Machine Learning on Oracle Autonomous Data Warehouse

Last week I wrote a blog post about how long it took to create machine learning models on Oracle Database Cloud service. There was some impressive results and some surprising results too.

I decided to try out the exact same tests, using the exact same data on the Oracle Autonomous Data Warehouse Cloud service (ADW).

When creating the ADW service I took the basic configuration and didn't change anything. The inbuilt machine learning for the Autonomous service will magically workout my needs and make the necessary adjustments, Right? It can handle any data volume and any data processing requirements, Right?

Here are the results.

* You will notice that there is no time given for creating a SVM model for the 10M record data set.  After waiting for 4 hours I got bored and gave up waiting (I actually did this three time to make sure it wasn't a once off) 

Update: The SVM Model for the 10M record data set eventually finished after 547 minutes! That's 9.1 hours !

[I also had a 50M record data set. I just didn't waste time trying that.]

[Neural Networks algorithm hasn't been ported onto ADW at this point in time]

If you look back at the results from using the DBaaS you will see it was significantly quicker than the ADW.  (for some it would be quicker using Python on my laptop)

Before you believe the hype, go test it yourself and make sure it measures up.

I re-ran my test cases over a number of days to see if the machine learning aspect of the Autonomous kicked in to learn from the processing and make any performance improvements. Sadly the results were basically the same or slightly slower. Disappointing.

When some tells you, you should be using this, ask them have they actually used and tested it themselves. And more importantly, don't believe them. Go test it yourself.

Understanding, Building and Using Neural Network Models using Oracle 18c

I recently had an article published on Oracle Developer Community website about Understanding, Building and Using Neural Network Machine Learning Models with Oracle 18c. I've also had a 2 Minute Tech Tip (2MTT) video about this topic and article.

Oracle 18c Database brings prominent new machine learning algorithms, including Neural Networks and Random Forests. While many articles are available on machine learning, most of them concentrate on how to build a model. Very few talk about how to use these new algorithms in your applications to score or label new data. This article will explain how Neural Networks work, how to build a Neural Network in Oracle Database, and how to use the model to score or label new data. What are Neural Networks?

Over the past couple of years, Neural Networks have attracted a lot of attention thanks to their ability to efficiently find patterns in data—traditional transactional data, as well as images, sound, streaming data, etc. But for some implementations, Neural Networks can require a lot of additional computing resources due to the complexity of the many hidden layers within the network. Figure 1 gives a very simple representation of a Neural Network with one hidden layer. All the inputs are connected to a neuron in the hidden layer (red circles). A neuron takes a set of numeric values as input and maps them to a single output value. (A neuron is a simple multi-input linear regression function, where the output is passed through an activation function.) Two common activation functions are logistic and tanh functions. There are many others, including logistic sigmoid function, arctan function, bipolar sigmoid function, etc.

Continue reading the rest of the article here.

R vs Python vs SQL for Machine Learning (Infographic)

Next week I'll be giving several presentation on machine learning at Oracle Open World and Oracle Code One. In one of these presentation an evaluation of using R vs Python vs SQL will be given and discussed.

Check out the infographic containing the comparisons.

Click here to download the PDF version.

Bringing Neural Networks to Production using GraphPipe

Machine learning is a fascinating topic. It has so much potential yet very few people talk about using machine learning in production. I've been highlighting the need for this for over 20 years now and only a very small number of machine learning languages and solutions are suitable for production use. Why? maybe it is due to the commercial aspects and as many of the languages and tools are driven by the open source community, one of the last things they get round to focusing on is production deployment. Rightly they are focused at developing more and more machine learning algorithms and features for developing models, but where the real value comes is will being able to embed machine learning model scoring in production system. Maybe this why the dominant players with machine learning in enterprises are still the big old analytics companies.

Yes that was a bit a of a rant but it is true. But over the summer and past few months there has been a number of articles about production deployment.

But this is not a new topic. For example, we have Predictive Model Markup Language (PMML) around for a long time. The aim of this was to allow the interchange of models between different languages. This would mean that the data scientist could develop their models using one language and then transfer or translate the model into another language that offers the same machine learning algorithms.

But the problem with this approach is that you may end up with different results being generated by the model in the development or lab environment versus the model being used in production. Why does this happen? Well the algorithms are developed by different people/companies and everyone has their preferences for how these algorithms are implemented.

To over come this some companies would rewrite their machine learning algorithms and models to ensure that development/lab results matched the results in production. But there is a very large cost associated with this development and ongoing maintenance as the models evolved. This would occur, maybe, every 3, 6, 9, 12 months. Somethings the time to write or rewrite each new version of the model would be longer than its lifespan.

These kind of problems have been very common and has impacted on model deployment in production.

In the era of cloud we are now seeing some machine learning cloud solutions making machine learning models available using REST services. These can, very easily, allow for machine learning models to be included in production applications. You are going to hear more about this topic over the coming year.

