Industries consistently demand high efficient and cost-effective solutions without compromising human and machine safety. Artificial Intelligence enabled by emerging digital technologies plays a vital role in addressing these business needs. Most AI solutions today are “centralized” in nature i.e. they require massive datasets, expensive computing resources, periodic tuning and optimization of complex AI models. Centralized models in the long run can gradually lead to the monopolization of AI marketspace, eventually confining the participation of other organizations in AI innovation.
Nevertheless, the advent of technologies like mobile and edge computing, on-device analytics, has a huge potential to enable faster decision-making through direct machine to machine (M2M) communication in a proficient way without the need for a centralized hub. The decentralized AI, when rightly exercised, helps in democratization of AI marketspace.
The notion of decentralized AI is extended to collaborative AI illustrating how a very “high availability” feature can be achieved with democratized intelligence, concisely called as the HAWDI platform, representing High Availability with Democratized Intelligence. This proprietary IP platform features functionalities like fault tolerance and smart-scale up. The platform leverages machine learning and machine vision technologies at Edge. In this blog series, we will highlight potential of the platform and subsequently discuss technical details with real-time applications.
Opportunity landscape for HAWDI platform
HAWDI platform opens-up the possibilities of tapping $550bn [Gartner, April 2018] of forecasted global business value for smart and intelligent products (edge devices) upcoming in next 5 years.
As per Gartner, smart machines will enter mainstream adoption by 2021 and enterprises will look for service providers to help them deploy AI technologies. The growing number of edge intelligence and IoT devices open-up potential applications across multiple industries.
Material movement is a frequent chore in manufacturing industries. Industry cranes used in this context are usually manual-controlled and demands lot of coordination on shop floor. A complete situational aware, intelligent and collaborative AI solution can avoid accidents while mobilization on shop floor. This will not only improve productivity but also enhance human and machine safety.
Autonomous haul trucks in mining industry are self-driving but remotely monitored by operators from a central control room. These operators closely monitor various activities and communicate constantly with personnel in the pit. There is a good scope of improving productivity by increasing the autonomy in operations. A direct collaboration between haul trucks, shovels and other mining vehicles can be established.
Automotive and Transportation
When a truck carrying goods is broken down on its way, it impacts the delivery schedule. By employing collaborative intelligence model, another truck from vicinity can be deployed to transport goods to the delivery location. High availability is an important factor for fleet management and transportation. This solution can also be implemented in other applications such as connected cabs, courier services etc.
Use of AI in healthcare is growing from managing medical records of the patients to assisting doctors with medication during surgeries. One such example is the area of nursing which involves continuous monitoring of patient, consultation with doctors in critical situation, etc. To avoid human errors in such scenarios, an AI agent can seamlessly monitor the health parameters, assess the criticality and coordinate with doctors for necessary action or intervention.
We have identified few challenges in certain sectors where the HAWDI platform can be well leveraged to improve efficiency. Needless to mention there are further potential opportunities. The use of decentralized architecture in AI applications is going to become more prevalent and inevitable in future.
In my upcoming blogs, we will detail some technicalities of the platform and with an example to illustrate the concept of collaborative intelligence with HAWDI platform. We hope this will help readers in appreciating the potential of the platform for achieving larger goals and embark upon AI-as-a-service model.