Benjamin Aunkofer

Data Science vs Data Engineering

/in , , , , , , /by

The job of the Data Scientist is actually a fairly new trend, and yet other job titles are coming to us. “Is this really necessary?”, Some will ask. But the answer is clear: yes!

There are situations, every Data Scientist know: a recruiter calls, speaks about a great new challenge for a Data Scientist as you obviously claim on your LinkedIn profile, but in the discussion of the vacancy it quickly becomes clear that you have almost none of the required skills. This mismatch is mainly due to the fact that under the job of the Data Scientist all possible activity profiles, method and tool knowledge are summarized, which a single person can hardly learn in his life. Many open jobs, which are to be called under the name Data Science, describe rather the professional image of the Data Engineer.

Read this article in German:
“Data Science vs Data Engineering – Wo liegen die Unterschiede?“

What is a Data Engineer?

Data engineering is primarily about collecting or generating data, storing, historicalizing, processing, adapting and submitting data to subsequent instances. A Data Engineer, often also named as Big Data Engineer or Big Data Architect, models scalable database and data flow architectures, develops and improves the IT infrastructure on the hardware and software side, deals with topics such as IT Security , Data Security and Data Protection. A Data Engineer is, as required, a partial administrator of the IT systems and also a software developer, since he or she extends the software landscape with his own components. In addition to the tasks in the field of ETL / Data Warehousing, he also carries out analyzes, for example, to investigate data quality or user access. A Data Engineer mainly works with databases and data warehousing tools.

A Data Engineer is talented as an educated engineer or computer scientist and rather far away from the actual core business of the company. The Data Engineer’s career stages are usually something like:

  1. (Big) Data Architect
  2. BI Architect
  3. Senior Data Engineer
  4. Data Engineer

What makes a Data Scientist?

Although there may be many intersections with the Data Engineer’s field of activity, the Data Scientist can be distinguished by using his working time as much as possible to analyze the available data in an exploratory and targeted manner, to visualize the analysis results and to convert them into a red thread (storytelling). Unlike the Data Engineer, a data scientist rarely sees into a data center, because he picks up data via interfaces provided by the Data Engineer or provides by other resources.

A Data Scientist deals with mathematical models, works mainly with statistical procedures, and applies them to the data to generate knowledge. Common methods of Data Mining, Machine Learning and Predictive Modeling should be known to a Data Scientist. Data Scientists basically work close to the department and need appropriate expertise. Data Scientists use proprietary tools (e.g. Tools by IBM, SAS or Qlik) and program their own analyzes, for example, in Scala, Java, Python, Julia, or R. Using such programming languages and data science libraries (e.g. Mahout, MLlib, Scikit-Learn or TensorFlow) is often considered as advanced data science.

Data Scientists can have diverse academic backgrounds, some are computer scientists or engineers for electrical engineering, others are physicists or mathematicians, not a few have economical backgrounds. Common career levels could be:

  1. Chief Data Scientist
  2. Senior Data Scientist
  3. Data Scientist
  4. Data Analyst oder Junior Data Scientist

Data Scientist vs Data Analyst

I am often asked what the difference between a Data Scientist and a Data Analyst would be, or whether there would be a distinction criterion at all:

In my experience, the term Data Scientist stands for the new challenges for the classical concept of Data Analysts. A Data Analyst performs data analysis like a Data Scientist. More complex topics such as predictive analytics, machine learning or artificial intelligence are topics for a Data Scientist. In other words, a Data Scientist is a Data Analyst++ (one step above the Data Analyst).

And how about being a Business Analyst?

