“ Artificial intelligence and Big Data are two burgeoning technologies, full of promise for businesses in all industries. However, the real revolutionary potential of these two technologies is probably their convergence. Discover the possibilities offered by the alliance between Big Data and AI. “
source : www.lebigdata.fr
Introduction to Machine Learning Unit-3 for II MECH
Artificial Intelligence and Big Data
1. University of Burgundy – Dijon
Master’s Degree DB & AI
November 2018
Artificial Intelligence and Big Data
a revolutionary convergence
“ Artificial intelligence and Big Data are two burgeoning technologies, full of promise for businesses in all
industries. However, the real revolutionary potential of these two technologies is probably their convergence. Discover the
possibilities offered by the alliance between Big Data and AI. “
Hatim El-Qaddoury
hatim-elqaddoury@outlook.fr
2. University of Burgundy
Master’s Degree DB & AI
Artificial Intelligence and Big Data
a revolutionary convergence
2
Outline
The revolution of AI and Big Data: what is artificial intelligence?.....................................................................3
Artificial Intelligence definition of a technology that inspires.............................................................................3
Big Data and AI are the next digital disruptions..................................................................................................3
What are the challenges for Big Data and Artificial Intelligence?......................................................................4
Can Big Data solve the problems and dangers of artificial intelligence? ...........................................................5
Example of Artificial Intelligence feeding on Big Data .......................................................................................6
The tools of artificial intelligence for the ABI.................................................................................................6
Data Crawlers...................................................................................................................................................6
Natural language processing ............................................................................................................................6
Machine Learning ............................................................................................................................................7
Google, Huawei, Apple, IBM ... who are the giants of artificial intelligence and Big Data................................7
Google and DeepMind .....................................................................................................................................7
Amazon ............................................................................................................................................................7
Apple................................................................................................................................................................8
Facebook ..........................................................................................................................................................8
Microsoft..........................................................................................................................................................8
Artificial Intelligence example: 5 big data trends that will lead the evolution of AI in 2019 ..............................8
Big Data self-service tools available on the web .............................................................................................9
Analytical technologies are struggling to adapt ...............................................................................................9
Data cleaning becomes an industry..................................................................................................................9
The democratization of data.............................................................................................................................9
3. University of Burgundy
Master’s Degree DB & AI
Artificial Intelligence and Big Data
a revolutionary convergence
3
• The revolution of AI and Big Data: what is artificial intelligence?
We are at the dawn of a technological revolution of greater magnitude than the internet and mobile
communication technologies. In 1965, Gordon Moore, co-founder of Intel, theorized that computing power
would be able to double every 18 to 24 months. For the next 50 years, his theory proved to be accurate. The
High-Tech sectors of robotics or biotechnology have made incredible progress.
Today, however, technologies like AI and Big Data are evolving even faster. The respective exponential
growths of these two technologies are about to come together, allowing each to grow even faster. Artificial
intelligence is no longer simply a film or a book. Elon Musk, Stephen Hawking and even Cédric Villani are
some of the personalities to discuss the consequences on a large scale.
• Artificial Intelligence definition of a technology that inspires
In 1990, a group of scientists began to decode the human genome. A process that would take them no
less than 13 years, and would cost them $ 2.7 billion. This decryption would not have been possible without the
help of immense computer power and custom software. Thanks to the fall in computer prices, the researchers
were then able to undertake editing the genome using the CRISPR technique. At present, analytical technologies
of big data will make it possible to develop medical treatments adapted to each one according to his genetic
code.
Autonomous cars have always held a special place in science fiction. Today, reality is catching up with
the imagination. In 2009, many luxury brands incorporated assisted navigation systems and adaptive data-driven
channel change software. More recently, Tesla has used Big Data and AI to create autopilot functionality.
For their part, Nvidia and Alphabet use artificial intelligence to make real-time detailed maps used by
their test vehicles to visualize the world. They are based on deep learning which itself is based on a network of
neurons. The commerce industry is also evolving. Product development and marketing are now driven by AI
and Big Data. All these fascinating innovations have been made possible by the encounter between computing
power, big data and artificial intelligence.
• Big Data and AI are the next digital disruptions
The big data and artificial intelligence technologies are both inextricably linked, so that a Big Data
Intelligence can speak. AI has become ubiquitous in companies in all industries in which decision-making is
transformed by intelligent machines. The need for smarter decisions and big data management are the criteria
that drive this trend.
