POSTER Velibor Božić

Machine Learning vs Deep learning

Machine learning
Machine learning (ML) is a type of artificial intelligence (AI) that allows computers to learn without being explicitly programmed. It's like teaching a
child by showing them examples and letting them figure out the rules themselves.

Machine learning algorithms are trained on data, which can be anything from text and images to numbers and sound. The algorithms learn to
identify patterns in the data, and then use those patterns to make predictions about new data. For example, a machine learning algorithm could be
trained on a dataset of emails, and then used to identify spam emails.

There are many different types of machine learning algorithms, each with its own strengths and weaknesses. Some common types of machine
learning algorithms include:

• Supervised learning: This is the most common type of machine learning, where the algorithm is trained on labelled data. Labelled data
means that each piece of data has a label associated with it, which tells the algorithm what the correct output should be. For example, an
algorithm that is learning to classify images of cats and dogs would be trained on a dataset of images that are labelled as either "cat" or
"dog."
• Unsupervised learning: This type of machine learning is used when the data is not labelled. The algorithm is left to find its own patterns in
the data. Unsupervised learning can be used for tasks such as dimensionality reduction and clustering.
• Reinforcement learning: This type of machine learning is used when the algorithm learns by trial and error. The algorithm is given a reward
or penalty for its actions, and it learns to take actions that maximize its reward. Reinforcement learning is often used in games, where the
algorithm learns to play the game by trial and error.

Machine learning is a powerful tool that can be used for a wide variety of tasks. It is being used in industries such as healthcare, finance, and
manufacturing to automate tasks, improve decision-making, and develop new products and services.

Deep learning
Deep learning is a specific type of machine learning inspired by the structure and function of the human brain. It utilizes artificial neural networks
with multiple layers, often referred to as deep neural networks, to process information and learn from data.

Here's a breakdown of key aspects of deep learning:

• Multi-layered processing: Unlike simpler machine learning models, deep learning algorithms involve data passing through several layers
of interconnected nodes, progressively extracting increasingly complex features from the input data. Imagine recognizing an object in an
image. Lower layers might detect edges and shapes, while higher layers combine these features to identify the object, like a car or a face.
• Representation learning: Deep learning models excel at automatically learning representations of the data they are processing. This
eliminates the need for manual feature extraction, which can be a time-consuming and challenging task in traditional machine learning
approaches.
• Applications: Deep learning has revolutionized various fields due to its ability to handle complex data like images, text, and audio. Some
prominent applications include:
o Image recognition: Deep learning models power facial recognition systems, self-driving cars, and image classification for medical
diagnosis.
o Natural language processing: Chatbots, machine translation, and sentiment analysis heavily rely on deep learning techniques.
o Speech recognition: Virtual assistants like Siri and Alexa use deep learning to understand and respond to spoken language.

While deep learning offers significant advantages, it also comes with challenges:

• Computational cost: Training deep learning models often requires significant computing power and large amounts of data.
• Data dependency: The performance of deep learning models heavily relies on the quality and quantity of data used for training.
• Interpretability: Understanding how deep learning models arrive at their predictions can be complex, making it challenging to explain their
decisions in certain scenarios.

Overall, deep learning represents a powerful tool within the machine learning landscape, constantly evolving and pushing the boundaries of what
AI can achieve.

Sources
o "Deep Learning" by Ian Goodfellow, Yoshua Bengio, and Aaron Courville: Often referred to as the deep learning bible, this book
provides both theoretical knowledge and practical insights. Suitable for those who enjoy a structured and academic approach.
o "Hands-On Machine Learning with Scikit-Learn, Keras & TensorFlow" by Aurélien Géron: A practical guide that gets you started with
building and training deep learning models using popular libraries like TensorFlow and Keras.
o Deep Learning Specialization by deeplearning.ai on Coursera: This comprehensive specialization offered by deeplearning.ai and
taught by Andrew Ng covers the fundamentals of deep learning and its applications in various domains.
o Fast.ai Practical Deep Learning for Coders by Jeremy Howard and Rachel Thomas: A project-based course that focuses on teaching
the practical aspects of deep learning using real-world coding examples.
o The TensorFlow Blog: Stay updated on the latest developments and tutorials related to TensorFlow, a popular deep learning
framework.