Business Analysts can (but need not) be Data Analysts. In any case, they have a very strong relationship with the core business of the company. Business Analytics is about analyzing business models and business successes. The analysis of business success is usually carried out by IT, and many business analysts are starting a career as Data Analyst now. Dashboards, KPIs and SQL are the tools of a good business analyst, but there might be a lot business analysts, who are just analysing business models by reading the newspaper…

Benjamin Aunkofer

Data Science Knowledge Stack – Abstraction of the Data Science Skillset

/in , , , , , , , , /by

What must a Data Scientist be able to do? Which skills does as Data Scientist need to have? This question has often been asked and frequently answered by several Data Science Experts. In fact, it is now quite clear what kind of problems a Data Scientist should be able to solve and which skills are necessary for that. I would like to try to bring this consensus into a visual graph: a layer model, similar to the OSI layer model (which any data scientist should know too, by the way).
I’m giving introductory seminars in Data Science for merchants and engineers and in those seminars I always start explaining what we need to work out together in theory and practice-oriented exercises. Against this background, I came up with the idea for this layer model. Because with my seminars the problem already starts: I am giving seminars for Data Science for Business Analytics with Python. So not for medical analyzes and not with R or Julia. So I do not give a general knowledge of Data Science, but a very specific direction.

A Data Scientist must deal with problems at different levels in any Data Science project, for example, the data access does not work as planned or the data has a different structure than expected. A Data Scientist can spend hours debating its own source code or learning the ropes of new DataScience packages for its chosen programming language. Also, the right algorithms for data evaluation must be selected, properly parameterized and tested, sometimes it turns out that the selected methods were not the optimal ones. Ultimately, we are not doing Data Science all day for fun, but for generating value for a department and a data scientist is also faced with special challenges at this level, at least a basic knowledge of the expertise of that department is a must have.

Read this article in German:
“Data Science Knowledge Stack – Was ein Data Scientist können muss“

Data Science Knowledge Stack

With the Data Science Knowledge Stack, I would like to provide a structured insight into the tasks and challenges a Data Scientist has to face. The layers of the stack also represent a bidirectional flow from top to bottom and from bottom to top, because Data Science as a discipline is also bidirectional: we try to answer questions with data, or we look at the potentials in the data to answer previously unsolicited questions.

The DataScience Knowledge Stack consists of six layers:

Database Technology Knowledge

A Data Scientist works with data which is rarely directly structured in a CSV file, but usually in one or more databases that are subject to their own rules. In particular, business data, for example from the ERP or CRM system, are available in relational databases, often from Microsoft, Oracle, SAP or an open source alternative. A good Data Scientist is not only familiar with Structured Query Language (SQL), but is also aware of the importance of relational linked data models, so he also knows the principle of data table normalization.

Other types of databases, so-called NoSQL databases (Not only SQL) are based on file formats, column or graph orientation, such as MongoDB, Cassandra or GraphDB. Some of these databases use their own programming languages ​​(for example JavaScript at MongoDB or the graph-oriented database Neo4J has its own language called Cypher). Some of these databases provide alternative access via SQL (such as Hive for Hadoop).

A data scientist has to cope with different database systems and has to master at least SQL – the quasi-standard for data processing.

Data Access & Transformation Knowledge

If data are given in a database, Data Scientists can perform simple (and not so simple) analyzes directly on the database. But how do we get the data into our special analysis tools? To do this, a Data Scientist must know how to export data from the database. For one-time actions, an export can be a CSV file, but which separators and text qualifiers should be used? Possibly, the export is too large, so the file must be split.
If there is a direct and synchronous data connection between the analysis tool and the database, interfaces like REST, ODBC or JDBC come into play. Sometimes a socket connection must also be established and the principle of a client-server architecture should be known. Synchronous and asynchronous encryption methods should also be familiar to a Data Scientist, as confidential data are often used, and a minimum level of security is most important for business applications.

Many datasets are not structured in a database but are so-called unstructured or semi-structured data from documents or from Internet sources. And again we have interfaces, a frequent entry point for Data Scientists is, for example, the Twitter API. Sometimes we want to stream data in near real-time, let it be machine data or social media messages. This can be quite demanding, so the data streaming is almost a discipline with which a Data Scientist can come into contact quickly.

