The document is a presentation on artificial intelligence (AI) and data stream mining by Albert Bifet, covering various aspects such as the definition of AI, its challenges, and machine learning for data streams. It discusses the importance of real-time analytics, characteristics of AI systems, and the need for adaptive models that can learn from changing data streams. Additionally, it introduces open-source tools like MOA and SAMOA for online learning from data streams and highlights the significance of the Internet of Things (IoT).

1. TITRE DE LA PRÉSENTATION - MENU « INSERTION / EN-TÊTE
ET PIED DE PAGE »
30/01
/2018
!2
ARTIFICIAL INTELLIGENCE AND
DATA STREAM MINING
ALBERT BIFET
15 FEVRIER 2018
#Futur&Ruptures

2. TITRE DE LA PRÉSENTATION - MENU « INSERTION / EN-TÊTE
ET PIED DE PAGE »
30/01
/2018
!3
1. ARTIFICIAL INTELLIGENCE
2. ARTIFICIAL INTELLIGENCE
CHALLENGES
3. MACHINE LEARNING FOR DATA
STREAMS
4. OPEN SOURCE TOOLS
5. SUMMARY
SOMMAIRE

3. TITRE DE LA PRÉSENTATION - MENU « INSERTION / EN-TÊTE
ET PIED DE PAGE »
30/01
/2018
!4
ARTIFICIAL
INTELLIGENCE

4. • Big Data
• Internet of Things
• Data Science
• Artificial Intelligence
Artificial Intelligence is the new Electricity
real time analytics
2

5. What is AI?
• Artificial intelligence
(AI) is an area of
computer science
that emphasizes the
creation of intelligent
machines”.
9

6. TITRE DE LA PRÉSENTATION - MENU « INSERTION / EN-TÊTE
ET PIED DE PAGE »
30/01
/2018
!5
ARTIFICIAL
INTELLIGENCE
CHALLENGES

7. Artificial Intelligence
3

8. Artificial Intelligence
5

9. European AI
3

10. Big Data
• GAFAM: Google, Apple, Facebook, Amazon,
Microsoft
• Personal Information
• Google, Facebook, Twitter, Linkedin,..
• All personal communications in Europe are
managed by non-European companies
3

11. 4

12. AI Systems
• According to Nikola Kasabov, AI systems should exhibit the
following characteristics:
• Accommodate new problem solving rules incrementally
• Adapt online and in real time
• Are able to analyze itself in terms of behavior, error and
success.
• Learn and improve through interaction with the environment
(embodiment)
• Learn quickly from large amounts of data (Big Data)
• Have memory-based exemplar storage and retrieval capacities
• Have parameters to represent short and long term memory,
age, forgetting, etc.
2

13. AI Challenges
• La vision de la France doit donc consister à
développer simultanément une IA plus verte et
une IA au service de la transition écologique.
5

14. TITRE DE LA PRÉSENTATION - MENU « INSERTION / EN-TÊTE
ET PIED DE PAGE »
30/01
/2018
!6
MACHINE LEARNING
FOR DATA STREAMS

15. Machine Learning
• Machine learning is a type of
artificial intelligence (AI) that
provides computers with the
ability to learn without being
explicitly programmed.
• Machine learning focuses on
the development of computer
programs that can teach
themselves to grow and
change when exposed to new
data.
3

16. Analytic Standard Approach
Finite training sets
Static models
11
Data Set
Model
Classifier Algorithm
builds Model

17. Data Stream Approach
Infinite training sets
Dynamic models
12
D
M
Update
Model
D
M
D
M
D
M
D
M
D
M
D
M
D
M
D
M
D
M
D
M
D
M

18. Importance$of$O
• As$spam$trends$change
retrain$the$model$with
Pain Points
• Need to retrain!
• Things change over time
• How often?
• Data unused until next
update!
• Value of data wasted
13

19. IoT Stream Mining
• Maintain models online
• Incorporate data on the fly
• Unbounded training sets
• Resource efficient
• Detect changes and adapts
• Dynamic models
14

