Adapting to a Dynamic World through Responsive Intelligence

Going forward, software will need to be increasingly and inherently adaptive to cater to the intelligent systems of the future. With a number of applications across industries, Adaptive Systems are poised to revolutionize the business and IT landscape in the near future.

‘Intelligence is the ability to adapt to change’ – Stephen Hawking

In today’s world, change is not only constant but constantly accelerating. There is a dire need for advanced intelligent systems and processes that can adapt not only to the growing dynamism in the world but also to the ever increasing expectations of technology consumers. Systems of the future will need to know, understand and respond in real-time to real-world conditions. We strongly believe that software will need to be inherently adaptive to cater to these new and modern requirements.

Adaptive Systems are multi-agent systems with advanced algorithms that optimally sense, reason and respond to changes in their environment in such a way that performance improves through continuous interaction with the surroundings.

The ability to adapt to dynamic conditions would allow a complete rethink and redesign of the physical-digital realm across sectors including Retail, Healthcare, Education, Urban Management, Surveillance and many others. For example, in a retail store, adaptive capabilities would assist the retailer to enhance both customer engagement and operations. It would help to identify customers, track them throughout their shopping journey, monitor their behavior, provide relevant customized offers based on their buying intent & purchase history and offer contextualized assistance. This would enable the retail store to provide hyper-personalized experiences, wherein the experiences for a 13-year-old in the store would be different from those for a 50-year-old, without any changes to the IT systems. These intelligent systems would encompass back-end processes as well, ranging from fulfillment, warehouse management and other processes across the supply chain to the point where the manufacturers could dynamically alter production based on real-world cues.

How is it different from cognitive systems?

Cognitive computing involves self-learning systems that use data mining, pattern recognition and natural language processing to mimic the way the human brain works. A cognitive system would leverage cognitive computing to enable machines and humans to interact more naturally. For example, the Infosys NIA Chatbot leverages machine learning, natural language processing and an intelligent conversation engine to provide contextual, natural engagement with users.

An adaptive system on the other hand, would consist of multiple agents that communicate with each other (be it autonomous agents, smart devices, sensors etc.) and leverage numerous technologies (such as Edge computing, Cognition, Sensor fusion, Cloud, Robotics etc.) to enable it to learn and respond to a dynamic environment. The retail scenario that was explained earlier would consist of various cognitive systems, autonomous agents, smart devices and sensors that work together and communicate with each other to provide a seamless personalized experience for each customer. This system on the whole is intelligent and adaptive to the needs of the customer.

Making systems adaptive

Adaptive systems leverage sensor fusion, advanced multi-agent algorithms and event based architectures to learn, reason and respond in real-time to dynamic real-world conditions. As these architectures extend into the physical realm, Edge Computing infrastructure, Sensor Fusion capabilities and Robotic Systems will become more prevalent. These systems, while adapting to changes in their environment, will have to be compliant with the high-level policies of the organization.

Characteristics of Adaptive

To illustrate further, some of the policies or rules would seek to set boundaries and limits for the operation of such systems, such as prohibiting the identification of customers using facial recognition or other means of tracking or using collected data without due consent, limiting a robotic agent to certain physical boundaries, stipulating that a drone should neither fly above 400 feet nor be operated out of the operator’s sight, mandating that when an unforeseen event affects the operation of robotic agents, they should gracefully disengage and go back to their base stations, and so on.

Adaptive systems promise new business value

We can relate to many instances where systems with some level and form of adaptive capabilities are being used across industries. They range from applications in drones, robots and vehicles to surveillance systems and conversational interfaces, to name a few.

Autonomous drones today have the capability to continuously adapt to changes in wind, visibility, and air or ground situations, and automatically make changes in flight plans. These are being deployed to find anomalies during inspection routines in mining, oil and gas, and other such industries. As per industry news, they not only perform the surveys up to 20 times faster but also save around 70-90% of costs. IdeaForge, Skydio, and Boeing are developing solutions around autonomous drones.

Intelligent robots also have adaptive capabilities through which they can recognize human emotions, adapt to their behavior and respond accordingly to their interlocutors. They are being widely used in the retail, banking, health & wellness, and hospitality sectors. Pepper, a robot from SoftBank, is a prime example. Self-driving vehicles also showcase adaptive capabilities which are enabled by context-aware systems, allowing them to sense and respond dynamically to their surroundings. They are employed in the supply chain, warehouse, logistics, retail and automotive sectors.

Another area where adaptive capabilities are applied is intelligent surveillance systems, such as those developed by NEC. They use multiple data sources such as cameras, sensors and records of public authorities to identify threats and ensure safety. Conversational interfaces also have an adaptive element to them, in that they sense intent, context, and emotion during a conversation. These are being used in customer service and according to a leading IT services company are shown to reduce up to 30% of the costs by eliminating the need for human agents to address repetitive customer concerns, and allocating them for other requirements.

Apart from these, adaptive systems are also being applied in other areas such as traffic management (Alibaba), education (Knewton, Fulcrum) and retail traffic management (Percolata). The promise of adaptive systems has triggered a keen interest in academia as well, with the likes of Cornell, UCLA, Boston University and SNN offering courses on adaptive and advanced intelligent systems.

Infosys’s play in this space

Infosys has developed substantial expertise and experience in many of the building blocks that proffer adaptive capabilities to IT systems. The Infosys NIA platform converges Big Data/Analytics, Machine Learning, Knowledge Management, Cognitive Automation capabilities, Optical Character Recognition (OCR) and Natural Language Processing (NLP) to offer a unified, flexible and modular cognitive computing platform. Our Cyber Security services leverage AI, Security Analytics and Machine Learning to combat cyberattacks against digital systems. The IoT and Data management platforms offer solutions that can help collect, manage, monitor, and analyze large amounts of data that is generated by the connected infrastructure and agents. Another critical need is the assurance processes for such systems, which are required to mitigate risks and ensure compliance with regulatory statutes as well as organizational policies. In the past, Infosys has built many infrastructure monitoring solutions and command centers for specific purposes that are coming together to manage the complex Adaptive Systems.

With an increasing number of applications across industries, and interest from startups, academia and incumbents alike, Adaptive Systems are gaining traction and are poised to revolutionize the business and IT landscape in the near future. This not only opens up numerous business opportunities but also plays a key role in improving operational efficiency in enterprises, leading to lower costs, better services and growth through innovation. It is clearly a space that calls for attention. Investment in time and resources will surely help enterprises stay ahead of the pack, and reap rewards.