Programming Language Knowledge

Programming languages ​​are tools for Data Scientists to process data and automate processing. Data Scientists are usually no real software developers and they do not have to worry about software security or economy. However, a certain basic knowledge about software architectures often helps because some Data Science programs can be going to be integrated into an IT landscape of the company. The understanding of object-oriented programming and the good knowledge of the syntax of the selected programming languages ​​are essential, especially since not every programming language is the most useful for all projects.

At the level of the programming language, there is already a lot of snares in the programming language that are based on the programming language itself, as each has its own faults and details determine whether an analysis is done correctly or incorrectly: for example, whether data objects are copied or linked as reference, or how NULL/NaN values ​​are treated.

Data Science Tool & Library Knowledge

Once a data scientist has loaded the data into his favorite tool, for example, one of IBM, SAS or an open source alternative such as Octave, the core work just began. However, these tools are not self-explanatory and therefore there is a wide range of certification options for various Data Science tools. Many (if not most) Data Scientists work mostly directly with a programming language, but this alone is not enough to effectively perform statistical data analysis or machine learning: We use Data Science libraries (packages) that provide data structures and methods as a groundwork and thus extend the programming language to a real Data Science toolset. Such a library, for example Scikit-Learn for Python, is a collection of methods implemented in the programming language. The use of such libraries, however, is intended to be learned and therefore requires familiarization and practical experience for reliable application.

When it comes to Big Data Analytics, the analysis of particularly large data, we enter the field of Distributed Computing. Tools (frameworks) such as Apache Hadoop, Apache Spark or Apache Flink allows us to process and analyze data in parallel on multiple servers. These tools also provide their own libraries for machine learning, such as Mahout, MLlib and FlinkML.

Data Science Method Knowledge

A Data Scientist is not simply an operator of tools, he uses the tools to apply his analysis methods to data he has selected for to reach the project targets. These analysis methods are, for example, descriptive statistics, estimation methods or hypothesis tests. Somewhat more mathematical are methods of machine learning for data mining, such as clustering or dimensional reduction, or more toward automated decision making through classification or regression.

Machine learning methods generally do not work immediately, they have to be improved using optimization methods like the gradient method. A Data Scientist must be able to detect under- and overfitting, and he must prove that the prediction results for the planned deployment are accurate enough.

Special applications require special knowledge, which applies, for example, to the fields of image recognition (Visual Computing) or the processing of human language (Natural Language Processiong). At this point, we open the door to deep learning.

Expertise

Data Science is not an end in itself, but a discipline that would like to answer questions from other expertise fields with data. For this reason, Data Science is very diverse. Business economists need data scientists to analyze financial transactions, for example, to identify fraud scenarios or to better understand customer needs, or to optimize supply chains. Natural scientists such as geologists, biologists or experimental physicists also use Data Science to make their observations with the aim of gaining knowledge. Engineers want to better understand the situation and relationships between machinery or vehicles, and medical professionals are interested in better diagnostics and medication for their patients.

In order to support a specific department with his / her knowledge of data, tools and analysis methods, every data scientist needs a minimum of the appropriate skills. Anyone who wants to make analyzes for buyers, engineers, natural scientists, physicians, lawyers or other interested parties must also be able to understand the people’s profession.

Engere Data Science Definition

While the Data Science pioneers have long established and highly specialized teams, smaller companies are still looking for the Data Science Allrounder, which can take over the full range of tasks from the access to the database to the implementation of the analytical application. However, companies with specialized data experts have long since distinguished Data Scientists, Data Engineers and Business Analysts. Therefore, the definition of Data Science and the delineation of the abilities that a data scientist should have, varies between a broader and a more narrow demarcation.


