We have a belief that we are seeing an inflection point in machine learning opportunities across enterprises. Technologists has been talking about machine learning for decades but four underlying shifts have made it so now is the time these companies will start to take off.
First, the availability of data has skyrocketed. Over the last 10 years enterprises started to save more and more of the data they created, leading to “big data” and consequently a trove of data for machine learning algorithms to learn on. It is estimated that more data will be created in 2017 than all prior years combined and, according to Domo, 90% of all data was created in the last two years.
Second, cloud computing has been widely adopted. The transition to cloud computing has revolutionized the world and allowed for billions of data points to be created. Additionally, users have access to large amounts of computing power for a fraction of the cost. As a result, the machine learning algorithms can be trained quickly. Even as edge computing starts to grow, cloud computing will continue to be the key place where low cost computation occurs.
Third, open source machine learning infrastructure became available. Machine learning has not been left out of the explosion of open source software, which has made machine learning more accessible to a wide variety of users. There are current open source projects that help with a number of aspects of machine learning, from projects that make it faster to train machine learning systems, to others that provide a framework to work within, to some that optimize workflow on GPUs.
Fourth, the advancements in machine learning algorithms cannot be understated. With a renewed focus in the field of machine learning there are a number of different algorithms that have evolved over last 10 years that enable people to better train what they are doing. The three broad categories (supervised learning, unsupervised learning, and reinforcement learning) have all seen significant progression.
These factors have led to a number of machine learning companies. More specifically they have provided the ability for companies to integrate machine learning into the problems they are trying to tackle, whether that be data liberation or optimization of articles on a webpage.
We saw in ClearGraph (formerly Argo) a team that was applying machine learning toward solving a large problem, enterprise search. The company was trying to solve a broad issue in a relatively crowded market, but we felt that the team as a whole stood out due to their technical expertise, clear vision, and ability to effectively execute. They succeeded in not only making data searchable through natural language, but also in solving the complex backend problems of data aggregation and normalization of data from disparate data sources. These backend problems are a key pain point for many enterprises today. The ability to accomplish all these goals, while also remaining focused on the growth of the company, helped propel ClearGraph to a terrific exit to Tableau.
Congrats to the team—we are happy to have been part of such a great company.