20. Approximation Algorithms
• General idea, good for streaming algorithms
• Small error ε with high probability 1-δ
• True hypothesis H, and learned hypothesis Ĥ
• Pr[ |H - Ĥ| < ε|H| ] > 1-δ
15

21. Hoeffding Adaptive Tree
• Replace frequency counters by estimators
• No need for window of instances
• Sufficient statistics kept by estimators separately
• Parameter-free change detector + estimator with
theoretical guarantees for subtree swap (ADWIN)
• Keeps sliding window consistent with
“no-change hypothesis”
55
A. Bifet, R. Gavaldà: “Adaptive Parameter-free Learning from Evolving Data Streams” IDA (2009)
A. Bifet, R. Gavaldà: “Learning from Time-Changing Data with Adaptive Windowing”. SDM ‘07

22. ADWIN
56

23. ADWIN
57

24. ADWIN
58

25. ADWIN
59

26. Adaptive Random Forest
• Why Random Forests?
• Off-the-shelf learner
• Good learning performance
Adaptive random forests for evolving data stream
classification.
Gomes, H M; Bifet, A; Read, J; Barddal, J P; Enembreck, F;
Pfharinger, B; Holmes, G; Abdessalem, T.
Machine Learning, Springer, 2017.
• Based on the original Random Forest by Breiman
60

27. TITRE DE LA PRÉSENTATION - MENU « INSERTION / EN-TÊTE
ET PIED DE PAGE »
30/01
/2018
!7
OPEN SOURCE TOOLS

28. MOA
• {M}assive {O}nline {A}nalysis is a framework for online learning
from data streams.
• It is closely related to WEKA
• It includes a collection of offline and online as well as tools for
evaluation:
• classification, regression
• clustering, frequent pattern mining
• Easy to extend, design and run experiments
{M}assive {O}nline {A}
MOA (Bifet et al. 20
{M}assive {O}nline {A}nalysis is a framework
learning from data streams.
It is closely related to WEKA

29. APACHE SAMOA
8
http://samoa-project.net
Data
Mining
Distributed
Batch
Hadoop
Mahout
Stream
Storm, S4,
Samza
SAMOA
Non
Distributed
Batch
R,
WEKA,…
Stream
MOA
G. De Francisci Morales, A. Bifet: “SAMOA: Scalable Advanced Massive Online Analysis”. JMLR (2014)

30. SAMOA ARCHITECTURE
An adapter for integrating Apache Flink into Apache SAMOA was implemente
n scope of this master thesis, with the main parts of its implementation bein
addressed in this section. With the use of our adapter, ML algorithms can b
executed on top of Apache Flink. The implemented adapter will be used for th
evaluation of the ML pipelines and HT algorithm variations.
Figure 20: Apache SAMOA’s high level architecture.

31. StreamDM
10
http://huawei-noah/github.io/streamDM

32. scikit-multiflow
14

33. TITRE DE LA PRÉSENTATION - MENU « INSERTION / EN-TÊTE
ET PIED DE PAGE »
30/01
/2018
!8
SUMMARY

34. INTERNET OF THINGS
IoT: sensors and actuators connected by networks to
computing systems.
• Gartner predicts 20.8 billion IoT devices by 2020.
• IDC projects 32 billion IoT devices by 2020

35. IoT versus Big Data
10

36. Applications IoT Analytics
9

37. IOT AND INDUSTRY 4.0
Interoperability: IoT
Information transparency: virtual copy of the physical
world
Technical assistance: support human decisions
Decentralized decisions: make decisions on their own

38. 6

39. Data, Intelligence and Graphs
(DIG)
3

40. Data, Intelligence and Graphs
(DIG)
4

41. 4

42. 5

43. Thanks!
60
@abifet

44. TITRE DE LA PRÉSENTATION - MENU « INSERTION / EN-TÊTE
ET PIED DE PAGE »
30/01
/2018
!2
ARTIFICIAL INTELLIGENCE AND
DATA STREAM MINING
ALBERT BIFET
15 FEVRIER 2018
#Futur&Ruptures