Google’s DeepMind not only beat a master player in the ancient strategy game of Go, it did so with unexpected results. Performing moves that no human had ever done. Moves that were graceful, levels above human in strategy, and by one account, beautiful.
This same power can be utilized by any organization using Open Source Software, specific machine learning models, and a production ready dataset. Maybe not simple, but more achievable than ever before. Including solutions for such use cases as: sales, customer service, maintenance, supply chains, human resources, and more.
How Business is using Deep Learning
Even small business is using Deep Learning. For example, a hardscape contractor in Tennessee is using Deep Learning to empower their field sales team. Holding a tablet, the salesperson now shows potential customers what their back yard will look like at various levels of investment. It does so by detecting what is concrete, rock, grass, and trees. Then superimposing images of their products onto the structures.
Larger companies are using Deep Learning to improve operational efficiency. Real Estate companies are now able to determine when a property is most likely to become available. Retailers are better at forecasting. Manufacturers save tremendous amounts on energy costs. Sales organization are perfecting their pitch based upon their best producers work.
The point is that Deep Learning is capable of driving business value beyond that of traditional software. In the past, many companies utilized software to automate, organize, and improve core functions. Today, these same organizations can augment human capabilities with Deep Learning. Going beyond improving existing processes to accomplishing new ones.
Creating Deep Learning Solutions
DeepMind uses a technology called Deep Learning. This is a type of Machine Learning that has proven itself for commercial applications. Everything one needs to build Deep Learning applications is available for free. Including a Google developed framework called TensorFlow that provides an end-to-end solution.
The missing piece is data, and lots of it. For large corporations, this is probably not an issue. However, for smaller companies, those with less than 700 employees for example, there is a large assortment of public domain and open source data. Find something meaningful to your organization and all that needs to be done is cleaning, labeling, and formatting it. Still a potentially large undertaking.
Otherwise, a dataset will need to be created. This can be difficult as it usually requires large scale data entry. People entering data, labeling that data, formatting it for a specific model, and cleaning it. However, this is no longer issue due to companies like White Rabbit AI that provide datasets on demand through automation — reducing the cost by 90%.
Once the data and model are created, one has the choice of hosting it on Google Cloud, Amazon Web Service, or most any cloud provider. Serving the model to web or mobile applications as needed. In short, most Python Developers and Data Scientist are capable of doing this.
While it is not magic. Deep Learning is perhaps the most powerful technology available today. Limited only by the quality and quantity of data available, it is changing how many organizations operate. Allowing companies to solve problems at a scale previously considered impossible.
In the interest of full disclosure, I am part of White Rabbit AI. A company that provides datasets on demand. You can learn more and contact me at http://www.whiterabbitai.com.