Can machines think?” asked Alan Turing, known as the father of artificial intelligence (AI), in a seminal paper on the topic of computing machinery and intelligence in 1950. Turing did not coin the term ‘Artificial Intelligence’ but his work laid the foundations for a new research area to be termed ‘Artificial Intelligence’ by John McCarthy, one of the organizers of the 1956 conference held at Dartmouth College, UK to delve into the fundamental task of developing an electronic brain. However, by 1973, disappointed by the progress of work, funding dried up in the UK and USA and AI plunged into a long ‘winter’.
In the 20th century, AI was an idea for the future. It needed much more computing power and a greater variety of digital data sources than was available at that time. Today, the picture has changed. Computing power has reached petaflops levels, distributed on the Cloud and accessible to personal devices like smartphones. As much as 90% of the existing data has been created in the last two years; 2.5 quintillion of data is generated per day from sensors, mobiles, online transactions and social media. The challenge is how to harness this huge data flow to return actionable information without storing this data for future analysis because the growth of storage capacity has long been surpassed by the growth of data volume and there is no possibility of the gap being covered. This is why AI has again gained prominence. Big Data Analytics enables the analysis of data as it streams and stores only the intelligence gathered for future reference. Big Data Analytics is one of the applications of Artificial Intelligence.
Atanu Sinha, Director – India and SAARC, Hexagon Geospatial, says: “Big Data Analytics is more to do with past analysis and future trends based on which an organization can make informed decisions. Artificial intelligence automates the processes to create scalable insights from large amounts of data.”
Deep learning is a methodology used in artificial intelligence where computers can be trained to take all the unstructured Big Data and make sense of it using various methods like Artificial Neural Networks, which partially mimics the way the human brain works
Adds Sundara Ramalingam Nagalingam, Head – Deep Learning Practice, NVIDIA, India, “Big Data refers to massive datasets brought on by new technologies like the Internet of Things, social media data and genomics. These datasets are so large and so complex that traditional methodologies cannot be used to analyze them.” Deep learning is a methodology used in artificial intelligence where computers can be trained to take all the unstructured Big Data and make sense of it using various methods like Artificial Neural Networks, which partially mimics the way the human brain works. Deep learning uses algorithms to look for complex relationships in all that Big Data and then we further refi ne those algorithms as they go along to make them better.
There are two types of AI. The first is a top-down approach which concentrates on building intelligent machines that can replicate instances of intelligent behavior. The data and programs are supplied by humans. These are exemplified by Expert Systems which aid humans in making decisions and by machines like Big Blue which can play and beat Grandmasters at chess. The second is the bottom-up approach that ‘learns’ from the data presented to it almost mimicking the process of human intelligent behavior. The keyword is ‘almost’. Cognitive computing, machine learning, deep learning, genetic and evolutionary computation, neural networks, natural language processing, knowledge discovery are some of the branches of AI falling in this category. Today machine learning is at the cutting edge of AI and deep learning is the cutting edge of machine learning! However, mimicking the human brain will require a neuromorphic computing device the size of a Volkswagen Beetle and will consume about 20W of power. Hence the stress on ‘almost’!
Self-driving trucks are some of the transportation solutions using AI. Integrated with IoT, such solutions are poised to bring a paradigm shift in the world of transportation. Disaster management and military ops also need real-time data analysis, which is possible using AI
How AI is disrupting everything?
AI has hit big time in the business world of retail. In a recent announcement India’s top online retailer, Flipkart has teamed up with Microsoft to use their AI tools for a better user experience. Apple’s Siri, Google’s Assistant and Microsoft’s Cortana are all AI apps on mobiles and tablets to understand and personalize the owner’s experience. The assistant you may chat with on a helpline might well be an avatar which uses Natural Language processing to help you. The financial sector uses AI to detect potential fraud, streamline back office operations, and assist in stocks and properties management. Health care is another major area of applications. Apart from aiding clinical diagnosis it also is used in medical image interpretation and interpretation of sound patterns like heartbeats.