But, despite all the claims and wonders and benefits of cloud solutions, it isn't for everyone. Maybe at some time in the future but it mightn't be for some months or years to come.

So, how can we easily add machine learning model scoring/labeling to our production systems? Well we need some sort of middleware solutions.

Given the current enthusiasm for neural networks, and the need for GPUs, means that these cannot (easily) be deployed into production applications.

There have been some frameworks put forward for how to enable this. Once such framework is called Graphpipe. This has recently been made open source by Oracle.

Graphpipe is a framework that to access and use machine learning models developed and running on different platforms. The framework allows you to perform model scoring across multiple neural networks models and create ensemble solutions based on these. Graphpipe development has been focused on performance (most other frameworks don't). It uses flatbuffers for efficient transfer of data and currently has integration with TensorFlow, PyTorch, MXNet, CNTK and via ONNX and caffe2.

Expect to have more extensions added to the framework.

Graphpipe website

Graphpipe getting started

Graphpipe blogpost

Graphpipe download

Lessor known Apache Machine Learning languages

Machine learning is a very popular topic in recent times, and we keep hearing about languages such as R, Python and Spark. In addition to these we have commercially available machine learning languages and tools from SAS, IBM, Microsoft, Oracle, Google, Amazon, etc., etc. Everyone want a slice of the machine learning market!

The Apache Foundation supports the development of new open source projects in a number of areas. One such area is machine learning. If you have read anything about machine learning you will have come across Spark, and maybe you might believe that everyone is using it. Sadly this isn't true for lots of reasons, but it is very popular. Spark is one of the project support by the Apache Foundation.

But are there any other machine learning projects being supported under the Apache Foundation that are an alternative to Spark? The follow lists the alternatives and lessor know projects: (most of these are incubator/retired/graduated Apache projects)

Flink	Flink is an open source system for expressive, declarative, fast, and efficient data analysis. Stratosphere combines the scalability and programming flexibility of distributed MapReduce-like platforms with the efficiency, out-of-core execution, and query optimization capabilities found in parallel databases. Flink was originally known as Stratosphere when it entered the Incubator. Documentation (graduated)
HORN	HORN is a neuron-centric programming APIs and execution framework for large-scale deep learning, built on top of Apache Hama. Wiki Page (Retired)
HiveMail	Hivemall is a library for machine learning implemented as Hive UDFs/UDAFs/UDTFs Apache Hivemall offers a variety of functionalities: regression, classification, recommendation, anomaly detection, k-nearest neighbor, and feature engineering. It also supports state-of-the-art machine learning algorithms such as Soft Confidence Weighted, Adaptive Regularization of Weight Vectors, Factorization Machines, and AdaDelta. Apache Hivemall offers a variety of functionalities: regression, classification, recommendation, anomaly detection, k-nearest neighbor, and feature engineering. It also supports state-of-the-art machine learning algorithms such as Soft Confidence Weighted, Adaptive Regularization of Weight Vectors, Factorization Machines, and AdaDelta. Documentation (incubator)
MADlib	Apache MADlib is an open-source library for scalable in-database analytics. It provides data-parallel implementations of mathematical, statistical and machine learning methods for structured and unstructured data. Key features include: Operate on the data locally in-database. Do not move data between multiple runtime environments unnecessarily; Utilize best of breed database engines, but separate the machine learning logic from database specific implementation details; Leverage MPP shared nothing technology, such as the Greenplum Database and Apache HAWQ (incubating), to provide parallelism and scalability. Documentation (graduated)
MXNet	A Flexible and Efficient Library for Deep Learning . MXNet provides optimized numerical computation for GPUs and distributed ecosystems, from the comfort of high-level environments like Python and R MXNet automates common workflows, so standard neural networks can be expressed concisely in just a few lines of code. Webpage (incubator)
OpenNLP	OpenNLP is a machine learning based toolkit for the processing of natural language text. OpenNLP supports the most common NLP tasks, such as tokenization, sentence segmentation, part-of-speech tagging, named entity extraction, chunking, parsing, language detection and coreference resolution. Documentation (graduated)
PredictionIO	PredictionIO is an open source Machine Learning Server built on top of state-of-the-art open source stack, that enables developers to manage and deploy production-ready predictive services for various kinds of machine learning tasks. Documentation (graduated)
SAMOA	SAMOA provides a collection of distributed streaming algorithms for the most common data mining and machine learning tasks such as classification, clustering, and regression, as well as programming abstractions to develop new algorithms that run on top of distributed stream processing engines (DSPEs). It features a pluggable architecture that allows it to run on several DSPEs such as Apache Storm, Apache S4, and Apache Samza. Documentation (incubator)
SINGA	SINGA is a distributed deep learning platform. An intuitive programming model based on the layer abstraction is provided, which supports a variety of popular deep learning models. SINGA architecture supports both synchronous and asynchronous training frameworks. Hybrid training frameworks can also be customized to achieve good scalability. SINGA provides different neural net partitioning schemes for training large models. Documentation (incubator)
Storm	Storm is a distributed, fault-tolerant, and high-performance realtime computation system that provides strong guarantees on the processing of data. Storm makes it easy to reliably process unbounded streams of data, doing for realtime processing what Hadoop did for batch processing. Storm is simple, can be used with any programming language. Documentation (graduated)
SystemML	SystemML provides declarative large-scale machine learning (ML) that aims at flexible specification of ML algorithms and automatic generation of hybrid runtime plans ranging from single node, in-memory computations, to distributed computations such as Apache Hadoop MapReduce and Apache Spark. Documentation (graduated)

