Scribd
Upload
9 views

Chapter1 Introduction To Machine Learning Summary Final

Machine learning works similarly to how humans learn by recognizing patterns from experiences over time. For example, when tasting apples we build experiences that allow us to recognize good apples from bad ones. Machine learning can be applied across many industries, such as healthcare where models can predict protein structure from amino acids. It enables tasks like data analysis, automation, and handling large amounts of data in faster and more efficient ways than previously possible. To develop a machine learning model, data is fed into the system and transformed, mathematical algorithms are applied, and outputs are computed using input data and parameters. The model is trained by minimizing differences between predicted and expected outputs over iterations to learn parameters. Performance is evaluated on test data not used for training to assess the model

Uploaded by

Zubair Najim
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
9 views

Chapter1 Introduction To Machine Learning Summary Final

Machine learning works similarly to how humans learn by recognizing patterns from experiences over time. For example, when tasting apples we build experiences that allow us to recognize good apples from bad ones. Machine learning can be applied across many industries, such as healthcare where models can predict protein structure from amino acids. It enables tasks like data analysis, automation, and handling large amounts of data in faster and more efficient ways than previously possible. To develop a machine learning model, data is fed into the system and transformed, mathematical algorithms are applied, and outputs are computed using input data and parameters. The model is trained by minimizing differences between predicted and expected outputs over iterations to learn parameters. Performance is evaluated on test data not used for training to assess the model

Uploaded by

Zubair Najim
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 8

Machine learning is similar to human

processing

• Over time with experience, we learn


to recognize different patterns.
• E.g., recognizing good apples from bad
apples. As we taste different apples, we
build experiences (data) of what we like
and gather these experiences in our brain
and create a model. This model allows us
to recognize good apples from bad ones.

Image source: LinkedIn


1
Machine learning is a really powerful tool
for many industries

For example, in the healthcare industry,


ML models can be used to predict protein
structure based on amino acids. This is
something that we can use in the future
not only for for detecting diseases, but it is
a good starting point for transfer learning
for all other questions that are similar in this
area.

Image source: DeepMind


2
What makes machine learning so exiting?

Machine learning is a way to Machine learning enables more


explore data and make predictions automation and better decision
that was not possible in the past. making.

Machine learning is versatile and Machine learning enables fast


can be applied to a variety of and efficient problem solving.
industries.

3
The practical advantages of machine
learning

More data analysis in less time

Automation of repetitive tasks

Improved accuracy & efficiency

Handling of large amounts of data

Faster adaptation to data changes

4
Defining a machine learning pipeline

1 2 3 4

Transforming data Performing a series Computing an


Feeding the system in a form that is of mathematical output with input
with raw data understandable steps data using different
by a computer (ML algorithm) parameters

The system = Machine learning model

5
Training process

How does the ML training process work?


The machine tries to minimize the difference between
the predicted and the expected output, the so-called
loss, by adjusting the learned parameters over a
number of iterations. This entire process is called
training.

6
For a well-functioning ML model an
appropriate dataset is a must

We split the dataset into two categories

Training set Test set

Used for training Used for evaluating the


the model performance of the model

7
Evaluating the performance of a
trained ML model

We can use evaluation metrics to


differentiate a good model from a bad one
The performance of a machine learning model
indicates how accurately it predicts the expected
output:
Evaluation metrics are calculated using the
predicted and actual values and indicate whether
the ML model learned the relations between
different features in input.

You might also like