RECENT ADVANCES IN

ARTIFICIAL INTELLIGENCE
 Ice Breaker
Can you guess if the image is a real human or A.I. generated?

page 2
Image #1

page 3
Image #2

page 4
Image #3

page 5
Image #4

page 6
Image #5

page 7
Image #6

page 8
Image #7

page 9
Image #8

page 10
Image #9

page 11
Image #10

page 12
 Results?
thispersondoesnotexist.com

page 13
Generating fake images
with StyleGAN
• products of a state-of-the-art
generative adversarial network (GAN)
• GANs perform adversarial training by
generating images based from sample
images
• Involves two neural networks (generator
and discriminator)
• released by NVIDIA
• what would be the consequences of this
AI ability?

page 14
Generating fake images with StyleGAN

http://www.openculture.com/2018/12/artificial-intelligence-creates-realistic-photos-of-people-none-of-whom-actually-exist.html
page 15
Generating fake images with StyleGAN

https://qz.com/1115353/new-research-from-nvidia-shows-that-the-era-of-easily-faked-ai-generated-photos-is-quickly-emerging/
page 16
Artificial Intelligence (A.I.)
• How machines exhibit intelligence. In contrast to the natural intelligence
displayed by humans and other animals.
• “Any device that perceives its environment and takes actions that maximize
its chance of successfully achieving its goals.” (Poole, Mackworth, and
Goble, 1998).
• “A system’s ability to correctly interpret external data, to learn from such
data, and to use those learnings to achieve specific goals and tasks through
flexible adaptation.” (Kaplan and Heinlein, 2018)
• AI is used to describe the cognitive functions of the human brain such as
learning and problem solving. (Russel and Norvig, 2009)

page 17
Artificial Intelligence (A.I.) in simpler words
A machine exhibits artificial intelligence if:
1. It can assess the environment
2. It can perform decision-making or
learning.
3. Use what it learned/reasons to achieve its
goal.

page 18
Terminator cyborg from The Terminator (1984)

page 19
Metalhead dogs from Black Mirror Season 3 (2017)

page 20
When there is AI, there is algorithm
• Algorithms are the backbone of any AI-integrated machines.
• An algorithm is a set of unambiguous instructions that a mechanical
computer can execute.
• Complex algorithms are built on top of simple ones.

page 21
Tic-Tac-Toe algorithm
1. if someone has a "threat" (that is, two in a row),
take the remaining square.
2. if a move "forks" to create two threats at once,
play that move.
3. take the center square if it is free.
4. if your opponent has played in a corner, take
the opposite corner.
5. take an empty corner if one exists.
6. take any empty square.

page 22
AI innovations in various fields
1. Healthcare
2. Agriculture
3. Education
4. Games
5. Navigation
6. Communication

page 23
AI Innovations in Healthcare

page 24
Beating Diabetic Blindness with AI
• Instant diagnosis with an eye scanner
without the help of a physician
• Behind the new screening methods
are neural networks, complex
mathematical systems that can learn
tasks by analyzing vast amounts of
data.
• By analyzing millions of retinal scans
showing signs of diabetic blindness,
a neural network can learn to identify
the condition on its own.
• Practiced in hospitals in India,
partnered with Google AI
https://www.nytimes.com/2019/03/10/technology/artificial-intelligence-eye-hospital-india.html
page 25
Beating Diabetic Blindness with AI
• Google’s eye system is certified
for use in Europe but awaiting
approval in the United States.
• The system performs on a par
with trained ophthalmologists
but doctors can sometimes
make a diagnosis when faced
with cataracts and blurry eye
scans.
• The Google system still struggles
to do this.

https://www.nytimes.com/2019/03/10/technology/artificial-intelligence-eye-hospital-india.html
page 26
Google AI on Healthcare

https://www.nytimes.com/2019/03/10/technology/artificial-intelligence-eye-hospital-india.html
page 27
Detecting Skin Cancer with AI
• The small study, published this month in
the journal Annals of Oncology, compared
the performance of a form of AI called
deep learning convolutional neural
network (CNN) to a group of 58
dermatologists.
• Researchers found that the CNN was
more reliable at identifying melanoma, an
aggressive form of skin cancer, and benign
moles (called nevi) than even experienced
dermatologists.

