How advanced OCR found new life in big data systems by Anna Johansson via @infomgmt

Today, optical character recognition, in combination with natural language processing, allows businesses to perform complex data extraction tasks.

A great idea - use OCR to scan in old paper documents to fill the gaps in your online data - you will never get accurate results on analytics if you are missing data.

o succeed at digital transformation, do a better job of data governance by Darren Cooper via @infomgmt

To set the stage for initiatives like AI and machine learning, companies need a rock-solid governance framework.

Great suggestions by Darren in this article.

GDPR compliance the perfect opportunity to modernise data architecture by Amandeep Khurana via @infomgmt

Compliance with the data privacy and security mandate enables organisations to become more agile in their product and service development and rollouts, and more efficient and effective in their ability to respond to market trends and competitive threats.

Yes this is exactly right - everything has to turn onto it's head and be data centric not application centric. I think we need to concentrate on:

WHERE is the data created
WHERE is it also stored (so where is it interfaced to)
HOW it is updated
WHAT changes when it is updated
HOW do you delete the data in ALL systems?

I would suggest you do something like a data flow diagram so you can document all of this for every piece of data.

The bias problem with artificial intelligence, and how to solve it by Sanjay Srivastava via @infomgmt

AI bias may come from incomplete datasets or incorrect values. Bias may also emerge through interactions overtime, skewing the machine’s learning. Moreover, a sudden business change, such as a new law or business rule, or ineffective training algorithms can also cause bias.

I agree - you need good quality and representative training data if you want to get good results from any AI and ML you want to use. My advice would be:

1.  Take your time - rushing always leads to mistakes so be realistic with plans.
2.  Be careful with the methodology you use to create and split your data into Training and Data.
3.  Try to use separate teams to test the same piece of code - the hope being that it will help to avoid the bias. Think of it as a human version of a small parallel ML solution.
4.  Check, check and check again.

Community lenders tell big tech vendors to get up to speed by Nathan DiCamillo via @infomgmt

Small banks and credit unions say slow responses and outdated products from the establishment tech vendor can become a drag on their innovation efforts.

I partially agree with him - yes large organisations move slow (particularly when you are a small customer and therefore your business is not a big loss to them if you move on) but small ones are less stable and sometimes that can be an unacceptable risk to the business (particularly in the financial sector where you just cannot afford an issue). So do really careful risk management and have SLAs to protect yourself.

WEBINAR: Getting Data Down to a Science – Code-free and Code-friendly ML - 5th September 2018

Data Science helps answer some of the most basic - and the most complex - business questions. In this latest Data Science Central webinar you will learn how to get data down to a science with code-free and code-friendly self-service analytics platforms. Decisive Data’s Lead Data Scientist Tessa Jones will use a sample data set from a global corporation to answer some of the most common data science questions applicable across businesses. Learn how to use code-free and code-friendly Machine Learning: Dive – Swim in the data and dive into a few common business questions with answers in data science including demand forecasting and customer segmentation. Build – Walk through two data science models including code-free time series and clustering machine learning models. Customize – Implement custom R code into models. Refine – Enhance your methods with rapid self-service techniques. Display – Creatively display information visually in Tableau and tell a story that makes the findings clear and captivating using the Art + Data methodology. Speakers: Tessa Jones, Lead Data Scientist -- Decisive Data Scott Trauthen, Director of Marketing -- Alteryx Hosted by: Bill Vorhies, Editorial Director -- Data Science Central   Title:  Getting Data Down to a Science – Code-free and Code-friendly Machine Learning Date:  Wednesday, September 5th, 2018 Time:  9 AM - 10 AM PDT

Join here

WEBINAR: Production ML for Data Scientists: What You Can Do and How to Make it Easy - 22 August 2018

Production ML for Data Scientists: What You Can Do and How to Make it Easy August 22, 2018  |  10am PT/1pm ET

For many data scientists in the enterprise, the deployment of machine learning into production environments has become a second job - and one that most do not want. Current IT and operations teams and tools can't account for the complexities of deploying, managing and scaling ML applications, leaving data science and data engineering teams on the hook for the success - or failure - of ML and AI initiatives. In this webinar, data scientists will be introduced to MLOps - an approach for machine learning operationalization that: Breaks down the silos between data science and IT Streamlines deployment and orchestration Adds advanced functionality like ML Health, governance and business metrics Get Your ML Experiments to Production On August 22 at 10am PT/1pm ET, join Nisha Talagala, CTO, and Craig Michaud, Sales Engineer, from ParallelM - the MLOps Company - for a look at how much easier machine learning can be with the right technology and processes in place. You'll see how to upload code from your existing data science platforms, run it in a sandbox against production data, conduct AB tests and perform timeline captures.  Register here