Community Spring Cleaning week is here! Join your fellow Maveryx in digging through your old posts and marking comments on them as solved. Learn more here!

Data Science

Machine learning & data science for beginners and experts alike.
NeilR
Alteryx Alumni (Retired)

@SydneyF gave us the lowdown on neural network fundamentals. The latest feature from Quanta Magazine highlights recent research that attempts to relate a neural network's form to its function: Foundations Built for a General Theory of Neural Networks.

Neil Ryan
Sr Program Manager, Community Content

Neil Ryan (he/him) is the Sr Manager, Community Content, responsible for the content in the Alteryx Community. He held previous roles at Alteryx including Advanced Analytics Product Manager and Content Engineer, and had prior gigs doing fraud detection analytics consulting and creating actuarial pricing models. Neil's industry experience and technical skills are wide ranging and well suited to drive compelling content tailored for Community members to rank up in their careers.

Neil Ryan (he/him) is the Sr Manager, Community Content, responsible for the content in the Alteryx Community. He held previous roles at Alteryx including Advanced Analytics Product Manager and Content Engineer, and had prior gigs doing fraud detection analytics consulting and creating actuarial pricing models. Neil's industry experience and technical skills are wide ranging and well suited to drive compelling content tailored for Community members to rank up in their careers.

Comments
SydneyF
Alteryx Alumni (Retired)

Neural networks are being used in an increasing number of real-world applications, but what makes them work is still not well understood. Before we hand over our steering wheels and medical diagnoses to neural networks, there are researchers trying to move away from trial and error development, and work towards understanding them on a theoretical level. This article highlights some of the recent developments in their work.

MarshallG
8 - Asteroid

I appreciate the context literature but it's starting to seem like Alteryx is dancing around the fact that this article (and most popular press about ML / AI) are referring to deep learning models with multiple hidden layers, while Alteryx's neural network tool is based on the single hidden-layer perceptron in use since the 80's.

 

It is probably true that deep learning has gotten overhyped and most customers aren't that advanced yet (certainly true for our organization) and I can also imagine that it might be quite difficult to build a resource-efficient, flexible RNN, CNN, or other deep learning tool, but I also think that Alteryx should be clear that what they currently offer is not the same thing as what's being described in most current literature about 'neural networks'.

 

I'm guessing (and hoping) that Alteryx is working on building this but there's also a Community Idea to support that work: https://community.alteryx.com/t5/Alteryx-Designer-Ideas/Neural-Network-Predictive-Tool-Deep-Learning...

NeilR
Alteryx Alumni (Retired)

Hi @MarshallG,

 

Posts like this and others mean to inform and entertain our readers and don’t necessarily speak to specific product functionality. It is our goal to help our audience understand the broader spectrum of data science and so we want to make sure that we’re taking part in conversations around these types of topics.

 

When we do refer to our Alteryx Designer Neural Network tool in documentation and blogs, its capabilities and limitations are made clear. Via the Alteryx Designer R Tool and Python Tool, any number of deep learning open source libraries can be used to train models. We also have working examples of deploying deep learning models to Alteryx Promote that we’d like to share with our readers on this blog in the future.

 

Thanks for your comment and for making sure we continue to remain honest. Would love to have a discerning voice like yours contribute to this blog in the future, if you’re interested!