https://www.healthline.com/health-news/ai-may-be-better-at-detecting-skin-cancer-than-your-derm#1
page 28
Detecting Skin Cancer with AI
• During the first portion of the study, the dermatologists identified 86.6 percent of
melanomas. On average, the group accurately identified 71.3 percent of the benign
moles.
• When artificial intelligence was put to the same task, it correctly identified 95 percent
of melanomas.
• During the subsequent stage of the study, when doctors were given additional
information, their accuracy improved. Average detection of melanomas increased to
88.9 percent, and benign moles to 75.7 percent.
• But, even with that additional information, the performance of human
dermatologists was still worse than the CNN.

https://www.healthline.com/health-news/ai-may-be-better-at-detecting-skin-cancer-than-your-derm#1
page 29
AI Innovations in Agriculture
page 30
Using AI for Automatic Strawberry Harvesting
• A shrinking in labor force in strawberry picking is seen in Wish Farms,
Florida.
• The task is back-breaking labor as the pickers are required to fold their
bodies to lean over the plants. Their hands skim through the leaves to
find the bright ripe strawberries. Workers should make sure not to tear
the plant.
• 600 workers are needed to harvest 600 acres every two to three days.
• Five years later, Harvest CROO Robotics has a robotic picker which is
able to look at a strawberry plant and determine which among the berries
on the plant is ripe, all through artificial intelligence.

https://www.cbsnews.com/news/artificial-intelligence-could-revolutionize-farming-industry/
page 31
Using AI for Automatic Strawberry Harvesting

It knows all this, thanks to super-accurate GPS. Its computer brain contains a map showing the exact locations of every
strawberry plant in the field. When it gets to a strawberry plant, bright lights flash; cameras spin in a circle.
https://www.cbsnews.com/news/artificial-intelligence-could-revolutionize-farming-industry/
page 32
Crop Detection and Diagnosis with Microsoft AI

https://www.nytimes.com/2019/03/10/technology/artificial-intelligence-eye-hospital-india.html
page 33
Bringing AI to the Sea

https://www.youtube.com/watch?v=aFx-Zb_oVrk
page 34
AI Innovations in Education
page 35
Attendance Monitoring with Facial
Recognition in China
• A Chinese high school in Hangzhou is using facial-recognition technology
that scans students every 30 seconds.
• The "intelligent classroom behavior management system," scans the
room logging both the behavior of the students and their facial
expressions.
• The system can identify seven moods, including happy, sad, afraid and
angry, by simply analyzing a student's face.
• A camera, perched atop the blackboard at the front of the classroom,
also tracks six types of behavior: reading, writing, hand raising, standing
up, listening to the teacher, and leaning on the desk.
https://www.healthline.com/health-news/ai-may-be-better-at-detecting-skin-cancer-than-your-derm#1
page 36
Attendance Monitoring with Facial Recognition in China

https://www.youtube.com/watch?v=k39oR2fLFPg&feature=youtu.be
page 37
Transforming Classrooms with AI and
Augmented Reality

https://www.youtube.com/watch?v=WPOvMIvvZi8
page 38
AI Innovations in Games

page 39
Mastering Super Mario with Reinforcement
Learning

https://www.youtube.com/watch?v=1qa7oRhZvbM
page 40
Beating the Best World’s Dota 2 Players with AI