A closer look at the more narrow definition shows, that a Data Engineer takes over the data allocation, the Data Scientist loads it into his tools and runs the data analysis together with the colleagues from the department. According to this, a Data Scientist would need no knowledge of databases or APIs, neither an expertise would be necessary …

In my experience, DataScience is not that narrow, the task spectrum covers more than just the core area. This misunderstanding comes from Data Science courses and – for me – I should point to the overall picture of Data Science again and again. In courses and seminars, which want to teach Data Science as a discipline, the focus will of course be on the core area: programming, tools and methods from mathematics & statistics.

Thomas Blanchard

Data Science and Predictive Analytics in Healthcare

/in , , , , /by

Doing data science in a healthcare company can save lives. Whether it’s by predicting which patients have a tumor on an MRI, are at risk of re-admission, or have misclassified diagnoses in electronic medical records are all examples of how predictive models can lead to better health outcomes and improve the quality of life of patients.  Nevertheless, the healthcare industry presents many unique challenges and opportunities for data scientists.

The impact of data science in healthcare

Healthcare providers have a plethora of important but sensitive data. Medical records include a diverse set of data such as basic demographics, diagnosed illnesses, and a wealth of clinical information such as lab test results. For patients with chronic diseases, there could be a long and detailed history of data available on a number of health indicators due to the frequency of visits to a healthcare provider. Information from medical records can often be combined with outside data as well. For example, a patient’s address can be combined with other publicly available information to determine the number of surgeons that practice near a patient or other relevant information about the type of area that patients reside in.

With this rich data about a patient as well as their surroundings, models can be built and trained to predict many outcomes of interest. One important area of interest is models predicting disease progression, which can be used for disease management and planning. For example, at Fresenius Medical Care (where we primarily care for patients with chronic conditions such as kidney disease), we use a Chronic Kidney Disease progression model that can predict the trajectory of a patient’s condition to help clinicians decide whether and when to proceed to the next stage in their medical care. Predictive models can also notify clinicians about patients who may require interventions to reduce risk of negative outcomes. For instance, we use models to predict which patients are at risk for hospitalization or missing a dialysis treatment. These predictions, along with the key factors driving the prediction, are presented to clinicians who can decide if certain interventions might help reduce the patient’s risk.

Challenges of data science in healthcare

One challenge is that the healthcare industry is far behind other sectors in terms of adopting the latest technology and analytics tools. This does present some challenges, and data scientists should be aware that the data infrastructure and development environment at many healthcare companies will not be at the bleeding edge of the field. However it also means there are a lot of opportunities for improvement, and even small simple models can yield vast improvements over current methods.

Another challenge in the healthcare sector arises from the sensitive nature of medical information. Due to concerns over data privacy, it can often be difficult to obtain access to data that the company has. For this reason, data scientists considering a position at a healthcare company should be aware of whether there is already an established protocol for data professionals to get access to the data. If there isn’t, be aware that simply getting access to the data may be a major effort in itself.

Finally, it is important to keep in mind the end-use of any predictive model. In many cases, there are very different costs to false-negatives and false-positives. A false-negative may be detrimental to a patient’s health, while too many false-positives may lead to many costly and unnecessary treatments (also to the detriment of patients’ health for certain treatments as well as economy overall). Education about the proper use of predictive models and their limitations is essential for end-users. Finally, making sure the output of a predictive model is actionable is important. Predicting that a patient is at high-risk is only useful if the model outputs is interpretable enough to explain what factors are putting that patient at risk. Furthermore, if the model is being used to plan interventions, the factors that can be changed need to be highlighted in some way – telling a clinician that a patient is at risk because of their age is not useful if the point of the prediction is to lower risk through intervention.

The future of data science in the healthcare sector

The future holds a lot of promise for data science in healthcare. Wearable devices that track all kinds of activity and biometric data are becoming more sophisticated and more common. Streaming data coming from either wearables or devices providing treatment (such as dialysis machines) could eventually be used to provide real-time alerts to patients or clinicians about health events outside of the hospital.

