Scribd
Upload
6 views

Introduction to Machine Learning

Machine learning is defined as computational methods that improve performance or make predictions based on experience, utilizing past data. It involves designing algorithms for various tasks such as classification, regression, and clustering, and can be categorized into supervised, unsupervised, and reinforcement learning scenarios. Key concepts include the learning process, generalization versus specialization, and the distinction between parametric and non-parametric models.

Uploaded by

ibk2007
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
6 views

Introduction to Machine Learning

Machine learning is defined as computational methods that improve performance or make predictions based on experience, utilizing past data. It involves designing algorithms for various tasks such as classification, regression, and clustering, and can be categorized into supervised, unsupervised, and reinforcement learning scenarios. Key concepts include the learning process, generalization versus specialization, and the distinction between parametric and non-parametric models.

Uploaded by

ibk2007
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 15

INTRODUCTION TO

MACHINE LEARNING
What is Machine Learning?
• Machine learning can be broadly defined as
computational methods using experience to improve
performance or to make accurate predictions.

• Learning by Human – a change in the cognitive structure


that occurs as a result of experience
• Humans acquire knowledge through experience

• Machines acquire knowledge through experience shared


in the form of past data [training samples].
What is Machine Learning?
• Example of a learning problem
• how to use a finite sample of randomly selected documents, each
labeled with a topic, to accurately predict the topic of unseen
documents.

• Machine learning consists of designing efficient and


accurate prediction algorithms.
Machine Learning Algorithms Vs. Traditional
Algorithms
• Traditional Algorithms

Input Logic- Output


code
• Machine Learning Algorithms

Input Output Logic-


model
Machine Learning Algorithms Vs. Traditional
Algorithms
• computer science
• measures of the quality of the algorithms are their time and space
complexity.
• Machine learning
• Additional quality – sample complexity - to evaluate the sample
size required for the algorithm to learn a family of concepts.

• ML - data-driven methods combining fundamental


concepts in computer science with ideas from statistics,
probability and optimization.
Examples of Learning Tasks
Prediction is the major task of ML
• Text: document classification, spam detection.
• Language: NLP tasks (e.g., morphological analysis, POS
tagging, context-free parsing, dependency parsing).
• Speech: recognition, synthesis, verification.
• Image: annotation, face/object detection, face/object
recognition, OCR, handwriting recognition.
• Computational biology: protein function prediction,
identification of key sites, or the analysis of gene and protein
networks.
• Unassisted control of vehicles (robots, car), Medical
diagnosis, fraud detection, network intrusion, Games (e.g.,
chess, backgammon, go).
Broad ML Tasks
• Classification: assign a category to each item (e.g., document
classification).

• Regression: predict a real value for each item (prediction of stock


values, economic variables).

• Ranking: order items according to some criterion (relevant web


pages returned by a search engine).

• Clustering: partition data into ‘homogenous’ regions (analysis of


very large data sets).

• Dimensionality reduction: find lower-dimensional manifold


preserving some properties of the data.
Learning stages
• Example: item, instance of the data used.

• Features: attributes associated to an item, often represented as a vector


(e.g., word counts).

• Labels: category (classification) or real value (regression) associated to


an item.

• Hyperparameters: Free parameters that are not determined by the


learning algorithm, but rather specied as inputs to the learning algorithm.

• Data:
• training data (typically labeled).
• test data (labeled but labels not seen).
• validation data (labeled, for tuning parameters).
Learning stages
• Loss function: A function that measures the difference, or
loss, between a predicted label and a true label.
• Misclassification loss
• Squared loss

• Hypothesis set: A set of functions mapping features


(feature vectors) to the set of labels Y.
Typical Learning Process
Learning Scenarios
• Unsupervised learning: no labeled data.

• Supervised learning: uses labeled data for prediction on


unseen points.

• Semi-supervised learning: uses labeled and unlabeled


data for prediction on unseen points.

• Transduction: uses labeled and unlabeled data for


prediction on seen points.
Learning Scenarios
• Online learning: involves multiple rounds where training
and testing phases are intermixed

• Reinforcement learning: The training and testing phases


are also intermixed

• Active learning: The learner adaptively or interactively


collects training examples, typically by querying an oracle
to request labels for new points.
Generalization vs. Specialization
Overfitting Vs. Underfitting
• The trade-off between the sample size and complexity
plays a critical role in generalization.

• Overfitting
• When the sample size is relatively small, choosing from a too
complex a family may lead to poor generalization.

• Underfitting
• with a too simple a family, it may not be possible to achieve a
sufficient accuracy.
Parametric Vs. Non-parametric model
• Assumptions can greatly simplify the learning process, but
can also limit what can be learned.
• Parametric
• A learning model that summarizes data with a set of parameters of
fixed size (independent of the number of training examples)
• The algorithms involve two steps:
• Select a form for the function.
• Learn the coefficients for the function from the training data.

• Non-parametric
• do not make strong assumptions about the form of the mapping
function
• good when you have a lot of data and no prior knowledge, and when
you don’t want to worry too much about choosing just the right
features.

You might also like