Scribd
Upload
16 views6 pages

Transfert Learning Techniques

Transfer learning is the process of using knowledge learned from one task to improve performance on another related task. It can be useful when limited data is available, as knowledge from a source task can help train models on a target task. There are three main types of transfer learning: instance, feature, and model-based. Transfer learning offers advantages like faster training, improved performance on small datasets, and ability to leverage knowledge across tasks.

Uploaded by

mahdi souki
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
16 views6 pages

Transfert Learning Techniques

Transfer learning is the process of using knowledge learned from one task to improve performance on another related task. It can be useful when limited data is available, as knowledge from a source task can help train models on a target task. There are three main types of transfer learning: instance, feature, and model-based. Transfer learning offers advantages like faster training, improved performance on small datasets, and ability to leverage knowledge across tasks.

Uploaded by

mahdi souki
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

MANEF BGUIGA

&
MAHDI SOUKI
INTRODUCTION

Transfer learning is the process of using knowledge


learned from one task to improve the performance of
a machine learning model on a different but related
task.
Transfer learning can be particularly useful in cases
where we have a limited amount of data available for
training a model on the target task. By using
knowledge learned from a source task with a larger
dataset, we can "transfer" this knowledge to the
target task, potentially improving model performance.

2
TYPES OF TRANSFER LEARNING
There are three main types of transfer learning: instance-based,
feature-based, and model-based.
Instance-based transfer learning involves transferring individual examples
from the source task to the target task. This can be useful when the
two tasks have a similar input space and the examples from the source
task are relevant to the target task.
Feature-based transfer learning involves transferring only the features
learned from the source task to the target task. This can be useful when
the two tasks have a different input space but share some common
features.
Model-based transfer learning involves transferring the entire trained
model from the source task to the target task. This can be useful when
the two tasks have a similar input and output space and the model
trained on the source task can be directly applied to the target task.

2
ADVANTAGES OF TRANSFER

LEARNING
Transfer learning can offer several advantages over training a model from
scratch on the target task:
Faster training times: Transfer learning allows us to use the knowledge
learned from the source task to initialize the model for the target task,
potentially reducing the amount of training data needed.
Improved performance on small datasets: In cases where we have a limited
amount of data available for the target task, transfer learning can help to
improve model performance by leveraging the knowledge learned from the
source task.
Ability to leverage the knowledge learned from one task to improve
performance on a related task: By transferring knowledge from a related
source task, we can potentially improve the performance of the model on
the target task.

2
APPLICATIONS OF TRANSFER

LEARNING
Transfer learning has been applied in a wide range of domains, including:
Computer vision: Transfer learning has been used to improve the
performance of image classification models, object detection models, and
more.
Natural language processing: Transfer learning has been used to improve
the performance of text classification models, language translation
models, and more.
Speech recognition: Transfer learning has been used to improve the
performance of speech-to-text models and other speech processing
tasks.
Some successful applications of transfer learning include using a pre-trained
model on the ImageNet dataset to improve the performance of an object
detection model on a different dataset, and using a pre-trained language
model to improve the performance of a text classification model.
2
REFERENCES

"Transfer Learning" by Pan and Yang (2010)


"A Survey of Transfer Learning" by Weiss et al. (2016)
"Deep Transfer Learning for Natural Language Processing" by Howard and Ruder (2018)

You might also like