Currently, a major issue facing medical providers is that patients’ data tends to exist in silos. There is little integration across electronic medical record systems (both between and within medical providers), which can lead to fragmented care. This can lead to clinicians receiving out of date or incomplete information about a patient, or to duplication of treatments. Through a major data engineering effort, these systems could (and should) be integrated. This would vastly increase the potential of data scientists and data engineers, who could then provide analytics services that took into account the whole patients’ history to provide a level of consistency across care providers. Data workers could use such an integrated record to alert clinicians to duplications of procedures or dangerous prescription drug combinations.

Data scientists have a lot to offer in the healthcare industry. The advances of machine learning and data science can and should be adopted in a space where the health of individuals can be improved. The opportunities for data scientists in this sector are nearly endless, and the potential for good is enormous.

Vishal Bhalla

A “Dialogue” on the recent advances in Conversational Artificial Intelligence (AI)

/in , /by

How important is it to interact, converse and emote in a world that is getting closed and parochial? Conversational Artificial Intelligence (AI) offers a leeway to build agents that have the capability to learn and respond like humans and thereby align in bringing the long term goal of General AI to fruition.

Conversation with artificial assistants, be it Microsoft’s Cortana, Apple’s Siri, Google Now or Amazon’s Alexa is gaining prominence in the last few years. So lay back, relax and enjoy the simple conversational interface at offer, as I take you through a short tour!

In this 2 part blog-series, I cover the latest developments in the field of Dialogue and conversational Artificial Intelligence (AI). I give a brief overview of the current developments from this field, the many Language Understanding tools in the market and in particular, review one of them – IBM Conversation.

It’s a rat race – So act and don’t over think!

After the horrors of Tay tweets -Microsoft’s conversational AI tweet bot that was eventually rolled back due to its racist and sexist tweets early this year, AI enthusiasts have had some good news over the last few months.

nycitizen07-tweet

Microsoft hurried the launch of Tay tweets, its conversational AI bot only to shun it completely.

The Amazon Echo, Google’s Home and the smart home hub Apple has been preparing are good examples of how big companies are fighting tooth and nail to secure a place on your smart space. Here’s what Francis Chollet, researcher at Google and author of the popular framework – Keras has to say,

Whatever idea you started working on last week, a few other teams have probably been working on it for a month and are about to publish.
— François Chollet (@fchollet) October 5, 2016

Alexa Prize Competition

Just 4 weeks back, Amazon announced the Alexa Prize, an annual competition for university students dedicated to accelerating the field of conversational AI. This inaugural competition focuses on creating a social bot, using the Alexa Skills Kit (ASK) to converse coherently and engaging with humans on popular topics and news events. This gives student developer teams to explore a plethora of advanced topics in the realm of AI that include knowledge acquisition, natural language understanding, natural language generation, context modeling, commonsense reasoning and dialog planning. With a huge cash prize at stake, goodies at offer and support from the ASK team it would be worth an experience to build a socially coherent bot!  The last date of team submissions is October 28, 2016 and more details about the application process can be found here.

Say Allo!

Google Allo, a smart messaging app that has personalized recommendations with the Google Assistant to express yourself better with stickers, doodles, and HUGE emojis & text. Allo also allows you to get help from your Google Assistant without leaving the conversation. A one to one conversation can be initiated with your Assistant which gets better as you use it more by addressing it with the @google tag. More functional details on the blog Say hello to Google Allo: a smarter messaging app

IBM Pepper developer Conference

The IBM BusinessConnect 2016 on 4th October 2016 in Stockholm, Sweden showcased some of IBM Watson powered tools, and applications in humanoid robot of Pepper.

Yesterdays #IBMBCSE at Stockholm Waterfront was fantastic thanks to all IBMers, partners and customers, and thanks to #Pepper of course! pic.twitter.com/quZuaptu8Z
— IBM ClientCtr Nordic (@IBMCCNordic) October 5, 2016

IBM’s Pepper is powered by SoftBank robot and uses IBM Watson technology at its core.

