Machine Learning and Artificial Intelligence, Speaking to the Machine:
An outsider’s view as I do not see myself as an expert in this space.
Brief overview:
Machine Learning and Artificial Intelligence (ML and AI) has of late become the talk of the town where the excitement of AI has become a hype word. As explained by Shivon Zillis the conversation has moved from academics and founders to Investors and now the focus is in the crosshairs of companies looking not to be left behind.
Before I jump into the topic of this essay, I want to highlight four key ML and AI feature articles. The web links below will help build colour for those new to the topic. The list covers the biggest news items in the ML and AI space that I have read over the last couple of months:
Zero Shot — Google Translate teaching itself and allow for translations between different combinations of language.
GAN — Or Generative Adversarial Networks of 2 separate deep neural networks competing with each other.
AI Partnership — The joining forces of the largest Tech companies on AI
Poker Playing AI — Inside the 20-year Quest to build computers that play Poker.
Another great insight that grabbed my attention by mentioned in Rob May’s last weekly email. May noted in his email that Google’s 2016 Annual Report included the line item of Machine Learning and Artificial Intelligence.
The significance here is that 2016 is the first year that Google reported a line item for ML and AI May continues to discuss his thoughts on the hype of AI and concludes that “we may be close to a tipping point — not a tipping point towards general AI — but a tipping point where, if you don’t have AI in your business, you are left behind”.
My view is that the hype of AI could potentially not be a hype in 2017 — Perhaps we are in the initial ‘wave’ of this new technology frontier. I feel this ‘wave’ is still in the early stages where the corporate and research partnerships in the space will help drive the machine intelligence development.
Speaking to the machine:
One aspect of AI includes the AI assistant and the Podcast titled, Will Siri Ever Outsmart Us, by Bloomberg dives deeper into this question. The podcast looks at the beginning of the A.I interaction and the journey starting with one of the Godfathers, Terry Winograd.
In 1968 Terry created SHRDLU an Artificial Intelligence program with the ability to interact via text. SHRDLU could be classed as the first chatbot/interactive assistant, where the main tasks of SHRDLU involved stacking blocks where the interaction was based on standard language. You can sense the ‘aww’ and excitement from the journalist when reviewing past footage of SHRDLU exclaiming how amazing SHRDLU was as an interactive machine. Imagine how people would have reacted seeing SHRDLU in the 60’s for the first time.
Winograd, however, found that the understanding of text speech regarding language and context with humans is extremely tricky — Especially when trying to program basic interaction into an algorithm. The podcast on Decrypted walks over the complexity of speech and the difficulties in a demonstration with today’s assistants such as Alexa, Google, Cortana, and Siri. This segment is a good listen as it highlights where we are today in speech interaction with an AI — To sum up, human thought is not programmed the same way as computers, e.g: If A, then B, if B, then C etc.
It was with this difficulty that most pioneers in the space found themselves in an AI Winter — research projects defunded and startups failed from the mid-1980’s as the idea and future of A.I became frozen in time.
Since then we have seen the extrapolated growth in technology and the development of the next generation machine intelligence with the likes of, Google Assistant, Siri, Alexa, and Cortana. 2016 also saw the rise of ‘bots’ where social communication tools began to allow open access for developers to build ‘bots’ that sit on a chat service. The main aim here is to enable brands and companies with a ‘bot’ (AI) to communicate with clients and assist in the e-commerce space.
The corporate play:
AI and digital assistants are becoming more of a focus, and it is with the increase interest of corporations that I believe we will see a rise in investments and funding. To add to this, Accenture has recently released a 2017 Tech Vison report where the first trend discussed is dedicated to AI — The title, AI the new UI says it all. It is worthwhile to also highlight a key extract from the report. “AI is poised to act as the face of a company’s digital brand and a key differentiator — and become a core competency demanding of C-level investment and strategy”.
The report also shares that of the executives surveyed, the theme and importance of AI will be crucial. Companies will be investing extensively in AI related technologies over the next three years. Elsevier’s CTO, Dan Olley noted that “if CIOs invested in machine learning three years ago, they would have wasted their money. But if they wait another three years, they will never catch up”.
This and the recent breakthroughs in our understanding of ML and AI could be the sign that Winter is over and the first rays of the Spring sunshine are breaking out from the cold deep winter clouds of AI.
One point I want to raise from this article is that customer problems should not be lost. The more the CEO and CIO focus on AI the more the focus should increase on their customer’s challenges and the core value of the business and concentrate on solving problems. Michael Dolbec of GE Ventures helps raise this point with his investment theme in AI — “We fund valuable outcomes, not science projects.”
The rise of the machine:
But, is it scary to think this could actually be the rise of the machines — I leave this question for you to ponder. Are we witnessing the beginning of Sky Net?
Another concern people might have, is the idea that machines are taking our jobs. This is not a concern for me, where I am excited to see the next phase of this technology develop as I feel the removal of repetitive work will give people the chance to focus on bigger tasks and help improve creativity and productivity.
Now, let’s fast-forward into the future — How long will it be until we find ourselves interacting on a daily basis with a machine to complete most our tasks through conversation. Could it be too far-fetched to think one could even start connecting on a more personal level — The film ‘Her’ might be a glimpse into the future where the Joaquin Pheonix’s character falls in love with a sentient AI operating assistant.
