Earlier this year I made a prediction that Machine Learning (ML) would make it to the masses. How did I do?
Google Trends shows us that the term has been gaining steam for the past few years, reaching an all-time high in the Fall of 2018.
And while ML continues to be the most ‘invested’ of the Artificial Intelligence technologies, such as Natural Language Processing or Computer Vision, it still remains to be seen if it is a real game-changing technology.
So, how is ML currently being used? Netflix uses it to help generate recommendations. Yelp uses ML to determine if a photo that someone uploads of a restaurant was taken on the inside or the outside. Twitter uses it to curate your timeline.
Basically, ML is used in a variety of ways, which is great in terms of building solutions. From a development perspective, Clemens Mewald, formerly of Google, gives a great overview of the ML tasks you can do, such as:
- Recommender/ranking systems
- Event/action prediction, which refers to ML models that try to predict the likelihood of an event or user action
- Classification of images or arbitrary objects
- Generative models, which are ML models that can generate output in a similar form to the input with which they were trained
- Clustering, a common form of unsupervised ML in which objects that are similar are “clustered” together
Clemens also goes on to show how Google forms is using ML to predict the type of response needed for a particular question, as seen in this GIF.
This is a great example of how ML is helping us today. As a simple suggestion engine, its unobtrusive, helpful, but if wrong, can be changed easily. Would I consider this a mass market application of ML? Yes. Would I consider this groundbreaking? Probably not. Given all the hype, I want something remarkable!
Alas, as a product person I know that the best way to introduce new technologies is to use techniques like the MAYA principle from Raymond Loewy, which means, “Most Advanced Yet Acceptable.” This principle, when used appropriately, can help people adapt to changes incrementally In this case, I think Google’s use of predicting the most applicable answer to a question is a great way to introduce a prediction without scaring anyone.
At TrackVia, we are also using ML to help customers make predictions. The work we are doing is still early because unlike some of the other ML implementations mentioned above, we are putting the power of ML into our customers hands. We are creating ways for customers to use their own data to create ML models which can then be used for predictions. This is a groundbreaking use of ML, and will be a gamechanger for our customers.
Our vision is to make ML more approachable, to allow people to create and test models without having to write code or think about the mechanics of getting them trained or executed. We want to expose the value of this revolutionary technology and put it in the hands of the people who can directly benefit from it. This is what it means to bring ML to the masses.
So is ML a groundbreaking technology yet? No, but it’s getting there, and we’re excited about the journey.
We’d like to hear from you. How would you apply ML to your business and applications?
Learn more about building IT-approved, custom workflow applications in just weeks by reading Business Applications for Your Operations at the Speed of Now.