AIstralian Stories – Google AI explained

AIstralian stories

Google hosted an event titled AIstralian Stories. It was one of the most exciting yet entirely practical events to help media understand AI (Artificial intelligence) and ML (Machine learning).

Now I have a fair understanding of AI/ML, but I can’t begin to do to do this event justice. I strongly recommend you read Google’s blog from today.

What I hope to achieve is to explain the sense of wonder I felt sitting in the audience. I have paraphrased below what some speakers said. The use of the word Google below is for convenience and generally refers to Google’s AI systems.

Jason Pellegrino, Google’s Australian boss, set the scene.

It is all about AI. AI is the driver and is accelerating the incredible advances in technology. AI is being harnessed to help automate tasks that we don’t have the resources to do effectively. It is about harnessing the power of tech to solve apparently insurmountable problems.

He had several key points

  1. Your jobs will change drastically over the next 15 years. You may produce the same output. AI will underpin the tools you use and the skills you need.
  2. We need more human capital with the right skills to enable AI.
  3. Australia needs to speed up. At present only 8% of major companies are investing in AI compared to 20% in the US. We are part of the global economy and cannot afford to be left behind.

If we can capture the benefits of AI and ML, then it will determine our future prosperity.

Anil Sabharwal, Google Photos creator

AI is one of the most profound technology advancements and its happening in our lifetime. Applying AI

  • Makes goods and services more useful
  • Helps us to be more efficient and do much more
  • Can solve humanities the impossibly big challenges

AI is the science of making things smart. ML is about making machines that learn smarter without programming rules.

For example, we taught Google what a cat and dog are by showing it thousands of photos. It analysed these and set the attributes for a cat or a dog. But it advanced by drawing inferences. It can now recognise breeds and more.

We have bundled an AI that has learned a lot into Google Assistant. It will never stop learning from our human interaction. Soon it will be able to interact in a natural language.

Google provides open source TensorFlow AI language, Cloud AI via APIs and purpose-built ASIC Tensor processors to researchers and business to help solve problems.

AI is enormously processor heavy.

His first love is Google Photos, and he showed off a few things AI could do to improve images, share them and search them. While your grandfather may have had only a few cherished photos today, we have tens of thousands of significant moments. Being able to search them and share them is just the beginning.

Things like Google Lens will have improved recognition features, and real-time results are nearly there.

He finished, “The best part about creating tech is seeing what the world does with it.”

Dann Van Esch languages at Google

There are more than 6,000 languages in the world. Over 50% of the world wide web is English. AI is driving real-time translation to make content universally accessible.

But converting text from one language to another is a simple task compared to converting speech from one to another.

It is complicated model involving acoustic models, pronunciation models, language models and more. We are adding new real-time translations all the time. He referred to Google’s work with Professor Janet Wiles and Ben Foley, researchers with the ARC Centre of Excellence for the Dynamics of Language (CoEDL). They are working to transcribe and preserve endangered languages. There are over 300 Indigenous languages in Australia.

Wiles said that they had over 40,000 hours of recorded languages. It would take a researcher 1,200 years to transcribe it. Google can do it in a fraction of the time.

Dr Lily Peng – program manager Google

Peng was a real highlight simply because she was able to showcase work being done in Australia by Maxwell+ on using AI/ML to diagnose skin, breast and prostate cancer.

The problem is that globally there are not enough skilled clinicians (Doctors) to analyse the mountains of data and medical images to provide a timely and cost-effective diagnosis for all.

Skilled clinicians helped train Google to identify certain cancers in imaging and photography. Dr Elliot Smith CEO and founder of Maxwell+ said that AI aimed to transform healthcare from fixing to preventing.

AIstralian Stories summary – Anil Sabharwal

These are the key takeaways

  • Australia can, should, will be a leader in AI
  • AI is simply a tool to multiply human ingenuity

Media quizzed him on the issues of privacy. While we are trained never to take things at face value, his responses were comforting.

At the very core of Google is the will to protect its users. We provide users choice in what information they share – transparency. We let them know how it is used in simple, understandable language. Google is only a small part of the AI revolution. We do the right thing because it is part of our DNA.

GadgetGuy’s take –  AIstralian Stories was a great education

Most of us have heard of AI or machine learning. Few get to go behind the scenes and see it at work. Google’s AIstralian Stories was an incredible opportunity.

From a personal viewpoint, I am highly protective of my personal information – I was bought up that way. Yet the twenty and thirty-somethings in the audience embrace it. As one said, “I would rather Google use that information to show me something I appreciate than all crap out there.”

Google says the next 15 years are the tipping point, If so that is a lot less time than since the birth of the internet to get to this place.

For more on this eye-opening Aistralian Stories event I urge you to read Google’s blog.

AIstralian Stories