Machine Learning and Data Analysis Bias Challenges

One of the challenges of doing machine learning is getting lots of data. One of the other challenges of machine learning is having a lot of data.

This is a bit of a dichotomy with the data. A data scientist or machine learning hacker needs a lot of data to find accuracy in their model predicting something. The challenge can be with all the features (or columns in a data set) that represent potential items that influence the decision. Which of these do you choose?

I saw a post about looking over data and analyzing loans for risk. The post looks at how you might use a notebook to perform some analysis and does a good job of walking through the process with Databricks. It then builds, trains, and tests a model in the classic machine learning sense. The data is a public dataset from Lending Club, with actual loan data.

The post doesn't show the complete data set, and I didn't download it. I hope there isn't any PII data in here, but I'm sure that many people that actually analyze loans would have data related to a person's address, gender, and other PII type data. That makes sense for loan companies that need to create a financial contract with an individual, but it also allows humans to use their bias (intentionally or unintentionally) to affect decisions.

One of the promises of machine learning and AI is that this bias won't be present because the machines don't have feelings about individuals. Except, the data can reflect past feelings. If the data tends to show that people in a low-income neighborhood don't get loans, then the machine might pick this as a feature to look for in future data sets. Same for address or any other number of variables. Certainly, the person training the model may account for this and supervise the training, but I worry about this. Far too many people trust computers to do the right thing, but they are only as good as the data.

Garbage in, garbage out. A common phrase many of us have used, and it's applicable here. Often our data sets reflect the frailty and mistakes of past human decisions and may not be the best choices for how to train machines that will assist us in the future. They are useful, but we ought to be aware and cognizant of the bias that might exist in our data.

I do think machine learning and artificial intelligence are useful in our world. They improve our lives, and they can help reduce the common mistakes that humans can make and improve our lives. They also can be flawed, and we ought to be careful about how much we rely on them. Certainly the early deployment of any type of model should be controlled, limited, and carefully monitored. Whether this is for a model driving a car, approving loans, or identifying pictures. All your training and testing will likely be with limited data, so be wary when releasing these items into the wider world.

Steve Jones - SSC Editor

Join the debate, and respond to the editorial on the forums

Our series of ‘DBAs at work’ blog posts feature conversations with IT leaders and experts about the challenges of managing and monitoring their server estates. Episode 1 features Dennis Heitmann, Database Administrator at Atruvia, which provides IT services for banking clients in Germany.

The Estate Configuration reports SQL Server configuration across all your servers, so you can quickly investigate ad-hoc or unauthorized changes to any settings that might affect their performance, stability, or security.

Demonstrates a cross-database PowerShell callback script for reporting on and auditing Flyway migrations, telling you which scripts were used to create each version, when they were run, who ran them and more.

Attention is one of most successful deep learning network architecture in state-of-the-art NLP works in AI, and attention can also be applied in visual recognition. In this post, I'll show...

In its simplest form, AIOps (Artificial Intelligent Operations) is a marriage of AI technology that’s been around for decades. AIOps arrived at its current incarnation from two different directions. ...

While Finance may be a complex and multifaceted industry, the core goal of any financial institute is very straightforward: to detect and mitigate risks while maximizing profit. This objective...

Ok, I get it, scheduling queries can be complicated. See this and this and maybe this and this too but only if you have insomnia. I still thought I kinda understood it. Then I started seeing hundreds of query timeouts on a quiet server, where the queries were waiting on…what?

I got a call this morning from a coworker. One of the database instances was unreachable with the message that the TempDB log file was full. No processes could login and the only way to get things going again might be to restart either server or instance.

At Stedman Solutions, we do a lot of work with SQL...

Ok, I get it, scheduling queries can be complicate...

Many years ago, in the company I was working for, one junior DBA started a reindex operation in a SQL Server Standard Edition on the most busy day of...

This article will explore the Copy Data tool for importing data into Azure SQL Database from a web source. Introduction Suppose you need to import data into Azure database...

This article will help you understand Azure SQL and its deployment options available in the Azure cloud. This article would be helpful to you if you are planning to...

In case you are not aware Microsoft have now deployed a new change to SQL Managed Instances within the tier types. In certain regions ( shown later) you can... The...

This blog is authored by Kevin Barlett (Senior Customer Engineer, Customer Success Unit) and reviewed by Mohamed Kabiruddin (Senior Program Manager, Azure SQL)   In this article, we'll detail the process...

Querying Delta Lake files using T-SQL in Azure Synapse Analytics is now generally available. In this blog post, you will learn how to use serverless SQL pool and T-SQL language to analyze your Delta Lake files from your Azure Synapse workspace.

In this article learn about various options for in...

