Making Artificial Intelligence Applicable
Over the last couple of years, I’ve shown off numerous machine learning/ai models to friends, family, and business executives. From identifying objects in a video to bots that you can converse with, it nearly always elicits a positive response.
It is then typically followed up with some version of, “who can use something like that?”.
For all of the promise that machine learning and artificial intelligence provide, the majority of us see these tools as something that will be useful years down the road. The problem is, the largest organizations in the world are deploying these tools right now and widening their ever growing lead on their smaller competitors.
A study published this week by the Boston Consulting Group and MIT Sloan Management Review revealed that “Only about one in five companies has incorporated AI into some offerings or processes. Only one in 20 has extensively incorporated AI into offerings or processes. The research also reveals large gaps between leaders — companies that understand and have adopted AI — and laggards in every industry. One sizable difference is their understanding of, and approach to, AI data and training.”
For a technology that some argue will be the source of power for the future’s world powers, it is imperative that this set of tools (machine learning, deep learning, artificial intelligence, etc.) becomes more accessible for small businesses, consumers, and local government.
Get Matt Moody’s stories in your inbox
Join Medium for free to get updates from this writer.
While there has been tremendous growth in educational materials for programmers and engineers taking up machine learning/AI (Fast.ai, DeepLearning.ai, MachineLearningMastery.com, etc.), it would be a stretch to think these will be useful for non-technical executives and small businesses.
So, over the course of the next few weeks, I’ll be releasing a series of brief, practical and most importantly, simple articles explaining how a specific machine learning/AI model works and what you need to utilize it right now. Here’s a short list of topics currently in the works:
- Predicting which customers are likely to cancel/unlikely to return (Churn Prediction) — September 13th.
- Identifying vehicles and humans in security video (Object Detection & Classification).
- Early detection of machine breakdown (Anomaly/Fault detection).
- Automatically identify injured livestock (Video analytics).
- Predicting how many customers will walk through the door in three hours (Time-series forecasting).
If you have questions or requests for explanations on a particular model, leave a comment or email matt@bellwethr.com.
(For those who are looking for more technical guides, it’s not the priority right now but I plan to release sample code and datasets at a later point).
