Side notes
b
Notes
div
That's
div
Note that the second idea is not restricted to SDCA or even linear learning. In fact, we originally implemented this binary data loader for training large neural networks. However, it couples nicely with SDCA as the real strength of SDCA is on very large sparse datasets, for which the need for out-of-memory training arises.
div
See the
div
Machine learning researchers and practitioners often use one metric on the test set and optimize on a different metric when training on the train set. Consider the traditional binary classification problem, for instance. We typically use AUC on the test set for measuring the goodness of an algorithm while using another loss function, e.g. logistic loss or hinge loss, on the train set.
div
Why is that? The common explanation is that AUC is not easily trainable. Computing AUC requires batch training as there's no concept as AUC per example. Even in batch training, we just don't use it as a loss function [1].
div
I want to ask a deeper question. Why is AUC a good metric in the first place? It's not the metric that business people care about. Why don't we use the true business loss, which can be factored into loss due to false positives and loss due to false negatives, for testing a machine learning algorithm; and even for training it?
div
The major reason that AUC is favored as a proxy for business loss is that it is independent of the classification threshold. Why are we scared of the threshold? Why do we need to set a threshold in order to use a classifier model? Isn't it anti machine-learning that humans have to manually set the threshold?
div
So I'd like to propose that we shouldn't consider threshold as a parameter to tune. Instead, make it another parameter to learn. Here are a few challenges when doing so
div
ML research has always been challenging. Adding another layer of complexity shouldn't be an issue. Not modeling the business problem directly is more of an issue to me.
div
[1] Computing AUC is costly and computing the AUC function gradient is even costlier.
div
Deep Learning is hot. It has been achieving state-of-the-art results on many hard machine learning problems. So it's natural that many study and scrutinize it. There have been a couple of papers in the series of
li
A semi-asynchronous parallel SDCA algorithm that guarantees strong (linear) convergence and scales almost linearly with respect to the number of cores on large and sparse datasets.
li
A binary data loader that can serve random examples out-of-memory, off a compressed data file on disk. This allows us to train on very large datasets, with minimal memory usage, while achieving fast convergence rate (due to the pseudo shuffling). For smaller datasets, we even showed that this *out-of-memory* training approach can be even more efficient than standard in-memory training approaches [*].
li
If we are using a linear model, this addition of threshold parameter will make the model nonlinear. In fact, we will no longer have linear models.
li
The threshold parameter needs to be within 0 and 1. We can relax this constraint by applying a logistic function on a parameterized threshold variable.
li
Changing an image (e.g. of a lion) in a way imperceptible to humans can cause DNN to label the image as something else entirely (e.g. mislabeling a lion as library)
li
Or DNN gives high confidence (99%) predictions for images that are unrecognizable by humans
li
The whole data mining [2] pipeline is offline.
li
In some cases, companies just want to get insights from their landfill of data, to help them make more informed decisions. That’s it. They don’t need to or perhaps don’t know how to train/apply a predictive model.
li
Online prediction is not too critical. Some data scientists are willing to run prediction offline on new batches of data. Examples include customer retention analysis, risk analysis, fraud detection (many firms already do fraud detection online but there are still lots of firms running this offline).
li
Or after training and validating a model offline, the data scientists will ask other software engineers to implement the model decoder, typically in another language that is more efficient (C++/C#/Java), which can then be deployed as a live service.
li
The educational bridge between data scientists (or statisticians) and production engineers. Basically, the data scientists need to explain the model to the engineers. Depending on the complexity of the model and other factors, this process could be long and error prone.
h1
Practical Machine Learning
h2
Pages
h2
The disconnect between training models and putting them into production
h2
Pending Topics
h2
Labels
h2
Popular Posts
h2
Blog Archive
h2
Subscribe To
h2
Online tools

kentran.net IP Information:

1.

Ip Address:

173.194.79.121
2.

Country:

United States
3.

Status Code:

OK
4.

Region Name:

California
5.

City Name:

Mountain View
6.

Zip Code:

94043
7.

Speed test:

not available

kentran.net Alexa Information:

5 Websites linking to kentran.net:
microsoft.com
github.io
washington.edu
360doc.com
126kr.com

Websites related to kentran.net:

Top Keywords from Search Engines:
softmax vs sigmoid, implement convolutional neural network, l2 regularization, softmax sigmoid, l1 and l2 regularization

Site Screenshots:

Screenshot created on Wednesday 30. November 2016

This is a screenshot of kentran.net, created on Wednesday 30. November 2016
Click on thumb to enlarge the screenshot.

Google Ranking Screenshots:

Screenshot created on Wednesday 30. November 2016

This is a Google Top 10 results' screenshot for keyword Practical Machine Learning created on Wednesday 30. November 2016.
The position for keyword Practical Machine Learning is: Not in top 10
Google Ranking Position checked from US (United States of America) Ip Address on Wednesday 30. November 2016.
Click on thumb to enlarge the screenshot.