A Quest for MIRACLE
Published on : Tuesday 10-08-2021
Dipankar Roy offers a perspective on the quest for creating machines that can replicate human behaviour.
For billions of years, the Earth has witnessed numerous species coming into existence and many becoming extinct over time. Some of these had the privilege to be dominant for centuries and ruled the earth. Humans emerged around 2 million years back. Since then, we have tried to explore the universe and master it across several dimensions.
We may not be the best when compared to many other creatures when it comes to our size, strength, speed, stamina, senses, survivability, etc. But, one thing that has helped us compensate for all these weaknesses and gain substantially over all others is – the brain. Over centuries, the human brain has gotten trained by exploring new horizons, experimenting with the environment, and associating the experiment results with the prevailing conditions in the form of learning. We have come a long way in this learning journey mastering many aspects of life. There has been a major focus in engineering, life sciences, medicines, genetic engineering, space research, robotics, cloud computing, etc. in the past decades.
Such has been the magnitude of our success that humans were fascinated with the capabilities of their own neural system and this was regarded as the greatest MIRACLE that must be explored. Of late, the quest of mankind has been to understand the deep complexities of the human brain and simulate it – in an endeavour to establish the ultimate supremacy of creating machines that can replicate human behaviour. This started the race to invest into concepts and technologies that can help achieve the ultimate success.
Until now machines were pieces of great engineering work that could be driven by humans to perform wonders. But how could the machines think and drive themselves? How could they learn from experiences? Remember your childhood while trying to have berries – some sweet and some sour. First time you bit into the sour one, you learnt from the experience, associated the colour and texture of the berries with their taste and stored the knowledge in the brain. The next time you saw an unripe berry you just ignored it.
Yes, there had to be a way that machines could learn. Today Machine Learning is one of the proven technologies where engines use training algorithms to learn and identify hidden patterns from tonnes of data. It is a fact that similar experiences form deeper impressions on humans, and they form stronger opinions. Likewise, machines need to be trained with a real set of data to be able to create meaningful decision rules and store them. When one has varied experiences, it often becomes difficult to act as one cannot arrive at a decision due to diverse learning. Similarly, variations and outliers in the training data can lead to very low confidence on the rules.
While certain success was achieved in making the machines independent to decide on their own, the challenge was to have the right kind of sensors that could work like human sense organs to capture various inputs. In the end, the brain needs these inputs in the form of an image, odour, sound, taste or feel to be able to analyse it, scan through its knowledge base and send the right command to the body to react. The answer lied in the Internet of Things. IoT devices could sense the various input parameters from the environment and feed to the ML Engine for processing.
We talked about certain weaknesses of humans earlier – speed, stamina, strength. Humans have emotions. Over time, they seek change and novelty, get tired of repetitiveness and stop performing. While we tried to simulate the human behaviours in machines, it was important to get over the weaknesses. Robots were designed to overcome this. In fact, there is a lot of research happening in the field of Robotic Process Automation to enable machines to complete the repetitive mass actions without the element of ‘human error’.
The child today is privileged to get packaged sweet berries. Ever thought what goes behind is a machine using Artificial Intelligence including the IoT sensor to sense the colour and texture of berries, ML based rule engine to process the inputs and RPA based levers to segregate the sweet from the sour ones and package them. Who helped us in this a few decades ago? Probably, we asked our moms to identify the sweet ones. Today, we have also succeeded in creating digital assistants that can listen to our queries and pass on the command to such an artificially intelligent machine. Conversations with such digital assistants have become more and more human-like using Natural Language Processors. Conversational AI is slowly becoming an integral part of our lives, making us feel in command. So, with a combination of ‘M’, ‘I’, ‘R’, ‘A’, ‘C’ we can get the sweet berries by just a simple voice command like ‘Give me a pack of sweet berries’. It is all that simple.
Slowly humans are inching towards achieving the miracle they wished for. However, there are still lots of challenges which keep presenting themselves. The recent pandemic gave a big jolt to humanity and our entire infrastructure of intelligence and supremacy collapsed in front of an invisible virus. Millions of lives were lost, and this made us realise that we still are in a learning journey and need to continue our quest and innovations. This reminds us that while machines can start replacing humans in some or the other form, they are still far away. To combat more complex challenges, we need to have the right Leadership behaviour demonstrated by the nations and business organisations. Leaders need to be selfless and show the Empathy for one and all. They need to come together, complement each other, collaborate to develop innovative solutions and think like humans. After all, machines cannot imitate all human behaviour – some traits are best left to humans.
Dipankar Roy is Vice President for Business Application Factory in ERP4SME, at SAP Labs India. He has been in IT for 22 years, of which 20 years at SAP. He has contributed to multiple ERP products in CRM, FIN, Supply Chain and Retail domains. He has worked in Engineering Teams driving multiple platforms in Product and People leadership roles contributing to UI, Analytics, Data Mining and ML, Business Configuration, Office Integration, Process Integration, Identity Management. He is a Post Graduate in Computer Applications from NIT, Raipur.
