Machine Learning or Pattern Recognition

Machine learning and pattern recognition are related fields but have some differences.

Machine learning refers to the development of algorithms that allow computers to learn and make predictions or decisions without being explicitly programmed. It focuses on training models on data to recognize patterns and make accurate predictions or decisions based on new data.

Pattern recognition, on the other hand, is the process of identifying and classifying patterns or regularities in data. It involves techniques and algorithms for recognizing and analyzing patterns in various forms, such as images, signals, or sequences.

In summary, machine learning is a broader field that encompasses pattern recognition as one of its components. Machine learning algorithms are used to recognize patterns and make predictions, while pattern recognition focuses specifically on identifying patterns in data.

Pattern Recognition has its origin in engineering whereas ML grew out of computer science

Artificial Intelligence

• Deduction: from general rules to specific instances → Reasoning (= logic = implications)
• Induction: from examples to general rules → Machine Learning (= probabilistic)

What is Machine Learning?

• Machine: clearly
• Learning: Learn from experience E with respect to a task T and performance measure P if its performance at task T, as measured by P, improves with experience E

Types of learning

• Supervised learning: given a set of pair $[x,y]$, rule by $y=f(x)$ → goal: find an approximate function $\hat{f}(x)$
• Unsupervised learning: given a set of data $[x]$ → goal: find a function $f(x)$
• Reinforcement learning: given a set of triplets $[x,a,r]$ ruled by $r = f(a|x)$ → goal: find a mapping $x \rightarrow a$ such that $r$ is maximum

Pattern recognition system

ENV → Sensing → Feature Extraction → Decision → OUTPUT (action)

Env = take a picture with face

Sensing = camera → data

(Feature Selection - - ->) Feature Extraction = picture/contour

(Model Selection - - -> ) Decision = face or not face

Output = bounding box around the face

Traditional approaches:

Feature Selection / Model Selection = Prior knowledge

Decision = Machine Learning