Over a million developers have joined DZone.
{{announcement.body}}
{{announcement.title}}

Machine Learning in 2017

DZone's Guide to

Machine Learning in 2017

The VP of Engineering and the Director of Product Marketing at Splunk share their thoughts on the state of machine learning in 2017.

· AI Zone ·
Free Resource

Enable your enterprise to add AI to your existing infrastructure with EdgeVerve’s Business Applications built on AI platform Infosys Nia™. Register for our webinar to learn more.

Toufic Boubez, VP Engineering at Splunk, sees the following.

Machine Learning-Washing

Expect the market to be flooded with solutions that promise machine learning capabilities and grab headlines, but deliver no substance. There will undeniably be valuable machine learning solutions in the market, but it will become more difficult to separate the signal from the noise. It’s going to be a confusing year for companies attempting to parse through this noise and deploy real machine learning capabilities that help them achieve business goals.

The App-ification of Machine Learning

Although there will be a lot of overpromising and overuse of the term “machine learning,” there will also be some real value gained from machine learning under the hood of advanced applications. Machine learning will become more accessible to a broader user base and applied more broadly to standard IT and business activities through both turnkey and invisible application integration. Machine learning capabilities will start infiltrating enterprise applications, and advanced applications will provide suggestions — if not answers — and provide intelligent workflows based on data and real-time user feedback. This will allow business experts to benefit from customized machine learning without having to be machine learning experts.

Predictive Maintenance

Predictive analytics is beginning to evolve into preventive analytics in the enterprise, but for industrial environments predictive analytics is king. In 2017, industries will leverage machine learning to execute predictive maintenance. As automation is used to quickly and efficiently ensure business continuity, enterprises will turn to machine learning to up the ante.

The biggest industry opportunity for this will be across industrial environments. In a sector where money, efficiency, and safety is reliant on machines, the industrial industry will gain deeper performance insights from machine learning. By predicting maintenance cycles, enterprises can reduce machinery failure and error, and also save costs by limiting costly maintenance on less frequently used machines.

Deep Learning Creeps Into the Enterprise

While machine learning and AI are still relatively new in the enterprise, we can expect deep learning to start making an appearance. With deep learning, more layers of processing elements are added for the ability to aggregate not only textual data but also more complex data like voice and sensor data into meaningful patterns. Enterprises will start experimenting with deep learning for a greater range of activities. Larger sample sizes for data will continue to increase accuracy for recommendations and predictions, especially when it comes to potential infrastructure outages.

In the longer run, advanced learning capacities allow all data to become part of a neural network fabric, expanding beyond internal data and encompassing higher-level data from outside sources. Much like the mandated cybersecurity information sharing acts, such as the Cybersecurity Information Sharing Act, deep learning will help set the stage for organizations to create and use neural networks.

Joanna Schloss, Director of Product Marketing, sees the following.

Hadoop Distribution Vendors Will Have Crossed the Chasm

Unstructured data in Hadoop is a reality. But since the open source problem has not been addressed, they aren’t making much money. As such, there will be an acquisition of many of these vendors by bigger players. As well as the idea that bigger ISV Hadoop vendors will band together and create larger entities in hopes of capitalizing on the economy of scale.

Data Preparation Will Be a Feature

Data preparation will become more of a feature rather than a market as big data analytics continue to evolve both in product offerings and market share. As such, there may be a consolidation in the marketplace as companies start to acquire product offerings in this area as well as customer lists from small, niche vendors.

People Will Realize the True Potential of AI, Machine Learning, and Advanced Analytics

Artificial intelligence, machine learning, and advanced analytics will become more complex as people start to realize the true potential of these disciplines. All three areas require an excellent understanding of big data and big data analytics and they can eventually evolve into a master discipline of analytics — or maybe we'll coin a new term for it in the near future.

Deep Learning Will Mature

By the end of 2017, the idea of deep learning will have matured and true use cases will emerge. For example, Google uses it to look at faces and then determine if the face is happy, sad, etc. There are also existing use cases in which the police is using it to compare the “baseline” facial structure to "real-time" facial expressions to determine intoxication, duress or other potentially adverse activities.

Adopting a digital strategy is just the beginning. For enterprise-wide digital transformation to take effect, you need an infrastructure that’s #BuiltOnAI. Register for our webinar to learn more.

Topics:
machine learning ,ai ,deep learning ,predictive analytics ,advanced analytics

Opinions expressed by DZone contributors are their own.

{{ parent.title || parent.header.title}}

{{ parent.tldr }}

{{ parent.urlSource.name }}