Over a million developers have joined DZone.

Lots of Talk About Machine Learning Marketplaces

DZone's Guide to

Lots of Talk About Machine Learning Marketplaces

I feel like the question of, ''Where can I publish my Machine Learning models and begin selling them?'' is going to be a common thing of the 2017 ML hype.

· Big Data Zone ·
Free Resource

Hortonworks Sandbox for HDP and HDF is your chance to get started on learning, developing, testing and trying out new features. Each download comes preconfigured with interactive tutorials, sample data and developments from the Apache community.

I spent last week in San Francisco listening to Google's very Machine Learning-focused view of the future. In addition to their Google Next conference, I spent Tuesday at the Google Community Summit, getting an analyst look at what they are up to. Machine Learning (ML) was definitely playing a significant role in their strategy, and I heard a lot talk of machine learning marketplaces.

Beyond their own ML offerings like video intelligence and vision APIs, Google also provides you with an engine for publishing your own ML models. They also have a Machine Learning-advanced solution lab, throwing a Machine Learning hackathon, and pushing a Machine Learning certification program as part of their cloud and data offerings. As the Google machine learning roadmap was being discussed throughout the day, the question of, "Where can I publish my ML models, and begin selling them?" came up regularly — something I feel like is going to be a common theme of the 2017 ML hype.

I'm guessing we will see a relationship between the Google ML engine, Google Cloud Endpoints emerge, and eventually some sort of ML marketplace like we have with Algorithmia. We are already seeing this shift in the AWS landscape between their Lambda, ML, API Gateway, and AWS Marketplace offerings. You see hints of the future in the AWS serverless API portal that I wrote about previously.

The technology, business, and politics of providing retail and wholesale access to algorithms and Machine Learning models in this way fascinates me, but as with every other successful area of the API economy, about 90% of this will be sh*t and 10% will be actually doing anything interesting with compute and APIs.

I'm doing all my image and video texture transfer Machine Learning model training using AWS and Algorithmia. I then use Algorithmia to get access to the models I've trained, and if I ever want to open up partner level (wholesale) or public (retail) access to my ML models, I will use Algorithmia or an API facade on top of their API to open up access and make available in the Algorithmia ML marketplace. I'm guessing that at some point, I will want to syndicate my models into other marketplace environments with giants like Google and AWS but also other more niche, specialty ML marketplaces where I can reach exactly the audience I want.

Hortonworks Community Connection (HCC) is an online collaboration destination for developers, DevOps, customers and partners to get answers to questions, collaborate on technical articles and share code examples from GitHub.  Join the discussion.

big data ,machine learning ,google

Published at DZone with permission of

Opinions expressed by DZone contributors are their own.

{{ parent.title || parent.header.title}}

{{ parent.tldr }}

{{ parent.urlSource.name }}