For the last 20 years or so, Sri Lanka has relied on software development and finance & accounting as its two main pillars for IT/BPM export. But times are changing. We need to shift away and look at another pillar that is gaining traction. This is Data Science and AI. So how can we make use of AI and Data science? Well that’s what we were about to find out at the first ever AI Asia Summit 2018.
Organized by SLASSCOM and held at the Kingsbury Colombo, the AI Asia Summit boasted an impressive repertoire of industry experts in the fields of data science, machine learning and AI.
DHL’s journey of Artificial Intelligence
Kicking off things at the AI Asia Summit 2018 was a keynote speech by Timothy Koi – Head of Data Analytics and Innovation Leader at DHL. Timothy gave a brief overview about DHL and how they’re not just a postal delivery service. Rather, DHL is made up of 4 businesses. These are Express, Freight Forwarding, Supply Chain and Postal eCommerce. They look at everything from blockchain to robotics to see how they can innovate to make their services better for customers.
Timothy then spoke about Trend Research. This deals with identifying upcoming trends. If DHL is able to know what trends are popular, they can use that to their advantage. To that extent, Timothy explained that DHL has converted from forecasted to predictive operations. They’ve also changed from manual to autonomous processes and transport and standardized to personalized experiences.
As examples, Timothy spoke about the robots that DHL uses. These are able to detect people and are easy to train. They can also learn multiple tasks by means of AI. The robots are located across various locations in the warehouse. There are also robots who are tasked with assisted order picking. They act as personal luggage carriers so that employees can just stack packets onto these vehicles and they would follow the employee.
Timothy also touched on other technologies that DHL is using. These include drone, AR, IoT and a connected warehouse for health and safety.
Inspirational Applications for AI in Sri Lanka
Presented by Dr. Yasantha Rajakarunanayake, the topic at the AI Asia Summit 2018 was rather self-explanatory. Can machines think? An age-old question indeed. Machines can sort of think, he explained. It all depends on what you define as intelligence. As the AI matures and can understand better, we will have AI in all our lives. But how do we know that the AI is as good as human intelligence? Dr. Yasantha explained that we just have to ensure that the AI can match human performance.
For example, if an AI doctor performs as well as a human doctor, then that is essentially good enough. While AI Is slowly becoming available everywhere, should we invest in it? We should, Dr. Yasantha explained. He predicted that by 2019, 75% of developer teams will include AI functionality in one or more applications. Further, 40% of digital transformation initiatives will be enabled by AI.
He then spoke about the different types of machine learning. One such type is deep learning. From there, Dr Yasantha spoke about the various different applications of machine learning such as image recognition, colorization, text recognition, image caption generation, healthcare and autonomous driving.
One thing that Dr. Yasantha emphasized was that the thing about AI is that it’s not very robust. If an AI program comes up with an answer, even if it’s wrong, it will be completely sure about it. This in turn can lead to incorrect decisions.
Coming back to his original topic, Dr. Yasantha explained that an application for AI in Sri Lanka would be education. AI can make learning fun and productive. AI has the opportunity to diagnose the root cause and fill learning gaps of the past. Exams too can be gamified. This would be a better measurement of skill, knowledge and experience.
Data-driven decision-making using AI
The next session at the AI Asia Summit 2018 was by James G. Shanahan. James started off by talking about his work in AI and machine learning. “When we think of AI, we think of this big vision” he explained. There are terms like Machine learning and deep learning, all of which are associated with AI.
Using advertising as an example, James spoke about the conversion rate for customers. Using AI, companies can make ads reach targeted audiences better and more efficiently. All this is using deep learning, he explained. When it comes down to it, it’s all about identifying patterns and doing the math about what a customer wants.
James’s next topic was about IoT. IoT is failing, James explained. This is because IoT devices are very cloud-centric. They rely on the cloud to carry out their instructions. To that extent, James is working with a startup that proposes a new breed of chips that are tuned to process information without wasting time connecting to the cloud. The device would do the processing in real time.
Data Science & Machine Learning in Business
Carried out by Dr. Rukshan Baduwita, the session dealt with the key drivers of the rapid growth in Data Science. This growth was due to 3 reasons. They are the increasing amounts of useful data, advancement of computer systems and the advancement of algorithms and software tools.
The main elements of a business are customers, products and services, suppliers, processes and employees. A business also has key objectives. These range from acquisition to customer relations, improving quality of goods, having a diverse portfolio, optimizing cost reduction and improving productivity. A business would also have data such as customer data, products and services etc.
Using Machine learning, you can forecast overall demand, supply, sales, share prices. In addition, machine learning and AI can be used for optimizing business operations. This could be anywhere from route planning, to logistic management, scheduling and ensemble-line optimization.
In conclusion, Dr. Rukshan summarized his session at the AI Asia Summit 2018 by explaining that in order to use data science and machine learning in a business, we should first define a problem, collect data about it, extract the data we need, clean it up and then build a model to see how we can best overcome the problem. Once that is done, we can take the model to the real world and see how it functions.
Research Opportunities in AI
The next speaker at the AI Asia Summit 2018 was also a doctor. Dr. Chrisantha Fernado became interested in evolution from a very young age. Following his primary, secondary and tertiary education, he was obsessed with figuring out the origin of life.
He then became interested in how evolutionary algorithms can be implemented in the human brain. Needing to satisfy his curiosity, Dr. Chrisantha joined DeepMInd in 2014 when it was acquired by Google. So, what does Deep Mind to? Well, it’s to solve the problem of artificial general intelligence. Once that it solved, they can use it to solve everything else. So how would you do research? The answer for that, according to Dr. Chrisantha was to work on important, easy problems.
