MEDIUM DR. JANAK GUNATILLEKE

Keen to explore the latest developments in AI and to learn more about how it was being used in the region, I attended the inaugural AI Everything summit in Dubai. Having also recently been thinking about the current state of AI in Sri Lanka and what it could be (or should be) in the future, I found that the summit provided some very useful perspectives. The perspectives were in the context of two themes. The first was through understanding more about the United Arab Emirates (UAE) strategy for AI and progress with its implementation. The second was through experiences and guidance shared by the speakers from across government, industry and academia. I believe Sri Lanka should consider the following 5 key points in developing their national strategy for AI.Read the article on medium.com >

MEDIUM CD ATHURALIYA

This is going to be a short one. While others argue over GPT-2, we had to dig a little deeper into Google BERT. If you are a follower of deep learning research, specially on NLP side, BERT needs no introduction at this point in time. Since its initial release last October, dozens more resources have been added highlighting the significance of this work by Google. The official blog post itself is a very good resource to understand the basics. Besides the significance, I think the code release with good documentation (a not-so-common practice in research, sadly) immensely contributed to the popularity it gained in such a short period. The repository covered almost all the essentials, with very few exceptions. One was online prediction support.Read the article on medium.com >

MEDIUM DR. JANAK GUNATILLEKE

According to Gartner analysts, up to 85% of big data projects fail. I suspect that for AI projects, especially if you count the ones that never start, the failure rate is even higher. What are the reasons for the high failure rate with AI projects? Drawing on my technology consulting experience from working at Accenture UK and wider transformation project experience within the highly complex environment of the UK National Health Service (NHS), I wanted to identify the causes and potential solutions. As highlighted in a previous post, although there is a widespread belief among senior executives that AI can give their companies or help them maintain a competitive advantage, only a few had implemented AI in a meaningful manner.Read the article on medium.com >

MEDIUM SAMEERA RAMASINGHE

Convolution is an extremely effective technique that can capture useful features from data distributions. Specifically, convolution based deep neural networks have performed exceedingly well on 2D representation learning tasks, e.g. image analysis. Given this success, it is natural to investigate how to use this concept to capture features in different settings. However, most state-of-the-art deep neural networks work only on Euclidean geometries and extending this concept to other manifolds such as spheres and balls is an open research problem. Representing data in spheres/balls can be quite natural and effective in cases such as analyzing 3D data. However, achieving this task is not straightforward. The main difficulty of adapting convolution to such manifolds is that in contrast to planar data, the spaces between adjacent points are not uniform.Read the article on medium.com >

MEDIUM DR. JANAK GUNATILLEKE

A couple of weeks ago I attended the World Summit AI 2018 in Amsterdam. Coming from a non-technology background (I originally trained as a medical doctor) and as someone who had since worked as a management consultant and more recently have been working with various startups (including ConscientAI), I was keen to understand more about AI from a business point of view. WSAI 2018 was an amazing gathering of more than 6,000 attendees from across 160 countries who are enthusiastic about the potential of AI to solve various challenges. Thinking back at the various speakers, workshops and conversations over coffee, I wanted to share my thoughts and learnings.Read the article on medium.com >

MEDIUM DR. JANAK GUNATILLEKE

A 2017 survey of more than 3,000 executives by MIT Sloan Management Review and Boston Consulting Group revealed that almost 85% believed that Artificial Intelligence (AI) will give their companies or help them maintain a competitive advantage [1]. Non-public AI companies raised $15.2B funding in 2017, an increase of 141% compared to 2016 [2]. Large companies such as Google, Facebook, Amazon and Microsoft are all heavily investing in AI. Google spent $660m acquiring British AI startup DeepMind in 2014, which has been in the spotlight for beating the World Champion at Go and for an exciting (yet controversial [3]) research project with the UK National Health Service.Read the article on medium.com >

MEDIUM CD ATHURALIYA

Yes, that's correct. Well, I think so and let me explain. I think the biggest challenge of AI right now is making it an industry or an engineering field. All the other challenges were there for a while if you really think about it. There can be slow downs and booms in fundamental research, talent can be in short supply, still we will make progress. But I think we are yet to define AI as an industry. And we need to do it right away! But why..? We all can see how AI is changing the world; from photos we take to how we move around. It is becoming more common and ubiquitous everyday. From a technical point of view this means the systems we use in our day-to-day life have much more AI underneath.Read the article on medium.com >

MEDIUM SAMEERA RAMASINGHE

Since Ian Goodfellow first proposed the idea of GANs ( https://arxiv.org/abs/1406.2661), it has become a buzz word within ML community, simply because it works stunningly well (given that you came up with a perfect architecture). Many people, specially Yann LeCun, a who is considered as one of the giants in Deep Learning, stated at some point that GANs are a significant breakthrough in deep learning. One thing I have noticed is that many people who claim to be familiar with GANs lack the theoretical foundation that lies beneath it, which is important.Read the article on medium.com >