Banzai! (Live long) – Watch this first home robot commercial as the unforeseen future is coming!

The Watson Developer Conference is packed with technical talks, hands-on labs, and coding challenges to get you working with the tools that will make you a sought after developer and is going to be held in San Francisco from 9th to 10th November this year.

ibm-robot

The IBM Global Industry Solution is located in Nice, France.

Joie de vivre – Samsung buys Viv

And after Google’s Allo and IBM’s pepper it was Samsung to jump into the Dialogue based conversational AI bandwagon as it acquired Viv, creators of Apple’s Siri. Viv is a more powerful version to Siri that brings in ubiquity. With its self-generating software that is capable of writing its own code to accomplish new tasks and by dynamic program generation, Viv handles new user tasks and build plans on the fly!

In its demo video on “Beyond Siri: The World Premiere of Viv with Dag Kittlaus” (as in the embedded link/video below) earlier this year, Viv was eventually be partnered or sold to a mobile device.

With everyone wanting to invest heavily, the question was who and when! Hence, this announcement from Samsung doesn’t come as a big surprise.

Viv will ultimately provide services to Samsung and its platforms but remain an independent entity. Samsung hopes to disrupt the mobile market share with this acquisition. It can extend it to other home devices, after all it had purchased SmartThings for around $200M back in 2014. More details on the acquisition here: Samsung acquires Viv, a next-gen AI assistant built by the creators of Apple’s Siri

Don’t take it slow because there is Ozlo !

Ozlo launched few days back on iOS and the web is another of the many sprouting AI assistants which uses good memory of one’s previous interactions. Ozlo, at least by its name attempts to be different than all assistants of its competitors in the market at present that use repetitive female names. The best thing is that it is integrated with a plethora of services like Yelp, TripAdvisor,IMDB, among many others and use  Further Food, Authority Nutrition, Cookies, etc. to provide nutritional guidance. This is a huge boost than all of its rival companies which tend to prioritize their own services rather than integrating with existing services. An in-depth review can be found here: Ozlo AI assistant is the new underdog filling the void left by Viv

And there were rumors that Apple is going to buy McLaren, which set the eyeballs rolling as a big tech giant was entering a completely new domain of automobile industry and would lead others like Google, Microsoft and IBM to follow suit and invest heavily!

Conference workshops also wanting a dialogue!

There are in total 50 workshops at NIPS 2016 this year covering a range of different Machine Learning topics.

  1. The Dialog workshop, scheduled on the 10th of December focuses on building agents capable of mutually coordinating with humans via communication. And given the tremendous economic potential of the ability to converse intimately transcends to the overall goal of AI.
    For the call for papers, the deadline is extended to the midnight of October 23, 2016 and more details about the workshop schedule can be found at the chair website LET’S DISCUSS: LEARNING METHODS FOR DIALOGUE NIPS 2016 WORKSHOP The papers are on the below three high-level areas

    • Being data-driven especially the offline/online evaluation
    • Build complete applications or end-to-end systems
    • Model innovation to incorporate linguistic knowledge into the architecture
  1. Another workshop on Interactive machine learning (IML) is to be held on the 9th of December. It focuses on the adaptable collaboration of how autonomous agents solve a task by making use of interactions with humans. Designing and engineering fully autonomous agents is a difficult and there is a compelling need for IML algorithms that enable artificial and human agents to collaborate and solve independent or shared goals.
    The call for papers explores new ideas in interactive learning, reports on research in progress as well as discussions of open problems and challenges facing interactive machine learning with particular interest in the research on the practical application of interactive learning systems (for robotics, virtual agents, dialog systems, among others), and the ability of these systems to handle the complexity of real world problems. More details about the application process, requirements, application deadline, etc. is at the workshop portal Future of Interactive Learning Machines Workshop (FILM at NIPS 2016)

In the next part of this series on Conversational AI, I would cover the basics behind Language Understanding tools in the market that enable to build a Dialogue system.

