Large quantities of data are generated when humans interact with software systems (like you purchase history from an online store) or by logging the operations of an electromechanical system (like the history of operations from a plant). These are qualified as Big Data due to its size and the rate at which it gets generated. Our curiosity to process this data resulted in algorithms that can predict what could happen next (like what would you like to buy next) or diagnose (like when a mechanical part in a plant will fail). The success from such predictive systems attracted the attention of engineers in designing innovative products and services that we all use today. The effectiveness of the data in solving many problems is thus deeply proven.
We are now in a world of Artificial Intelligence (AI) where data enables the discovery, implementation and even refinement of solutions over time. In recent years, we have been talking about data that is much more complex to process than before. Examples could include what we say, hear, read or experience. Many of these are not even captured at this stage. These experiences characterise our personalities, intents and our preferences. However, computers fail to understand them.
Perception has been a highly important but less understood cognitive skill. The world has been eagerly looking for the development of algorithms that can mimic human capabilities in this space. We have made huge leaps in this area in the last five years, thanks to deep learning. This success can be credited to a number of factors including the large data sets, powerful computing infrastructures, and of course advances in efficient and effective algorithms for machine learning. Often these machine learning algorithms look at a number of examples and experiences and learn from them to generalise the performance into novel future situations.
What does it mean for us? Systems of the next decade will have some amount of perception and problem-solving skill embedded. You will find them to be more human than machine. We have already started to see such solutions in the form of Siri and Alexa. Such systems will become more personal, effective and omnipresent in the coming years.
This paradigm shift demands a change in mindset among students, practitioners and planners. Mathematical foundations, algorithmic thinking and experimental skills will become critical. Problem-solving skills will become more important than programming skills. Equally important will be the appreciation of the human mind. These demands are breaking the cultural barriers between software systems and human beings.
We should look at this more as an opportunity to design and implement a new class of products and services. Since these will also have to interact with humans more closely than the previous generation of solutions, a number of novel challenges will emerge from different parts of the world. Technological choices will have to be based on cultural, environmental and lifestyle choices of the people in a region. We can easily see how different parts and sects of India will throw up novel opportunities and challenges. Let’s get ready!
(C V Jawahar is an Amazon Chair Professor at IIIT Hyderabad, India. He received his PhD from IIT Kharagpur and has been with IIIT Hyderabad since 2000. At IIIT Hyderabad, Jawahar leads a research group that focuses on computer vision, machine learning and multimedia systems. His primary interest is in problems that overlap with vision, language, text and retrieval. He is a Fellow of IAPR. In the past, he has served as a chair for many national and international conferences. Presently, he is an area editor of CVIU and an associate editor of IEEE PAMI. Jawahar is especially interested in teaching and research that suits and is needed in the Indian setting. He looks at applied research problems with the same passion as that of basic research. In the recent years, he has been looking at Artificial Intelligence as an area that cuts across different conventional disciplines.)
