Notesale: Turn your study into money

Document Preview

Extracts from the notes are below, to see the PDF you'll receive please use the links above

---

Artificial Intelligence & Machine Learning
Artificial Intelligence (AI) deals with making machines that are capable of
doing tasks that humans usually handle
...
AI allows machines to review data, identify repeated
trends and get better with time without explicit human directions
...
Before AI was called Artificial Intelligence,
Alan Turing introduced the Turing Test and John McCarthy is known for
labeling the field with this term
...


Types of AI
• Narrow AI (Weak AI)
Weak AI is another term for Narrow AI which are systems only built to do
specific jobs and lack consciousness and general intelligence
...
While they handle their
specific tasks very well, they cannot translate their knowledge to other
tasks
...
General AI
is unique because it can think, reason and adapt as a human, letting it
manage all sorts of intellectual tasks people can do
...
General AI is meant to imitate human
understanding, but a Superintelligence could well be smarter than
humans and be crowned the brightest mind on Earth
...

Basically, the key idea of AI research is to try to build systems that can think,
learn and change just as humans do
...
The main
goal here is to develop agents that can interact with the world flexibly, use
what they learn and respond to new circumstances
...
AI systems make repetitive tasks and tasks
using much data faster, more accurate and more repeatable
...

Since AI can handle large quantities of information and work without tiring, it
enables human workers to concentrate on developing and coming up with
strategies
...
In healthcare, AI speeds up the search for medicines and
makes medical treatments more personal and it also improves handling
climate issues and enhances worldwide logistics
...

Thus, AI can be used in material science, city planning and environmental
protection to find useful knowledge and enable new methods
...
It is hard for many
people to pull together and figure out the meaning behind all the information
that is available these days
...
By using
actionable data, individuals and companies can make smart options related
to improving marketing efforts, controlling money matters or forming good
public policies
...

Achieving General AI (AGI) which is also known as strong AI, is the major longterm goal for many scientists working with AI
...
AGI would allow AI to become
much more like a human by doing things such as reasoning, creating and
learning by themselves
...

Taken together, these objectives paint a clear picture of how AI is set to change
several industries, fix difficult problems and help humans in new and
impressive manners
...
During the 1940s and 1950s, scientists started focusing
on finding out if machines could show intelligence
...
It introduced an idea where machines could be tested
to determine if they could act intelligently, just like a human
...

The defining event of the early days was the Dartmouth Summer Research
Project on Artificial Intelligence which took place in 1956
...
Numerous scientists focused on mathematics,
psychology and computer science met to consider ways to develop intelligent
machines
...
People from this period were clearly
carried away by the speed of technological progress and the belief that soon
machines would strongly resemble human intelligence
...

Still, after some time, the intense interest and positive views about AI
gradually ended in disheartening times commonly called the “AI Winters”
between the 1970s and 1990s
...
There were
various causes behind this period of recession
...
At first, researchers faced difficulties in expanding AI systems
since they struggled with real-life difficulties, thinking like humans and huge
quantities of data
...

Also, computers and software in the 1980s were not advanced enough to
handle the complex ideas and data processing needs of newer AI models
...
As it was difficult to show real-world examples
of AI technology and because of the technological hurdle, both investors
stopped and started investing less in AI
...
As a result of less hopeful
expectations, those working in AI started to focus more on making practical
applications and on firming up theory which also slowed down the rate of
advancements
...

Key AI Technologies and Techniques
ML plays a central role in Artificial Intelligence today and introduced a new
way of creating computer programs
...
An algorithm is fed with lots of data so
that it can, without our direct help, spot patterns and irregularities in a
dataset
...
When you visit online shops, for example, the products promoted to
you are often chosen by a recommendation system that inspects your past
actions to find things you might prefer
...
The main thought is to
train computers to discover and use knowledge just like humans do, but at a
much higher and quicker level
...

The main component of Deep Learning is artificial neural networks which are
designed based on the nervous system in our brain
...
Every layer learns to identify more abstract
features from the data which helps the network reach a detailed
understanding
...
Contrary
to expectations, Deep Learning has shown fantastic results where people
believed AI would not be able to help
...

Because of these new advancements, DL is now used in driverless cars,
medical scans, voice support and state-of-the-art surveillance
...
Thanks to NLP,
machines can understand human language and respond to us like we would
expect
...
It is mainly thanks to Deep Learning
that recent progress in Natural Language Processing brought about the
creation of highly complex technology such as large language models (LLMs),
with ChatGPT being a prime example
...
This area of technology
makes chatbots, automated customer service, search engines and content
creation tools possible, changing the way we deal with both information and
technology
...

The area covers making computers capable of analyzing and interpreting
images and videos, just like people do visually
...
Due to deep convolutional neural networks, AI can
now understand hierarchical details such as lines, shapes and whole objects
from information presented by data pixels
...
Face recognition helps security
systems and is also used to unlock smartphones
...
What’s more, this technology supports autonomous vehicles
in detecting other things such as pedestrians and various road signs
...

AI is mainly powered by complicated algorithms and models that allow
machines to learn, decide on their own and connect to their environment
...
This indicates that whenever you input something, the
correct response is given to you
...
The data is
processed by the algorithm in pairs to spot the main patterns and links
...
Spam detection is an
illustration of supervised learning, since the AI is taught through emails that
are marked as either spam or not spam
...
Other fields where machine learning can be used are image
classification, diagnosis in medicine and estimating housing costs, depending
on having enough well-labeled training examples
...
AI in this case is expected to uncover
structures, patterns or connections in the data by itself
...
Such an approach proves valuable
when it is tricky to label samples, when collecting tags costs too much or the
main aim is to find new insights from the data
...
Because of
this, businesses are able to improve their marketing strategies and the
products they sell
...
By using
unsupervised learning, AI can help discover data that isn’t right away obvious
to human analysts

---