Practical machine learning
Over the years I have taken a ton of Coursera courses related to machine learning.[0] It’s amazing how advanced the labs and experiences have become where you can spin up a rela environment in the cloud where you can actually do things with machine learning and curated data sets. That is one of the ways that practical machine learning has thrived over the last 5 years.[1] The John Hopkins University course on practical machine learning has had 141,649 people enrolled at the time this sentence was typed. Learning about machine learning and jumping into practical use cases eventually will represent the smallest use case for practical machine learning. Over time the most prolific use case will be where machine learning is built into the products people are using. The developers over at Google have released a formula suggestion feature into the Google Sheets product.[2] Microsoft developers have been working to build machine learning into the products related to office for years. Some of them extend Microsoft Excel using machine learning.[3] You can take a look at the videos Yannic Kilcher releases or the counts of people taking popular machine learning courses. You will quickly realize that an audience of around 10,000 people shows up to a lot of machine learning content. The number of researchers submitting papers to conferences and the audience of people who deeply care about machine learning is much smaller than they number of machine learning jobs that are open on LinkedIn right now which stands at around 153,481 results for machine learning in the United States.[4]
Outside of direct expertise within the field of machine learning users of products like Google Sheets or Microsoft Excel are seeing augmented experiences thanks to practical machine learning being applied to real world scenarios. At the point where machine learning is being actively distributed to the world of spreadsheets the incidence of applied machine learning rises exponentially by well over an order of magnitude. As of June 2021, the number of companies in the United States using Office 365 is over 700,000.[5] A lot of people use Microsoft Excel for all sorts of things. That really does change the dynamics for applied machine learning in the workplace and for people in general. Use cases that are the result of machine learning projects in companies building it into products are going to have a far greater reach than anything based on some of these major industry driving business tools. Between these two use cases that I’m considering of people doing machine learning and people benefiting from practical machine learning being into products this really is a golden age for machine learning.
Spending some time thinking about practical machine learning has made me really wonder what is next for machine learning within the space. I’m curious about what new and interesting things people will do with applied machine learning. More and more products are going to have it built into their core functions.
Links and thoughts:
Yannic dropped another edition of ML News this week, “[ML News] Microsoft combines Images & Text | Meta makes artificial skin | Russians replicate DALL-E”
This week Linus and Luke really got overly excited about YouTube dislike buttons. It got a little intense compared to their normal banter. “YouTube Made a Huge Mistake - WAN Show November 12, 2021”
Microsoft developers do produce a weekly AI Show and this week it was interesting, “AI Show Live - Episode 39 - Exciting updates from Managed Online Endpoints in Azure ML”
Top 3 Tweets of the week:
Footnotes:
[0] https://www.coursera.org/user/f794f67f2bb7e56d7609fbe2b586653d
[1] https://www.coursera.org/learn/practical-machine-learning
[3] https://venturebeat.com/2020/12/30/you-dont-code-do-machine-learning-straight-from-microsoft-excel/
[4] https://www.linkedin.com/jobs/machine-learning-jobs/
[5] https://www.statista.com/statistics/983321/worldwide-office-365-user-numbers-by-country/
What’s next for The Lindahl Letter?
Week 44: Machine learning salaries
Week 45: Prompt engineering and machine learning
Week 46: Machine learning and deep learning
Week 47: Anomaly detection and machine learning
Week 48: Machine learning applications revisited
I’ll try to keep the what’s next list forward looking with at least five weeks of posts in planning or review. If you enjoyed reading this content, then please take a moment and share it with a friend.