So, what are the important problems in AI? For starters, there’s the problem of continual learning. This is where rather than being programmed with algorithms, the AI would always be learning by itself. So what are the research opportunities for Sri Lanka? Well, according to Dr. Chrisantha, we need to have some serious community building. This can be done with meetups and practical groups. Once these are firmly in place, they can act as a hub to share knowledge and experience with relation to AI and machine learning.
What’s next in AI?
The first panel discussion at the AI Asia Summit 2018 consisted of the above speakers. As the topic suggested, the primary goal of the discussion was to see where AI can help us in the future. One key highlight of the discussion was made by Dr. Yasantha Rajakarunanayake.
“AI will flatten the landscape for education”, he explained. Dr. Yasantha, Dr Rukshan, Dr. James and Dr. Chrisantha all emphasized that an economy is made up of 3 parts: Labor, capital and Natural resources. If you get rid of labor, the other two factors will still stay.
In conclusion, AI will be a great way for SL to catch up to be one of the leading countries. If two guys can create Facebook in a garage, then two people from Sri Lanka can do something greater, added the panelists.
More on AI and Machine Learning
Dr. Sadeep Jayasumana spoke to us about Deep Learning and Computer Vision. He explained that a deep neural network is like a newborn baby. It knows nothing. Because of that, it needs to be trained. With time, the neural network learns how to perform. This is where deep learning comes into play.
In order to achieve maximum results from a deep neural network, we need 3 things. We need vast amounts of data. We also need an equal amount of processing power. Lastly, Dr Sadeep joked that we also need a lot of students. This was to mean that we need data scientists and those with an affinity for AI and machine learning to pursue further studies and careers in these fields. In conclusion, Dr. Sadeep emphasized that deep learning is dominating AI research. People are heavily investing in it and it is definitely too important to ignore.
The link between AI and Simulation
Even though you may not know it, simulations and machine learning are fundamentally linked. This is what Dr. Adam Easton was on stage to talk about. He started off by explaining data-driven machine learning. From there, he spoke about model driven machine learning.
So you can run a simulation, and take its problem data and try as many iterations as you like. From there, you can use AI to make decisions and control the simulation is based on the results. Dr. Adam also spoke about Reinforcement Learning. This is a type of machine learning technique that enables an agent to learn in an interactive environment by trial and error using feedback from its own actions and experiences. From there, Dr. Adam went on to give a number of case studies that made use of both AI and simulation such as medical treatment.
How Microsoft empowers businesses with AI
Carried out by Christopher Gunawan – Technical Lead Architect of Microsoft, the session was an in depth look at how Microsoft wants to empower developers to innovate. But that’s not all. Microsoft is also looking to empower organizations to transform industries. Lastly, the company also wants to empower people to transform society.
Christopher then went on to talk about the various products and services that fall under Microsoft’s AI umbrella. These include Azure AI Services, Azure Infrastructure, and Azure Tools. As an example, Christopher explained how these technologies help in not only object recognition by classification but on a more specific level as well. He then went on to explain about digital agents. These are essentially chatbots who would engage with customers. The bots would not only reply to FAQs but actually carry out meaningful conversations.
So how does Microsoft help businesses go about this? Well, after you build the AI model for your business, you can implement it into the desired applications. From there, you can use AI to transform your business processes. These can range from Financial services, to retail and manufacturing.
The Landscape of AI in Sri Lanka
Following a session on research case studies by Dr. Subha Fernando and Prof. Asoka Karunananda at the AI Asia Summit 2018, we had Svein Olaf Olsen – Co-Founder of ANZYZ Technologies LTD on stage as the final speaker of the AI ASIA Summit 2018. Svein spoke about their companies’ efforts to deal with AI.
In case you didn’t know about the company, ANZYZ is an Artificial Intelligence company. They deal with big data. As such, the company develops tools to understand and process big data. These tools are used by leading companies in Norway.
In comparing ANZYZ to other self-learning technologies, Svein emphasized that they are far ahead. ANZYZ CCL (Corpus Cube Linguistics) has been used in a variety of threat detection scenarios. A prime example of this was in the detection of bombs and possible terrorist attacks. Svein explained that the company is working on further potential cases in the legal and energy sector.
Opportunities for AI Startups in Sri Lanka
This was the last panel discussion at the AI Asia Summit 2018. Is Sri Lanka behind in terms of opportunities for AI startups? The panelists agreed that we are, but then again, there is no reason as to why we can’t be at the forefront. The mindset in Sir Lanka is that companies think they know all about this business and they don’t need the analytics. In reality, we need the analytics to stay ahead of the competition. This is because when disruption happens, it happens really fast.
With regard to brain drain and creating more opportunities for those seeking foreign employment, the panelists also agreed that we should create more opportunities in Sri Lanka. This would mean that even those who have gone abroad for their studies could be enticed to come back and work in Sri Lanka.
For startups, you need to always stick to your goal. You also need a good team. Thirdly, you should be smart enough to grab an opportunity and come up with a good proposal then people will take notice of you. We need to be a little bit more serious on AI. We also need to understand the importance of research and collaboration. After all, this is not a competition.
With the closing remarks being delivered, the first ever SLASSCOM AI Asia Summit 2018 drew to a close. Overall, we gained a wealth of information with regard to the latest trends in AI and machine learning. It also sets a stepping stone for future summits such as this as well. Onwards to the SLASSCOM AI Asia Summit 2019, then.