4. University of Burgundy
Master’s Degree DB & AI
Artificial Intelligence and Big Data
a revolutionary convergence
4
The convergence between Big Data and AI seems inevitable as the automation of smart decision-making
becomes the next evolution of Big Data. Rising agility, smarter business processes and higher productivity are
the most likely benefits of this convergence.
The evolution of data management did not go smoothly. Much of the data is now stored on a computer,
but there is still a lot of information on paper, despite the possibility of scanning paper information and storing
it on disks or in databases.
You just have to go to a hospital, an administration, a doctor's office or any business to realize that a lot
of information about customers, vendors, or products is still stored on paper. However, it is impossible to store
terabytes of data produced by streaming video, text and images on paper.
The mere fact of collecting or having access to large sets of data is not enough to produce a result. Most
of us are not sufficiently prepared for the knowledge extraction and the demand for rapid decision-making
required by customers and markets to maintain a competitive advantage.
Today, the use of machine learning, expert systems and analytical technologies in combination with Big
Data is presented as the natural evolution of these two disciplines. Convergence is inevitable.
The Internet of Things also represents a convergence between Big Data and Artificial
Intelligence. Without a digitized human brain intelligent enough to allow humans to use an IoT network that
can process, distribute and collect Big Data, it will not be possible to set up such a network.
Even the sensors, chips, network nodes and software that make IoT networks work on the cloud will be
related to artificial intelligence. This phenomenon is already in place in the field of Machine to Machine
communications.
The capture data to identify trends or patterns in the behavior of customers or employees can be very
helpful. However, the extraction of meaning, and its automation, to discover optimal methods of improving
productivity or problem solving could be even more useful.
Artificial intelligence will be used to extract meaning, determine better results, and enable faster
decision-making from massive Big Data sources. In a world where Big Data is ubiquitous, the extraction of
meaning, the monetization of data will be led by artificial intelligence for the future of business and the
development of the planet. The convergence of Big Data and AI could help overcome challenges such as
unemployment, the environment, the economy, security or health.
Automation of decision making is slowly becoming the norm. Many problems concerning the ethics of
artificial intelligence have yet to be solved. Systems capable of learning autonomously, responsible for
determining which Big Data should be identified and used, will require human management, at least initially.
In the fields of healthcare, law, banking, advertising, fair trade, security or finance, big data alone is not
enough. It is necessary to use artificial intelligence in addition.
It is therefore important not to make the mistake of perceiving these two technologies as two separate
tendencies. Your business might miss an opportunity. This convergence will have a direct impact on your
employees, your customers, your services and your market and must be considered.
• What are the challenges for Big Data and Artificial Intelligence?
5. University of Burgundy
Master’s Degree DB & AI
Artificial Intelligence and Big Data
a revolutionary convergence
5
For the moment, AI is not regulated specifically. Many people express security concerns. This problem
needs to be resolved quickly. Any information can be easily stolen by hackers. Highly sophisticated models
make us vulnerable to many threats.
Moreover, many worries about the control around this technology. The lack of laws to govern the sales
and purchase of artificial intelligence software. If these programs are intended to control traffic, health systems,
or the stock market, it is necessary to put in place governance laws.
There is no doubt that autonomous decision-making is the future. However, again, many fears are
emerging about the authenticity and ethics of artificial intelligence and Big Data. The accumulation of data on
cloud servers and its accessibility to fraudsters can be fatal for businesses.
All these challenges are daunting. They give rise to suspicion around this convergence between
Artificial Intelligence and Big Data. It is important to remember that technologies are only disruptive when we
are poorly prepared.
• Can Big Data solve the problems and dangers of artificial intelligence?
Over the last four years, agreements between large companies and startups dedicated to artificial
intelligence have increased significantly. This number increased from 160 in 2012 to 658 in 2016. Companies
use AI for a wide variety of uses, ranging from autonomous car development to remote emotion detection.
Apart from these uses, artificial intelligence can be even more useful for businesses through what is called
Account-Based Intelligence.
Account-Based Intelligence is the latest iteration of the dream of sales and one-to-one marketing. Today,
we are closer than ever to achieving this utopia.
6. University of Burgundy
Master’s Degree DB & AI
Artificial Intelligence and Big Data
a revolutionary convergence
6
• Example of Artificial Intelligence feeding on Big Data
First, we are generating more data today than ever before. Every second, humanity produces 6000
tweets, 40,000 Google searches, and 2 million emails. By 2019, global web traffic will surpass 2 zettabytes per
year.
