Building Smarter and Efficient Systems Using Analytics
Published on : Monday 30-11--0001
Seth DeLand elaborates upon impact of predictive analytics systems on industries like manufacturing and medical devices.
It’s well-known at this point that data analytics technologies can bring significant business benefits in areas such as predictive maintenance. However, system architecture for such applications remains an open question. Customers are hesitant to share their data with vendors, logging all of the data from a machine is often impossible given the volume of data created, and responses to events may be needed in milliseconds – much too short of a time to wait for a response from an Internet server.
All of these will drive innovation at “the edge”, or on the equipment itself. This will entail data reduction techniques such as signal processing algorithms that can transform high-frequency sensor data into a compressed form that can be easily transferred over a network. Design constraints may also result in running a machine learning model directly on the equipment. Software that makes it easy to develop an algorithm once and deploy it into these different scenarios will enable design teams to implement the optimal architecture for their systems.
As AI systems, wearables and other new technologies move from concept to reality, there is an overarching need to aggregate data that is scattered across multiple platforms, apply the latest analytics capabilities, and transform the data into something that’s actionable.
Predictive analytics systems will allow for more informed and personal relationships between patients and physicians and more effective diagnoses at point-of-care. It is quite possible that predictive analytics will also drive the progress of both preventative and therapeutic care with the data collected from wearables and shared on personal devices.
Adoption of Machine Learning and Deep Learning technologies
As it becomes easier and easier to apply machine learning techniques, more products and services will incorporate machine learning models. Embedded systems, typically used for controls and diagnostics, will incorporate machine learning models that can detect previously unobservable phenomena (e.g., detecting a driver’s style of driving, or classifying whether a machine is likely to breakdown or not). In 2018, we’ll continue to see machine learning models being incorporated in new places, especially in edge nodes and embedded processors.
While deep learning continues to look promising, there is still a lot of design and tuning necessary to train a useful deep network. Techniques such as automated hyperparameter tuning appear well-positioned to reduce this work, which should ramp-up the pace of adoption of deep learning.
Competitive advantage of data science technologies
As engineering and IT teams become more integrated, there will be increased demand for domain experts that understand the core products and services of the business. Partnering with (or serving as) data scientists, these domain experts will be critical to identifying areas where data science technologies can benefit the business. Empowering these domain experts to apply data science methodologies will enable the rapid integration of big data and machine learning technologies into the services and operations of the broader organisation, ultimately establishing a significant competitive advantage when offering the products and services that customers are demanding.