Read the second Part here: A review of Language Understanding tools – IBM Conversation

Mathias Golombek

Data Science on a large scale – can it be done?

/in , , , , , /by

Analytics drives business

In today’s digital world, data has become the crucial success factor for businesses as they seek to maintain a competitive advantage, and there are numerous examples of how companies have found smart ways of monetizing data and deriving value accordingly.

On the one hand, many companies use data analytics to streamline production lines, optimize marketing channels, minimize logistics costs and improve customer retention rates.  These use cases are often described under the umbrella term of operational BI, where decisions are based on data to improve a company’s internal operations, whether that be a company in the manufacturing industry or an e-commerce platform.

On the other hand, over the last few years, a whole range of new service-oriented companies have popped up whose revenue models wholly depend on data analytics.  These Data-Driven Businesses have contributed largely to the ongoing development of new technologies that make it possible to process and analyze large amounts of data to find the right insights.  The better these technologies are leveraged, the better their value-add and the better for their business success.  Indeed, without data and data analytics, they don’t have a business.

Data Science – hype or has it always been around?Druck

In my opinion, there is too much buzz around the new era of data scientists.  Ten years ago, people simply called it data mining, describing similar skills and methods.  What has actually changed is the fact that businesses are now confronted with new types of data sources such as mobile devices and data-driven applications rather than statistical methodologies.  I described that idea in detail in my recent post Let’s replace the Vs of Big Data with a single D.

But, of course, you cannot deny that the importance of these data crunchers has increased significantly. The art of mining data mountains (or perhaps I should say “diving through data lakes”) to find appropriate insights and models and then find the right answers to urgent, business-critical questions has become very popular these days.

The challenge: Data Science with large volumes?

Michael Stonebraker, winner of the Turing Award 2014, has been quoted as saying: “The change will come when business analysts who work with SQL on large amounts of data give way to data EXASOL Pipelinescientists, which will involve more sophisticated analysis, predictive modeling, regressions and Bayesian classification. That stuff at scale doesn’t work well on anyone’s engine right now. If you want to do complex analytics on big data, you have a big problem right now.”

And if you look at the limitations of existing statistical environments out there using R, Python, Java, Julia and other languages, I think he is absolutely right.  Once the data scientists have to handle larger volumes, the tools are just not powerful and scalable enough.  This results in data sampling or aggregation to make statistical algorithms applicable at all.

A new architecture for “Big Data Science”

We at EXASOL have worked hard to develop a smart solution to respond to this challenge.  Imagine that it is possible to use raw data and intelligent statistical models on very large data sets, directly at the place where the data is stored.  Where the data is processed in-memory to achieve optimal performance, all distributed across a powerful MPP cluster of servers, in an environment where you can now “install” the programming language of your choice.

Sounds far-fetched?  If you are not convinced, then I highly recommend you have a look at our brand-new in-database analytic programming platform, which is deeply integrated in our parallel in-memory engine and extensible through using nearly any programming language and statistical library.

For further information on our approach to big data science, go ahead and download a copy of our technical whitepaper:  Big Data Science – The future of analytics.

Dr. Robert Suhada

A quick primer on TensorFlow – Google’s machine learning workhorse

/in , , , , , /by

Introducing Google Brains‘ TensorFlow™

This week started with major news for the machine learning and data science community: the Google Brain Team announced the open sourcing of TensorFlow, their numerical library for tensor network computations. This software is actively developed (and used!) within Google and builds on many of Google’s large scale neural network applications such as automatic image labeling and captioning as well as the speech recognition in Google’s apps.

