As businesses look to cash in on the hype of predictive analytics, artificial intelligence is becoming one of the main drivers behind those efforts.
To better understand how organizations are embracing predictive analytics and for what goals, Information Management spoke with John Crupi, vice president of IoT analytics at Greenwave Systems. Crupi has over 25 years of experience implementing advanced visual analytics solutions. He oversees the development of edge-based visual analytics and real-time pattern discovery for Greenwave’s AXON Platform, which analyzes and manages critical data immediately from IoT-enabled devices at the edge.
Information Management: First, please tell us a bit about Greenwave Systems and how you are working with customers.
John Crupi: Greenwave Systems is a global software leader for managed services. The company is dedicated to empowering market-leading brands to profitably deploy their own managed services and products, foster deeper customer relationships and grow their businesses. Greenwave’s AXON Predict is an edge analytics platform focusing on solving Industrial IoT/predictive maintenance problems.
IM: Artificial intelligence is one of the biggest buzz words in IT right now. What is the reality of who is investing in AI, how are they using it, and for what goals?
Crupi: You are correct, AI is a big buzzword and it’s in the early adoption phase. However, there are many areas where AI has been applied and many areas where companies are trying to learn how it can benefit them.
It's easy to think that AI just suddenly arose, but in fact much of the “new” AI success is based on the confluence of decades of AI research; a tremendous amount of training data from the cloud guys and highly tuned GPU hardware and software.
AI is being used in many ways, but ultimately it’s trying to make software and things intelligent:
- Intelligent in recognizing and understanding speech to use voice as the first-class mode of communication with devices.
- Intelligent in identifying and classifying images for autonomous car object detection, recognition and crash avoidance.
- Intelligent for doctors to aid in identifying abnormalities and tumors in radiology images by recognizing image patterns too subtle and detailed for even the most trained professionals.
IM: How is AI serving as a driving force behind predictive analytics implementations?
Crupi: AI is beginning to make its way into predictive analytics, specifically in Industrial IoT. Traditional machine learning techniques used in the past turned out to be very data scientist labor intensive. However, newer deep learning AI algorithms and techniques such as a special form of recurrent neural networks (RNNs), called long short term memory networks (LSTMs) (which have been around since 1997), are showing great promise in their temporal prediction capability.
For example, massive amounts of industrial machine’s sensor data can be fed into an LSTM for training. The resulting model can then be inferenced in real-time at the edge to predict upcoming issues and anomalies. The beauty of this AI approach for predictive analytics is that AI models can continuously learn as more data arrives and the models refined to achieve increasing accurate results. Predicting abnormalities in minutes or hours in advance can save companies millions of dollars by avoiding catastrophic failures and downtime.
IM: As investments in IoT devices and technologies increase, how does that impact AI in the enterprise, and vice versa?
Crupi: Traditional means of IoT analytics in the cloud will change dramatically by introducing AI at the edge, essentially pushing “intelligence” to the device. This will allow devices to detect and respond to conditions not only based on their past performance, but also based on the current “fleet” performance and environmental changes without having to rely on the Cloud for decisions. Adding AI at the edge gives the devices even more capability to add predictive capabilities into their intelligence.
IM: How will all of these trends impact jobs in the IT and data management areas – what jobs will be created, lost, or changed?
Crupi: AI and IoT edge computing require highly specialized skills and architectures. AI will begin to affect all IT professionals in the next five years. IoT enterprise architects will have to advance their skills in highly distributed edge and cloud architectures which can scale to millions of devices with zero down-time. AI and Intelligence will become synonymous and will represent a new way of designing and implementing all styles of software, including software living in robots, drones and devices.
IM: What is your best advice to an organization that wants to capitalize on the use of AI to bolster their predictive analytics efforts?
Crupi: Create a small group of data scientists, architects and developers and begin with a few highly defined use cases that provide an integrated AI solution. This would be a solution that builds AI or analytical applications which incorporate AI into the mix to solve a problem.
Use open-source tools such as TensorFlow, Keras and MXNet and take advantage of Python, as well as all the open source libraries dedicated to advancing AI. Evangelize throughout the organization and provide ways in which others in the organization can provide input and get involved.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access