In heavy industry, robotics is a major user of AI. According to a research report from Forrester, Tesla’s modern car manufacturing facility has 160 multifunction robots that help process and manufacture cars alongside 3,000 humans. That is 20,000 customized cars per year. Tesla’s self-driving cars are a marvel of AI applications in transportation. Self-driving intelligent trucks are some of the transportation solutions using AI. Integrated with Internet of Things, such solutions are poised to bring a paradigm shift in the world of transportation. Disaster management and military operations also need almost real-time data analysis, which is possible using AI.
“Artificial intelligence may be the key to digesting the barrage of data from multiple sources. Crunching vast tracts of data is a growing task for defense, intelligence and security agencies,” underlines Sinha. Intelligence and security forces need fast analysis of what is going on in the air and on the ground to assess battle fields, secure environments, and decide when and how to deploy people or humanitarian aid. To unlock insights from this data, agencies are increasingly turning to use powered deep learning, with algorithms that can identify relevant content and patterns in raw data at machine speed”.
In an article in Forbes, Bernard Marr describes the Chinese activities in AI. Baidu, Alibaba and Tencent have at their command the data of a huge population plus data from their global customers. Such data is being leveraged through AI and deep learning to “roll out machine learning algorithms for voice and image recognition, as well as natural language processing, to help it return smarter, more useful and more personalized results”. These are open source solutions available to industry on the Cloud. Applications include health care and home automation. By integrating AI with Augmented Reality new tools have been created for advertising and tourism.
Budhendra Bhaduri, Director, Oak Ridge National Laboratory, USA, feels that there will be a tremendous impact on global economy from various different sectors. “I believe defense and intelligence are leading the way and so is the commercial business sector but there is general overall impact of AI on society and humanity as such that will change the way which we deal with society and societal problems making sure that societal services reach population at large.” This can include anything from agriculture to Climate Change science, security applications, enabling e-commerce and even the future of society through an interconnected infrastructure that is going to be deeply impacted by AI and deep learning.
So where does geospatial fit in?
Nigel Clifford, CEO, Ordnance Survey, United Kingdom, feels geospatial at one level is a huge dataset that can be utilized by machine learning and deep learning to help make its findings more accurate. “We look at aerial imagery, we look at auto change detection, classifying the features; we are comparing those features automatically with the previous versions of the map. Therefore we are able to spot the changed that have occurred. That is a form of machine learning and artificial intelligence. We sometimes call geospatial a golden thread which links so many datasets. It is going to be at the heart of making sense of the trillions of bits of data.”
In geospatial systems, one of the accepted methods of classification of remotely sensed data for thematic mapping is by using AI software, Neural Networks. As Sinha observes, “Typically, image processing has been utilizing the algorithms and programs that made use artificial intelligence at a very early stage itself. There are now numerous artificial intelligence techniques which allow symbolic information to be exploited in conjunction with numerical data to improve object classification performance.”
Nagalingam adds: “There have been examples where methodologies have delivered results that are even more accurate than an average human level of accuracy.”
However, the industry has not stopped at mapping and progressed from maps to apps. A map is a document which has to be used with other data to arrive at a planning decision, in other words a map is an input for an application.
“The biggest opportunity for geospatial industry is its core asset, which is geospatial data: 80% of datasets that we generate and deal with are geospatial in nature so exploiting that data using automation through AI and DL comes naturally to creating solutions for rest of the basic sectors,” points out Bhaduri. The application phase is deeply tied with the geospatial world, even if you think about agriculture, climate, connected autonomous vehicles, defense and intelligence, all the data that are being collected as part of these sectors are essentially geospatial in nature. He is confident that if there is any community that is poised to make a deep impact on AI and DL based solutions, geospatial industry is going to be a leading one.
Agrees Sinha when he says: “As the world churns out increasing amount of geospatial/Big Data, using intelligent analysis for actionable insights has become imperative to all sectors ranging from public safety, smart city, urban infrastructure, transportation etc.”