This huge amount of data is the first step towards Account-Based Intelligence, because the ABI requires
granular information about each target company. However, it also raises a new problem. Companies must find
how to turn this data into exploitable insights.
Indeed, this task is impossible to accomplish using traditional marketing tools or simple Google
searches. The web is too massive is disorganized to achieve it as well. Many companies spend millions of dollars
to mix data sources and solution points, which ultimately results in only a very low conversion rate. For good
reason, this method usually results in sending the wrong message to the wrong people at the wrong time.
• The tools of artificial intelligence for the ABI
Until recently, computers struggled to interpret unstructured data like Facebook content and YouTube
videos. However, with recent advances in cognitive computing and processing power, things are changing.
However, this change can benefit businesses for their sales and marketing. Indeed, information on
business leaders, the decisions they make, their attitude and demographics are not stored properly within small
databases. They are scattered in social media publications, browsing history and geolocation data. Today, new
tools allow startup leaders to make sense of this data.
• Data Crawlers
The Data Web Crawlers undermine autonomously in search of unstructured data. They examine entities,
establish relationships, and create customer profiles. With an estimated 70 percent increase in data per year, it
is critical that these programs continually scan the web for the most relevant information.
Startups can use them to deploy the ABI. For example, to find new customers, browsing the web can
reveal a niche of customers whose demographics match those of the best current customers.
In 2015, Microsoft acquired Mantanani for this purpose. By using crawlers, the startup can explore a large
amount of non-relational data. She then recovers insights from different sources faster and more accurately than
humans.
• Natural language processing
The natural language processing can examine the interactions between computers and humans to extract
meaning from conversations. By spotting some words or phrases, this technology helps to analyze feelings about
the brand. It also predicts which audiences will be more receptive to the company's message. This is essential
in order to communicate the right message to the right people, which is the primary criterion of the ABI.
If the company wants to know what people are saying about its products on social networks, natural
language processing can explore social media publications, link them with certain consumer groups, and find
out what's important the most for each group. This system can be used to respond to consumer criticism and
positive reviews, to solve problems, and to improve a product.
If you want to try this technology for yourself, be aware that the IV.AI startup allows anyone to try out
their natural language processing platform. Type any phase to know the emotion that corresponds to it.
7. University of Burgundy
Master’s Degree DB & AI
Artificial Intelligence and Big Data
a revolutionary convergence
7
• Machine Learning
The Machine Learning allows computers to learn and act without being programmed
explicitly. This technology looks for patterns within the data to drive the actions of an Artificial Intelligence
program, considering the context. The true ABI requires dynamic templates, and the Learning machine
automatically adjusts them as new data emerges.
Without even knowing it, new companies are already taking advantage of Machine
Learning. Facebook uses this technology to personalize the news feed based on clicks and likes. Other
companies use this technology to predict customer loyalty or purchasing behavior, predict product performance,
or anticipate risks.
Google Now is probably the most advanced Machine Learning app yet. She learns user habits, mimics
their conversation style, and provides them with smart recommendations. For example, if the user needs to go
to the airport for a flight that will take place in 30 minutes, Google Now can analyze the traffic delays and
schedule an Uber that will take him there on time.
Artificial intelligence is strong, it without a doubt a great technology. It can find data inaccessible to
humans, and distill meaning with great precision. Combined with the ABI, it can also guide the company to its
next best customers. This technology will be the biggest change of the century in the field of business, and the
revolution is just beginning.
• Who are the biggest players of artificial intelligence and Big Data?
Artificial intelligence is a technology in full swing, and many startups around the world are looking to
seize the opportunities it offers. However, as Prometheus seizes fire, the tech giants are determined to keep this
precious resource for them.
Thus, according to CB Insights, 115 of the 120 AI firms that left the market in 2017 did so through an
acquisition. The Silicon Valley behemoths are willing to spend billions to buy the most promising startups, and
AI is now the new high-tech war front.
• Google and DeepMind
Google is clearly determined to dominate the nascent AI industry. Over the past four years, Mountain View has
acquired 12 startups in this area. However, Big G is not currently seeking to market a product. His goal seems
to be to improve his various services through AI. In parallel, it is also developing its TensorFlow development
platform and its Tensor AI chip.