I will have a closer look that the following SQL based machine learning languages in a lager blog post:

- MADlib

- Storm

Oracle Code Online December 2017

This week Oracle Code will be having an online event consisting of 5 tracks and with 3 presentations on each track. This online Oracle Code event will be given in 3 different geographic regions on 12th, 13th and 14th December. I've been selected to give one of these talks, and I've given this talk at some live Oracle Code events and at JavaOne back in October. The present is pre-recorded and I recorded this video back in September. I hope to be online at the end of some of these presentations to answer any questions, but unfortunately due to changes with my work commitments I may not be able to be online for all of them. The moderator for these events will take your questions (or you can send them to me here) and I will write a blog post answering all your questions.

Part 5 - The right to be forgotten (EU GDPR)s

This is the fifth part of series of blog posts on 'How the EU GDPR will affect the use of Machine Learning'

Article 17 is titled Right of Erasure (right to be forgotten) allows a person to obtain their data and for the data controller to ensure that the personal data is erased without any any delay.

This does not mean that their data can be flagged for non-contact, as I've seen done in many companies, only for the odd time when one of these people have been contacted.

It will also allow for people to choose to not take part in data profiling. Meaning that these people cannot be included in any of the input data sets. And should not be the scenario where they are included but they are flagged as not to be contacted in any post ML process where the consumers are contacted, just like I've seen in lots of places.

Click back to 'How the EU GDPR will affect the use of Machine Learning - Part 1' for links to all the blog posts in this series.

Part 4b - (Article 22: Profiling) Why me? and how Oracle 12c saves the day

This is the fourth part of series of blog posts on 'How the EU GDPR will affect the use of Machine Learning'

In this blog post (Part4b) I will examine some of the more technical aspects and how the in-database machine learning functions saves the day!

Probably in most cases where machine learning has been used and/or deployed in your company to analyse, profile and predict customers, it is more than likely that some sort of black box machine learning has been used.

Typical black box machine learning will include using algorithms like Neural Networks, but these can extended to other algorithms, within the context of the EU GDPR requirements, such as SVMs, GLM, etc. Additionally most companies don't just use one algorithm to make a decision on a customer. Many algorithms and rules based decision make can be used together, using some sort of voting system, to determine if a customer is targeted in a certain way.

Basically all of these do not really support the requirements of the EU GDPRs.

In most cases we need to go back to basics. Back to more simpler approaches of machine learning for customer profiling and prediction. This means no more, for now, ensemble models, unless you can explain why a customer was selected. This means having to use simple algorithms like Decision Trees, at a push Naive Bayes, and using some well defined rules based methods. All of these approaches allows us to see and understand why a customer was selected and based on Article 22 being able to explain why.

But there is some hope. Some of the commercial machine learning vendors already for some prediction insights built into their software. Very few if any open source solutions have this capability.

For example, Oracle introduced a new function called PREDICTION_DETAILS in Oracle 12.1c and this was expanded in Oracle 12.2c to cover all their in-database machine learning algorithms.

The following is an example of using this function for an SVM model. When you examine the boxes in the following image you an see that a slightly different set of attributes and the values of these attributes are listed. Each box corresponds to a different customer. This means we can give an explanation of why a customer was selected. Oracle 12c saves the day.

select cust_id, 
       prediction(clas_svm_1_27 using *) pred_value, 
       prediction_probability(clas_svm_1_27 using *) pred_prob, 
       prediction_details(clas_svm_1_27 using *) pred_details 
from mining_data_apply_v;

If you have a look at other commercial machine learning solutions, you will find some give similar functionality or it will be available soon. Can we get the same level of detail from open source solutions. Not really unless you are using Decision Tress and maybe Naive Bayes. This means that companies that have gone done the pure open source for their machine learning may have to look at using alternative software and may have to folk out some hard earned dollars/euros to make sure that they are complainant with Article 22 of the EU GDPRs.

Click back to 'How the EU GDPR will affect the use of Machine Learning - Part 1' for links to all the blog posts in this series.