More Deep Learning Skills Needed

As deep learning becomes an increasingly important part of the artificial intelligence and business environments, it will inspire changes that affect other systems and processes as well as business requirements in software development.

Although machine learning has long been considered integral to the analytics toolbox, particularly as a component of predictive analytics, the increasing power, flexibility, and range of deep learning set it apart. [Editor's note: For more about deep learning, see "Styles of Deep Learning: What You Need to Know."] The growth of deep learning has created new possibilities and new ways to apply analytics and artificial intelligence to business problems.

We are in the midst of a skills shortage in big data analytics and data science. The scramble for talent has driven academic programs to scale up; many enterprises are gaining skills by acquiring innovative start-ups. With deep learning, we can expect similar tactics. Indeed, we are already beginning to see this with Microsoft acquiring Maluuba, TomTom acquiring Autonomos, and Uber acquiring Geometic just in the past several months.

The skills required for deep learning are similar to those for analytics with some important differences. Machine learning is not primarily about databases or applying statistics to structured and unstructured data. Instead, it is an approach that attempts to navigate, interrogate, dissect, and derive results from any stream of data using a set of discrete and autonomous algorithms derived from neural networking.

Who Steers the Deep Learning Ship?

Who will be responsible for assembling the proliferating deep learning projects? Data scientists focusing in this area will need to integrate results with the larger analytics framework to achieve desired results. Deep learning specialists, on the other hand, need to focus on deep learning, its advances, and general issues in application of machine learning and neural nets.

Understanding deep learning is about understanding the many deep learning algorithms. It is also critical to understand how deep learning can be used with other AI and analytics components. The field is developing and there are numerous threads of advancement across a wide spectrum of application areas. Practitioners in this area need to be familiar with these algorithms and with the tools that support them; they also need to have an extended understanding of cases in which these algorithms have been applied and the specific problems that arise in each of them.

These problems are not the same as the statistical problems raised by big data; they are problems of matching edges, selecting hierarchical traits, organizing networks of inquiry, and establishing output parameters. The skills needed are heavily related to programming and to the existing body of work in the machine learning area. Popular programming languages include MATLAB, R, and Python; special skills include an understanding of core concepts in pattern recognition, and ability to understand and apply an ever-broadening mix of analytic strategies.

The Coming Skills Shortage

The uniqueness of deep learning will likely mean that enterprises will have difficulty filling positions quickly. This will create another drain on recruitment processes and develop into a global competition as deep learning becomes popular for every organization wishing to expand capabilities in areas specific to robotics and process automation.

The applications for this technology are so critical for competitive advantage that the need for these skills is likely to explode during the next several years. In addition to a need for specialists in the general area of machine learning, there will be an increasing number of subspecialties as deep learning continues to develop new use cases and more powerful functions. It demands both new ways of thinking about data and new ways of modeling a solution.

Deep learning solutions do not exist in a vacuum and developing expertise in this area will entail issues of data access and streaming; sensors and robotics; and, above all, how to integrate autonomous and semi-autonomous deep learning systems with other technologies. Even within the general area of analytics, integrating these algorithms and architectures with other analytics approaches and processes will create demand for new skills.

As a rapidly evolving skill set, the greatest demand will be for experts with practical experience.

Finding the Talent

The rush for deep learning skills is already underway. These skills will be required across an increasing array of processes, and this will result in greater emphasis on deep learning in academic and training programs. As with big data and data science, however, there will likely be an immediate lag as available expertise is drawn from the workplace; Uber hired a whole research department from Carnegie-Mellon. Acquiring these skills early can provide an advantage if there are suitable applications in sight, but special skills are not inexpensive.

Meanwhile, training can fulfill many needs, particularly starting with data scientists who already have special interests in this area. It will require commitment because the field is developing quickly. Hopefully the errors of previous skills shortages can be avoided.

About the Author

Brian J. Dooley is an author, analyst, and journalist with more than 30 years' experience in analyzing and writing about trends in IT. He has written six books, numerous user manuals, hundreds of reports, and more than 1,000 magazine features. You can contact the author at bjdooley.query@yahoo.com.

Subscribe to Upside

Q&A with Jill Dyché

Find out what's keeping teams up at night and get great advice on how to face common problems when it comes to analytic and data programs. From head-scratchers about analytics and data management to organizational issues and culture, we are talking about it all with Q&A with Jill Dyche.

View Recent Article

Submit Your Questions to Jill

Powered by TDWI. Advancing All Things Data
A Division of 1105 Media, Inc.