Today geospatial technology and deep learning are becoming a focal area to create bolder and better decision making. It is evident that almost every vertical will be affected in this journey of digital transformation.
Data today comes to us in many different forms. Apart from imagery we understand and define an environment through different data resources like text, voice recordings, smell, texture, etc. “All these define an environment that we are trying to understand and explain. Geospatial is a key platform which uses geographic coordinates in bringing these kinds of different data together,” says Bhaduri.
The most important part that we are looking through AI and DL is to understand the context and not just the objects. Therefore it goes beyond identifying a particular object an environment but also characterize behavior the object. Thus, a imagery also has streaming data that mounted on streets. Data comes from the driver’s cellphone to cellphone towers. So, understanding this continuous process and integrating data is critical and that is where geospatial solutions are most important.
A look at the mergers and acquisitions that have been happening in the past few years shows that geospatial technology companies have been expanding their capabilities to move from data, hardware and software marketing silos to providing complete decision making solutions. Not only that, geospatial systems have collaborated and converged with other standalone systems to create end-to-end solutions in areas like Business and Knowledge Process Engineering, e-Governance and C4ISR.
What AI brings to this scenario is the next step — the ability to handle information in a manner to enable a personalized experience. AI when combined with geospatial and other systems will help institutions to improve their performance in health care, education, energy, environment, transportation, criminal justice, and economic inclusion in terms of response and efficiency.
Effect on the human society
Clifford thinks the impact of AI on the human society is going to be profound. “Technologies should help us in terms of decisions, it should help us with lots of the mundane activities that we are involved in. It should also come up with more answers because it is going to probe and analyze datasets that human mind will never be able to cope with,” he says.
That gives rise to some of the very interesting issues for insurance companies, for governments, for public bodies, private bodies. When you have got an artificial intelligence which is making decisions, how far do you need to go to explain those decisions to the citizens, customers or the buyers of the service?
“There is some huge potential there. But also there are some speed bumps that we are going to navigate before we get to the full speed on artificial intelligence,” Clifford adds. Nagalingam feels that, “There are multiple social innovations possible and once again these are things that will impact the life led by human beings today”. He goes on to quote possible applications in poverty alleviation, migration management, epidemiology and management of epidemics.
According to a report on Maps to Models, published by the National Academy Press, USA, to do this we need analytics which could be models of physical processes that affect human activities, social system models of human behavior in a geospatial context, models of combined physical and social systems, inverse methods to infer uncertain model parameters from measurements of the real-world system, spatial statistics, data mining, and machine learning to discover trends, patterns, and associations, and spatial network analysis to examine how patterns of relations affect behavior from the individual to state level.
Clifford predicts that the use of AI in predictive modelling is going to be one of the big steps for cities, individuals, companies and governments. “That is a powerful change that you don’t have to go and do something before you begin to understand the implications. You can model it out and model it in fi ne details, which never been done before.”
Agrees Sinha: “The rise of artificial intelligence and deep learning has brought tremendous opportunities for the geospatial industry across various businesses for faster, timely and informed decisions like the agriculture insurance business solutions, mining business solutions, border control and security solutions, smart Infrastructure solutions and dealing with traffic congestion etc.”
In the short term AI will result in increasing efficiency but may result in job loss in areas where AI-based systems replace human operators but it also will throw up demand for new types for expertise called new collar workers who will join the existing white collar and blue collar workers. “I would be optimistic and say that AI is going to provide more answers and provide more freedom for individuals to make a profound impact on their country,” underlines Cliff ord.
Bhaduri also feels that automation is not necessarily a threat. “It is well recognized that the rate of automation could perceivably dwarf the rate human beings can be retrained for particular trade. So, automation could be disruptive to society in terms of making some of these functionalities obsolete or not so much effective. At the same time automation also opens up opportunities for creating new ways of looking, analyzing our environment, our datasets for creating solutions for every possible sector which in turn require new sets of skills and diverse sets of talent. So, overall it may be a threat to part of the society who will have the challenge to keep up with their occupations but at the same time there would be different kinds of challenges that would create different kinds of career options and opportunities for people to continuously innovate. So, overall society is likely to benefit from it then not”.