I’ve architected and currently implementing a solution that uses Synapse (my last newsletter has the details, plus the architecture diagram). Synapse Serverless is the Microsoft answer to Amazon Athena but instead of using open-source tools like Presto, it’s built on SQL Server. In this project we extract many tables from 1,500 on-prem SQL Server databases and stage them in ADLS.

Azure Synapse Analytics is a limitless analytics service that brings together data integration, enterprise data warehousing, and big data analytics. It gives you the freedom to query data on your terms, using either serverless or dedicated resources—at scale. Azure Synapse brings these worlds together with a unified experience to ingest, explore, prepare, manage, and serve data for immediate BI and machine learning needs.

If you don’t get smarter soon, it’s your own damn fault. Oh sure, I’m running a Black Friday sale, but even if you’re as broke as SSMS’s live query...

Azure Charts is a site that a lot of people should know about but don’t. Alexey Polkovnikov came up with and developed the site, and I highly recommend checking...

If you’ve ever set up an Azure Network Gateway for Site to Site or Person to Site VPNing you’ve probably wanted to be able to see logging from the...

The PASS Data Community Summit 2021 is with us online this week, November 8 – 12, with over 300 speakers delivering over 300 sessions to, literally, thousands and thousands...

Attending PASS Data Community SUMMIT 2021

We held an in-person SQLSaturday here in Orlando last weekend (Oct 30th). We didn’t organize one last year, there was just too much risk and too much uncertainty, so... The...

I’d just like to say a quick thank you to Redgate for hosting the Pass Summit, all of the amazing ... Continue reading

Key Details What: SQL Server & Azure SQL Conference at DEVintersection.Where: Las Vegas, NV.When: Sunday, December 5th through Friday, December 10th.Admission is not free. Registration is available from the conference...

With so many recent contacts from possible clients who have been hit by ransomware, I have seen too many companies completely lose their entire SQL Server databases, their backups,...

Join Steve Jones, Kathi Kellenberger and Grant Fritchey as they each reveal their highlights, learnings and key takeaways from their 2021 Summit experience.

I’ve published a new project to the PowerShell G...

Sometimes when you are trying to figure something ...

In this previous post, we introduced the notion of...

Let’s assume that we have some distribution we want to estimate some quantity related to it (e.g. the mean of the distribution). A typical estimation strategy is to draw...

It’s well-known that the distribution has heavie...

In this previous post, we explored how heavy-tailed the distribution is through the question: “What is the probability that the random variable is at least x standard deviations (SDs)...

Happy T-SQL Tuesday! I’m really excited for this...

This month’s topic is something I wouldn’t hav...

Growing companies often find themselves floating on an “ocean” of underutilized or misused data – data that doesn’t reach the people who would most benefit from it or reaches...

 with David McGoveran  (Title inspired by Richard Feynman) In Part 1 we explained that all database relations are, mathematically, relations, but not all relations are database relations, which are in both...

Let’s consider the benefits and drawbacks of SQL views. Querying multiple tables can be easier with views. It’s shorter to write a query against a single view (SELECT *...

Reading Time: 3 minutes In this post I want to cover reasons to learn GitHub if using Azure DevOps. Because I think it is an important topic to...

An ETL process requiring SQL Server View definitio...

This is the second article in the series of Migrat...

Date: Tuesday, December 07, 2021 Time: 12:00 PM Ea...

SQL Server on Azure Virtual Machines is the best option to migrate your SQL Server workloads maintaining complete SQL Server compatibility and operating system level access. It is ideal...

Introduction On November 8th, 2021, AMD held an AM...

It happens often in Power BI calculations and reports that you need to concatenate a list of values from a column. You can do this concatenation in Power Query...

CONTAINSSTRING: Returns TRUE if one text string contains another text string. CONTAINSSTRING is not case-sensitive, but it is accent-sensitive. https://dax.guide/containsstring/ CONTAINSSTRINGEXACT: Returns TRUE if one text string contains another...

AND: Checks whether all arguments are TRUE, and returns TRUE if all arguments are TRUE. https://dax.guide/and/ OR: Returns TRUE if any of the arguments are TRUE, and returns FALSE...

Dos Puntos Look, I really like EXISTS and NOT EXI...

If you are new to using C# and the Tabular Object Model (TOM), please check out the previous blog post (https://dataonwheels.wordpress.com/2021/10/15/power-bi-meets-programmability-tom-xmla-and-c/) for both an introduction to the topic and...

What I did learn when working through the blog post is that I ran into some errors when trying to re-connect or trying to connect to the Azure Storage...

This video describes how to enable the Single Valu...

You should be reviewing performance of your Power BI Reports. This report from Smart Power BI helps you to visualize the results of Performance Analyzer. Let's check it out!...

