Follow Tweet This Facebook LinkedIn google+
Industry talking to customers What's this?

Deep learning and artificial intelligence: Making a big deal of big data

Published: December 11th, 2017 By: Suzanne Robicheau

AWS

Connect with AWS products and services that will help lead your business to 2025AWS DeepLens
Looking for a new way to learn machine learning? Let a machine teach you with AWS DeepLens, the world’s first deep learning enabled video camera for developers. Designed to connect securely to a variety of AWS offerings, including AWS IoT, Amazon SQS, Amazon SNS, and Amazon DynamoDB, AWS DeepLens uses Amazon Kinesis Video Streams to stream video back to AWS and Amazon Rekognition Video to apply advanced video analytics. Easy to customize and fully programmable with AWS Lambda, AWS DeepLens runs on any deep learning framework, including TensorFlow and Caffe.Amazon SageMaker
Amazon SageMaker offers developers and data scientists a quick and simple way to build, train, and deploy machine learning models at any scale. A fully managed service, it eliminates the heavy investment of time, specialized expertise, compute time, and storage space and removes barriers to machine learning. SageMaker includes hosted Jupyter notebooks to facilitate visualizing training data stored in Amazon S3 and comes pre-installed with the 10 most common machine learning algorithms. Models are deployed on an auto-scaling cluster of Amazon EC2 instances.

Amazon Rekognition

Rekognition is an artificial intelligence designed for deep, learning-based image recognition. Based on the same technology developed by Amazon to analyze billions of images each day for Prime Photos, Amazon Rekognition uses a probability-based model to detect and identify objects, scenes and faces in images. Why scroll through endless photos looking for shots of people, places and things when Rekognition can detect them in a matter of minutes and present you with numerical proof of the probability of accuracy. Better still, there are no minimum fees or upfront commitments. Pay for the number of images you analyze and the metadata you store. Common uses include celebrity recognition, demographic analysis, photo resource management, and privacy monitoring.

Deep Learning Artificial Intelligence: Making a Big Deal of Big Data

There are many predictions for new discoveries in the field of deep learning and artificial intelligence, with massive disruptions forecast by 2025 for a broad segment of industries, including education, healthcare, retail, transportation, aerospace, farming, and home maintenance. Skeptics argue that we still have a long way to go, yet there’s mounting evidence that we’re very close.

Ramy Sedra

“The thing to keep in mind is that we have the technology. What we’re doing is  already widely used by media and online services,” says Ramy Sedra, a data strategist and PwC Partner who leads a thriving Canadian Data Analytics practice with capabilities in business analytics, Artificial Intelligence (AI), Big Data, IoT, digital apps, and visualization.

Sedra points to the significant change going on today with smart cars, purchase prediction, speech recognition technology, robot-assisted medical interventions, and virtual assistants. “When we make predictions for the next decade, we’re talking about the adoption rate of AI,” he explains, “and what we’re seeing right now is significantly greater than anything experienced in previous technology revolutions.”

AI: the future of banking?

In the face of ever-increasing amounts of digital data, businesses are looking to AI to gain analysis and insights. Banks in the UK are already using chatbots to answer questions from customers and a March 2017 report from professional services giant Accenture (Accenture Banking Technology Vision 2017) found that bankers see Artificial Intelligence as the key to creating a more human customer experience. In fact, there is some thought that AI will soon become the main way that banks interact with their customers.

A $40B business by 2025

Based on research conducted in 12 developed economies, Accenture also predicts that by changing the nature of work, AI could increase labour productivity by up to 40 percent and double annual economic growth rates by 2035. Business is betting on a similar scenario, with findings in the MoneyTree Report from PwC Canada and CB Insights that AI attracted $162 million in investments in Canada alone for the first six months of 2017. Market intelligence firm Tractica shares this confidence in AI, forecasting that annual worldwide AI revenue will grow from 643.7 million in 2016 to almost $40 billion by 2025.

Canada’s competitive edge

In recent comments about his  government’s support of innovation, Prime Minister Justin Trudeau compared AI to electricity in terms of its impact, noting that AI is poised to cut across nearly every industry in the country. According to Prime Minister Trudeau, failing to embrace AI immediately would cost Canada a competitive edge, discourage top talent, and have dire consequences in terms of jobs.

The labour question

Ironically, it’s jobs that many people point to when they look to a future dominated by AI. Secretaries worry that digital assistants will answer phones and disburse information. Shop clerks are concerned that robots will stock the shelves and check inventory. Financial advisors see algorithms as the up and comers. Translators worry they’ll lose work to software, and cab drivers know that taxis will soon drive themselves.

“There’s no question that existing jobs will change dramatically,” says Ramy Sedra, “but there will also be significant employment in fields we haven’t really considered. “I understand the fear around AI taking jobs, especially those involving repetitive tasks, but when it comes to the future, I believe that there will be more new opportunities than old jobs lost.”

Sedra likens commercial success to observing the laws of gravity, saying that in order to stay in orbit, businesses have to adapt to emerging technologies. “Netflix is a good example of the disruptive potential of using AI and machine learning to mine and understand the customer base,” he explains. “Not only did it disrupt the industry, but from there it moved to creating and distributing content.”

Like Prime Minister Trudeau, Sedra predicts dire consequences for countries and businesses that choose to lag behind, encapsulating his point with one simple question: “When was the last time you rented a movie from Blockbuster?”


Analytics and oil: Ambyint monitors remotely

As the CTO of Ambyint, a Calgary-based leader in artificial lift optimization solutions for the oil and gas industry, Ryan Benoit is excited by the opportunity to use big data to enable intelligent and autonomously operating oil wells.

“Removing the manual component of data analytics reduces labour costs and offers our customers an easy way to monitor their wells 24/7,” says Benoit. “By automating pump optimization and augmenting remote capabilities, Ambyint decreases the frequency of visits to well sites and increases the number of wells an operator can manage. As a result, there are significant improvements in overall production efficiency.”

Real-time analysis, real-time improvements

Amazon Web Services (AWS) enables Ambyint to remotely monitor oil wells and provides the people who manage them with real-time analysis and recommendations for improvement. Deployment speed keeps customers agile, while integrated communications options, including a satellite network, provide secure access in even the most remote locations. A proprietary device equipped with sensor technology and machine intelligence detects problems and alerts operators to impending leaks before they pose any danger.

“AWS elevates Ambyint’s platform by allowing us to offer our customers high impact data and a comprehensive software suite with web applications,” says Benoit. “Equally important, it lets us focus on what we do best.”

Amazon’s Deep Learning Commitment

More than two decades of heavy investment in artificial intelligence has revolutionized everything at Amazon from supply chain forecasting and capacity planning to Amazon.com’s recommendations engine and the robotic picking routes in its fulfillment centers. Natural Language Understanding and Automated Speech Recognition drive Alexa, the Prime Air drone initiative, and Amazon Go.

Three layers comprise Amazon’s AI stack:

  1. Frameworks and Infrastructure with tools like Apache MXNet and TensorFlow
  2. API-driven Services for adding intelligence to applications
  3. Machine Learning Platforms for data scientists

 

Enjoy this? Read others in the What’s Possible series.

What’s Possible 2025 | Voice and picture recognition

What’s Possible 2025 | Smart government

What’s Possible 2025 | The citizen experience