A report on Preparing for the Future of AI by the Executive Office of the President, National Science and Technology Council, Committee on Technology, US states that rapid growth of AI has dramatically increased the need for people with relevant skills to support and advance the field. An AI-enabled world demands a data-literate citizenry that is able to read, use, interpret, and communicate about data, and participate in policy debates about matters affected by AI. Education should empower all students from kindergarten through high school to learn computer science and be equipped with the computational thinking skills they need in a technology-driven world.
Rapid growth of AI has dramatically increased the need for people with relevant skills to support and advance the field. An AI-enabled world demands a data-literate citizenry that is able to read, use, interpret, and communicate about data, and participate in policy debates about matters affected by AI
In an article entitled “Will Democracy Survive Big Data and Artificial Intelligence?” in Scientific American, Dirk Helbing et al conclude that “Big Data and artificial intelligence are undoubtedly important innovations. They have an enormous potential to catalyze economic value and social progress, from personalized healthcare to sustainable cities. It is totally unacceptable, however, to use these technologies to incapacitate the citizen. Big nudging and citizen scores abuse centrally collected personal data for behavioral control in ways that are totalitarian in nature. This is not only incompatible with human rights and democratic principles, but also inappropriate to manage modern, innovative societies. In order to solve the genuine problems of the world, far better approaches in the fields of information and risk management are required. The research area of responsible innovation and the initiative “Data for Humanity” provide guidance as to how Big Data and artificial intelligence should be used for the benefit of society.”
How are institutions responding to these needs?
“We are researching on AI. We are using AI. And that is the way you progress. It is not just about theory, but about practice. We are investing in academic research. So we are sponsoring some Ph.D in local universities looking at deep learning and machine learning,” reveals Cliff ord.
According to the report Outlook on Artificial Intelligence in the Enterprise 2016, presented by Narrative Science in partnership with National Business Research Institute, 38% of enterprises are already using AI technologies and 62% will use AI technologies by 2018. Adoption of AI across a broad range of industries will drive worldwide revenues from nearly $8.0 billion in 2016 to more than The new Worldwide Systems Spending from Data Corporation market for cognitive/solutions compound annual growth rate CAGR) of 55.1% over the 2016-2020 forecast period. Nearly half the revenue will come from software applications and platforms. In spite of such a rosy there are some doubts. According to a survey Forrester conducted in 2016, there are also obstacles to AI adoption as expressed by companies with no plans of investing in AI which range from 42% respondents who said ‘There is no defined business case’ to 3% who said ‘Not sure what AI means’. A study conducted in Japan and reported in the IEEE Technology and Society magazine shows that while there is a general appreciation of application of AI in self-driving cars, disaster prevention and military activities, the public is not so enthusiastic about AI in elderly care, health care, child nursing, creative activities and decision making for life events.
Another major area which affects the implementation of AI models is the issue of regulations. The regulatory process is designed to ensure safety for the general public and to the users of the technologies. Taking the examples of AI applications in driverless cars, the Future of AI report states that an approach to evolving the relevant regulations that is based on building expertise within the government, creating safe spaces and test beds for experimentation, and working with industry and civil society to evolve performance-based regulations that will enable more uses as evidence of safe operation accumulates is the way to go. These efforts must also address the issues of ethics as AI systems impact individuals. Decisions taken by AI-enabled systems should be transparent and easily understood by the common person.
In conclusion, the report states that AI holds the potential to be a major driver of economic growth and social progress, if industry, civil society, government, and the public work together to support development of the technology with thoughtful attention to its potential and to managing its risks.
As Clifford sums it up: “I would still say that we are very much looking at the tip of the iceberg what can be provided here. We are looking forward to it, we are investing into it, but we have way way way to go in terms of the full potential.”