https://www.youtube.com/watch?v=l92J1UvHf6M&feature=youtu.be
page 41
AI Video Games that Exploit your Fears
• Stealth horror PC game Hello Neighbor takes that spying, meddling
impulse to the extreme. Also extreme is your enemy, just a friendly
neighborhood AI who learns from your every move.
• In Hello Neighbor, an advanced AI will come after you if your snooping
catches his eye.Your mistakes will teach it how to best hunt you down.
• In Nevermind, the rules are simple: find the exit, and don't be afraid. You
shouldn't be afraid because the game knows when you are, and the more
frightened you become, the harder it gets.
• The game tracks your heart rate via a real-sense camera or heart monitor
to see how scared you are. The faster your heart is pumping, the more
frightened it thinks you are, the more difficult Nevermind becomes.
https://kotaku.com/a-horror-game-where-the-ai-learns-from-your-every-move-1787241885
page 42
AI Video Games that Exploit your Fears

page 43
AI Innovations in Navigation

page 44
Map Navigation with Augmented Reality
• Thanks to advances in AI, Google Maps will soon create Street View-style
visual guides for step-by-step directions overlaid onto the real world, as
viewed through the smartphone camera.
• The company plans to integrate its Assistant, equipped with the
computer vision platform Google Lens, into Maps. That way, you’ll be
able to pan over a city street and see pop-ups highlighting restaurants
and other locations in real time.

https://www.theverge.com/2018/5/10/17340004/google-ai-maps-news-secret-weapon-remaking-old-apps-products-io-2018
page 45
Map Navigation with Augmented Reality

https://www.youtube.com/watch?v=4F0gFpzsYLM
page 46
Drone Hunting with AI
• DroneHunter hunts for unauthorized drones and captures them. It's a
fully-autonomous air defense mechanism.
• Stray drones can be specifically targeted. Or DroneHunter can learn to
find rogue drones itself.
• DroneHunter captures the rogue drones with a net. Forcing it to land on
the ground.
• The goal is to stop threats at a safe distance from protected locations. It
is able to detect, monitor, and autonomously capture drones in real time
using object detection in AI.

https://www.youtube.com/watch?v=PpgFNSpANFM
page 47
Drone Hunting with AI

https://www.youtube.com/watch?v=PpgFNSpANFM
page 48
AI Innovations in Communication
page 49
Rise of ASR-Based Assistants
• Google Assistant has made incredible progress
since its 2016 launch and is probably the most
advanced and dynamic of the assistants out there.
• The "OK Google" or "Hey, Google" side covers voice
commands, voice searching, and voice-activated
device control, letting you do things like send
messages, check appointments and so on on your
Android device.
• In 2017, Google Assistant was installed on more
than 400 million devices.

https://www.youtube.com/watch?v=PpgFNSpANFM
page 50
Rise of ASR-Based Assistants

https://www.youtube.com/watch?v=FPfQMVf4vwQ&feature=youtu.be
page 51
Chatbot communicating with another Chatbot

https://www.youtube.com/watch?v=WnzlbyTZsQY
page 52
 Jobs threatened by AI , professions at risk
• Martin Ford, futurist and author of Rise of the
Robots: Technology and the Threat of a
Jobless Future, explains the jobs that are most
at risk are those which “are on some level
routine, repetitive and predictable”
• The jobs themselves won’t entirely vanish;
rather, they will be redefined.

https://www.theguardian.com/us-news/2017/jun/26/jobs-future-automation-robots-skills-creative-health

page 53
 Jobs safe from AI automation
• Jobs that involve “genuine creativity, such as
being an artist, being a scientist, developing a
new business strategy”.
• Jobs that involve building complex
relationships with people: nurses, for
example, or a business role that requires you
to build close relationships with clients.
• Jobs that are highly unpredictable – for
example, if you’re a plumber who is called out
to emergencies in different locations.

https://www.theguardian.com/us-news/2017/jun/26/jobs-future-automation-robots-skills-creative-health

page 54
What’s next in AI?
• We are just at the tip of the iceberg. We barely scratched the surface.
• Even with these ground-breaking innovations, we are nowhere near
harnessing the true potential of our envisioned artificial intelligence. As
of now, we are still at the stage of Artificial Narrow Intelligence or Weak
AI.
• The ultimate goal is to achieve Artificial General Intelligence (AGI)
where AI machines can solve other problems and exhibit general
intelligence - as a human would.

page 55