In 2014, Google bought DeepMind for $ 500 million. This London company also counts among the leaders of
AI. It has a predominant role in the field of artificial intelligence research and Machine Learning. In particular,
she created an AI capable of detecting eye disorders with the same precision as a human expert, and
developed an AI assistant for doctors and nurses.
https://youtu.be/rsN690cfWsM
• Amazon
The e-commerce giant is also involved in the field of AI. It offers products and services for individuals and
businesses. Thus, Amazon Echo brings the artificial intelligence in the households of the individuals via the
voice assistant Alexa. Similarly, the AWS Cloud offers three leading AI services for professionals: Lex is a
8. University of Burgundy
Master’s Degree DB & AI
Artificial Intelligence and Big Data
a revolutionary convergence
8
version of Alexa for business, Polly transforms text into speech, and Recognition is an image recognition
service. https://youtu.be/2DtyjC0UxTw
• Apple
Between 2016 and 2018, Apple acquired four artificial intelligence startups. One of them has enabled the
creation of Faced, the revolutionary facial recognition system of the iPhone X. Similarly, the Apple has been
devoting for several years its virtual assistant Siri. Recently, Apple has also taken over the former director of
Google's AI. There is no doubt that the Cupertino company has many surprises for the future ...
https://youtu.be/K72wjPomTe4
• Facebook
The Facebook AI Research (FAIR) brings together four artificial intelligence labs scattered around the
world. Their objective? Use AI to understand how humans communicate. Recently, the firm acquired four
startups of artificial intelligence. The most recent is Ozlo, which seeks to create a better virtual assistant for
Messenger. https://youtu.be/-CRJLam3BNc
• Microsoft
The creator of Windows also has many AI projects, both for the general public and for businesses. For
consumers, the Redmond company develops the virtual assistant Cortana for Windows or the chatbot Zo. For
professionals, Microsoft offers various AI services on its Azure cloud: chatbots, machine learning, Cognitive
Computing ... https://youtu.be/XopvSz4GpEc
• AI example: 5 big data trends that will lead the evolution of AI in 2019
The rise of Artificial Intelligence and Machine Learning is highly dependent on Big Data. The data
makes it possible to develop predictive models. The more data that are numerous, and representative of the
concepts to be learned, the more Machine Learning AI applications are completed.
In 2017, we should see more experts in this area, but demand should remain above supply. Machine
Learning promotes the adoption of Big Data solutions, just like the cloud that facilitates their deployment.
9. University of Burgundy
Master’s Degree DB & AI
Artificial Intelligence and Big Data
a revolutionary convergence
9
• Big Data self-service tools available on the web
With advances in data processing applications, there are many free online Big Data platforms
available. These cloud platforms make it easy to organize and synthesize data, even for beginners.
It is enough for the user to specify the amount of storage and computing power it needs, and the
databases appear in the cloud in minutes. No need to configure racks, networks or servers.
For Michael Cigarette, Director of Analytical Infrastructure at Ford Motor Company, this trend is
expected to continue in 2017. Big Data's cloud implementations are becoming increasingly popular as they
reduce the cost of accessing these technologies. For many, developing a Big Data stack is not cost effective, and
works best when most data can be hosted on an individual instance.
• Analytical technologies are struggling to adapt
Even with state-of-the-art tools and data warehouses such as Hadoop and Spark, data analytics remain
complex. Companies struggle to transfer their data from operational systems to analytical systems. This
difficulty directly affects productivity.
The data available is no longer numerous, and the algorithms are improving, allowing more automation
and better predictions. In fact, analytical technologies are struggling to adapt.
• Data cleaning becomes an industry
To transfer data to Machine Learning systems, it is necessary to clean them first. Cleaning up data means
looking for errors in the format or duplications within the database. The quality of Machine Learning systems
depends on the data on which they are based. The secret is to turn raw data into actionable data. For example,
knowing that someone has visited an online shoe store is helpful, but knowing when he or she visited is an
invaluable piece of information.
• The democratization of data
The data director of the Toyota Research Institute believes that data does not reside in data lakes but in
silos in which their mission is clear. Server-less and micro-service architectures make it much easier for owners
of these silos to access, analyze, and manage their data without having to rack up servers, configure virtual
machines, or even to the payment by the hour. Data owners can therefore focus on data enforcement and pay
for what they use by the minute.
Artificial intelligence is a sector with great potential to transform the fields of science, medicine and
technology. We can only advise companies to be ready to embrace this new technology.
The Big Data and AI are emerging technologies, and it is impossible to predict their effect on the long
term. It would be absurd, however, to ignore these technical advances. These two technologies are likely to
converge in the very near future.