Power BI's "Show data point as a table" feature is not a security hole - you need to use row-level security and object-level security to stop users from accessing...

Fighting Churn with Data by Carl Gold is an intere...

Store your favorite DAX expressions into repositor...

Azure Synapse Link makes it simple for data teams to unlock greater value in their data by eliminating barriers between Microsoft data stores – including both operational data stored in Azure Cosmos DB and business application data stored in Microsoft Dataverse – and the limitless cloud analytics capabilities available in Azure Synapse.

Two years ago I told you about the then-new G4 ins...

We have released a new early technical preview of the JDBC Driver for SQL Server which contains a few additions and changes.   Precompiled binaries are available on GitHub and also on Maven Central.   Idle...

Today and every Wednesday Data Exposed goes live at 9AM PT on LearnTV. Every 4 weeks, we’ll do a News Update. We’ll include product updates, videos, blogs, etc. as...

Get’em Daddy You know, you love it, you often wondered if it would ever get a new version! Someone JUST asked me yesterday if this has been updated recently. Thank you...

Announcement: Purview policy based access control for Azure SQL at scale – Private Preview   I am excited to present a new way to control access to resources by assigning policies in...

At this past virtual Ignite 2021 conference, Microsoft announced that their Microsoft Defender for Cloud service will provide native support for AWS. Multi-cloud protection has become a priority for...

Action BI Toolkit provides tools to bring Power BI projects under source control, enabling professional development workflows and enhanced governance. Available for free and with open source code (MIT...

A little over two months ago we announced public preview for the Azure SQL Migration extension for Azure Data Studio to ease your database migrations from SQL Server to...

Learn how to create performance charts for time series data using SQL Server and Python.

Long ago at a real-life meetup (remember those?), I received a t-shirt which said: “biggeR than R”. I think it was by microsoft, who develop a special version of R with automatic parallel work. Anyways, I was thinking about bigness (is that a word? it is now!) of your data. Is your data becoming to big?

Brent Ozar opines on an interesting topic: The que...

yes SQL Server 2022 is coming and would have many advance features. https://www.microsoft.com/en-us/sql-server/sql-server-2022 whats new in SQL Server 2022 would be know more here: https://techcommunity.microsoft.com/t5/microsoft-mechanics-blog/what-s-new-in-sql-server-2022/ba-p/2922227 Bob Ward would be... The...

Ransomware has threatened many organizations over the past few years. In this article, Robert Sheldon explains the history of ransomware and what needs to be done to protect against it.

Russian cybercrime group known as T505 is targeting SolarWinds Server-U systems that haven't been patched for a remote code execution vulnerability fixed this summer.

Wiz researchers who discovered a severe flaw in the Azure Cosmos DB database discussed the full extent of the vulnerability at Black Hat Europe.

ArsTechnica’s Sean Gallagher has a two–part article on “securing your digital life.” It’s pretty good.

Electronic Warfare Associates says an attackers in...

The total number of smart attacks (advanced DDoS attacks that are often targeted) increased by 31% when compared to the same period last year.

The Nokia 2021 Threat Intelligence Report announced today shows that banking malware threats are sharply increasing as cyber criminals target the rising popularity of mobile banking on smartphones, with...

In this article, we will discuss the performance d...

When messing with data, sometimes you want to chan...

Itzik Ben-Gan concludes his series on table expressions in SQL Server, explaining more internals of inline table-valued functions. The post Fundamentals of Table Expressions, Part 13 – Inline Table-Valued Functions,...

A customer I’ve been working with for a while now has a monolithic ASP.NET MVC web application which we are porting to .NET Core 3.1 (and then almost immediately...

It Was Written I was a bit surprised by this, because I thought the whole point of these new functions was to avoid errors like this. SELECT oops...

In this article we look at different ways to format currency output in SQL Server and most notably the different options the FORMAT function provides.

See the new overloaded STRING_SPLIT() function with its enable_ordinal parameter, now available in an Azure SQL Database near you.

This article covers various ways to split strings in SQL Server along with using OPENJSON and also a performance comparison of different approaches.

In this article, we will learn how to enable multi-lingual search on data hosted in Azure SQL Databases. Introduction Azure SQL Database is one of the most popular relational...

I show one way to run arbitrary SQL against objects in an arbitrary database - using nested dynamic SQL.

101ers This is still one of the most common problems I see in queries. People are terrified of NULLs. People are afraid to merge on freeways in Los Angeles. What results is...

Every time I write a post about splitting strings, I promise myself it's the last one. I need to stop making that promise. In this tip, I show how...

This article will explore Parameterized SQL notebooks in Azure Data Studio. Introduction SQL Notebook or the Jupyter notebook in the Azure Data Studio has excellent capabilities that include codes...