TensorFlow in bullet points

Here are the main features:

  • Supports deep neural networks – and much more machine learning approaches
  • Highly scalable across many machines and huge data sets
  • Runs on desktops, servers, in cloud and even mobile devices
  • Computation can run on CPUs, GPUs or both
  • All this flexibility is covered by a single API making the execution very streamlined
  • Available interfaces: C++ and Python. More will follow (Java, R, Lua, Go…)
  • Comes with many tools helping to build and visualize the data flow networks
  • Includes a powerful gradient based optimizer with auto-differentiation
  • Extensible with C++
  • Usable for commercial applications – released under Apache Software Licence 2.0

Tensor, what? Tensor, why?

„Numerical library for tensor network computations“ maybe doesn’t sound too exciting, but let’s  consider the implications.

Application of tensors and their networks is a relatively new (but fast evolving) approach in machine learning. Tensors, if you recall your algebra classes, are simply n-dimensional data arrays (so a scalar is a 0th order tensor, a vector is 1st order, and a matrix a 2nd order matrix).

A simple practical example of is color image’s RGB layers (essentially three 2D matrices combined into a 3rd order tensor). Or a more business minded example – if your data source generates a table (a 2D array) every hour, you can look at the full data set as a 3rd order tensor – time being the extra dimension.

Tensor networks then represent “data flow graphs”, where the edges are your multi-dimensional data sets and nodes are the mathematical operations on this data.

Example of of a data flow graph with multiple nodes (data operations). Notice how the execution of nodes is asynchronous. This allows incredible scalability across many machines. Image Source.

Looking at your data through the tensor formalism gives you a lot of powerful tools that were already developed for tensor algebra, allowing fast, complex computations.  

Tensor networks are also a natural fit for computations done on graphical processing units (GPUs) as they are built exactly for the purpose of very fast numerical operations on such a data – speeding up your calculations significantly compared to standard CPU execution!

The importance of flexible architecture & scaling

The data flow graph approach has also further advantages. Most notably, you can split the design of your data flows (i.e. data cleaning, processing, transformations, model building etc.) from its execution. You first build up the graph of your data flow and then you send it to for execution: either on the CPUs of your machines (and it can be your laptop just as well as cluster) or GPUs or a combination. This happens through a single interface that hides all the complexities from you.

Since the execution is asynchronous it scales across many machines and can deal with huge amounts of data.

You can count on the Google guys to build tools not only for academic use, but also heavy-duty operations in the industry!

Is this just another deep learning library?

TensorFlow is of course not the first library to embrace the tensor formalism and GPU execution. The nearest comparisons (and competitors) are Theano, Torch and CGT (Caffe to a limited degree).

While there are significant overlaps between the libraries, TensorFlow tries to provide a broader framework. It is not only a deep learning library – the Data Flow Graphs can incorporate any data processing/analysis applications. It also comes with a very powerful gradient based optimizer with automatic calculations of derivatives offering huge flexibility.

Given this broad vision the closest competitor is probably Theano (while Caffe and the existing Theano wrappers have a narrower focus on deep learning). TensorFlow’s distinguishing feature is that by design its focus is on large, scalable architectures with a complete flexibility in the hardware, best suited for industry/operational use, whereas the other libraries have more academic pedigrees.

Initial analyses also indicate that TensorFlow should bring also performance improvements compared to Theano, although no comprehensive benchmarks have yet been published.

As the other packages are out already for a while, they have large, active communities and often additional supporting software (examples are the very useful wrappers around Theano like Lasagne, Keras and Blocks that provider higher level abstractions to its engine).

Of course, with Google’s gravitas, one can expect that TensorFlow’s open source community will grow very fast and the contributors will quickly add a lot of additional features (and find hidden bugs).

Finally, keep in mind, that while Google provided us with this great data processing framework and some of its machine learning capabilities, it is likely that the most powerful machine learning algorithms still remain Google’s proprietary secret.

Nonetheless, TensorFlow is a huge and very welcome contribution to the open source machine learning world!

Where to go next?

You can find Google’s getting started guide here. The TensorFlow white paper is worth a read too. Source code can be found at the Github page. There is also a Vagrant virtual machine with TensorFlow pre-installed available here.