This pdf contains info about federated learning.

1. FEDERATED
LEARNING
Vaishakh K P

2. Machine learning
Machine learning is an application of artificial intelligence (AI) that
provides systems the ability to automatically learn and improve from
experience without being explicitly programmed. Machine learning
focuses on the development of computer programs that can access
Data and use it learn for themselves.
• Supervised machine learning: The program is “trained” on a pre-
defined set of “training examples”, which then facilitate its ability to
reach an accurate conclusion when given new data.
• Unsupervised machine learning: The program is given a bunch of data
and must find patterns and relationships therein.

3. Traditional
Programming
Machine
Data
Algorithm
Result

4. Machine
Learning
Machine
Data
Desired
Result
Algorithm

5. DATA
MLdepends heavily on data, without data, it is
impossible for an “AI” to learn.
It is the most crucial aspect that makes
algorithm training possible…
No matter how great your AI team is or the
size of your data set, if yourdata setis not
good enough, your entireAI project will fail!

6. Some of the
terms used in
Machine
Learning
Model:The representation of what
an ML system has learned from the
training data.
data set or dataset :A collection of
data contains one or
more features and possibly a label
Training: The process of determining
the ideal parameters comprising a
model.

7. Where do they
get data?
User data is a focus area all its own. From
consumer behavior to predictive analytics,
Companies regularly capture, store and
analyze large amounts of data on their Users
base every day.
Some companies have even built an entire
business model around consumer data,
User data is big business

8. User data can be
collected in
three ways
By directly asking customers,
By indirectly tracking customers
By appending other sources of
customer data to your
own,"

9. Data privacy
• It is essential to have users’ data to develop the Smart
technologies.
• But as they gather and keep all this data, it becomes a
liability for these companies.
• If a person has pictures on their phone that they do not
want anyone else to see, then if Apple or Google
collects that picture, their employees could have access
and abuse the data.
• Even if these companies protect against its own
employees having access to the data, a privacy breach
could occur, and then hackers would have access to
people’s private data.

10. SOME DATA PRIVACY
ISSUES..
• Cambridge Analytica
• Facebook Reportedly Gave Tech Companies
Access to User Data Beyond Disclosure
• Google CEO To Face Congress Over Data Privacy
• An Amazon Echo device recorded a woman's
conversation in Portland and "shared it with one
of her husband's employees in Seattle,

11. FEDERATED
LEARNING
• A distributed machine learning
approaches which trains machine
learning models using decentralized
data residing on end devices such as
mobile phones
• To train a machine learning model,
traditional machine learning adopts a
centralized approach which requires
the training data to be aggregated on a
single machine or in a datacenter.

13. Cloud-centric ML for Mobile

14. Current Model
Parameters
The model lives in
the cloud.

15. trainin
g
data
We train models in
the cloud.

16. Make predictions in
the cloud.

17. trainin
g
data
But how do we continue
to improve themodel?

18. Local
Training
Data
We gather to the
cloud.

19. And make the model
better.

20. And make the model
better. (for
everyone)

21. Federated Learning

22. Mobile
Device
Local
Training
Data
Current Model
Parameters
Cloud
Service
Provider

23. Mobile
Device
Local
Training
Data
Many deviceswill be offline.
Current Model
Parameters
Cloud
Service
Provider

24. Mobile
Device
Local
Training
Data
Current Model
Parameters
Many deviceswill be offline.
1. Server selects
a sample of e.g.
100 online
eligible devices.

25. 2. Selected devices
download thecurrent
model parameters.

26. 3. Devices compute an
update using localtraining
data

27. ∑
4. Server aggregates
users' updates into a
new model.

28. ∑
Federated
Learning
5. Repeat until
convergence

29. To make the model
better. (for
everyone)

30. And
personalize
it, for every
one.

31. Training
and
Testing

32. How is Federated
Learning an
improvement to
AI?
Personal data never leaves the user’s device,
only updates made to the model are
transferred.
This data is encrypted making it impossible
for anyone to intercept the data and retro
engineer it
The updates are lighter than the original
users’ data.
The model is located in the user’s device,
allowing for real time inferences with no
latency problems

33. Is data secure?
• Secure Aggregation
enables the server to
combine the encrypted
results, and only decrypt
the aggregate

34. ADVANTAGES
Federated learning provides a privacy-preserving
mechanism
Less battery and band width consumption
Model personalization
Cost saving
Improved accuracy

35. Conclusion
Federated learning and
analytics are new fields with
established roots and
tremendous room to grow.
And they allow us to test and
train all kinde of devices- not
just phones and tabets!
Training from self driving
cars on aggregated real-
world driver behavior.
Helping hospitals to improve
diagnostics while
maintaining patient privacy

36. Any Questions?
“We can learn from every one
without learning about any one.”