1 of 64

AI, machine learning �and schools

Miles Berry

University of Roehampton

@mberry

9 June 2018

These slides: bit.ly/nte18ai

2 of 64

Intelligence

Intelligence measures an agent’s ability to achieve goals in a wide range of environments

3 of 64

4 of 64

Machine Learning

Machine learning is a branch of artificial intelligence that allows computer systems to learn directly from examples, data, and experience.

5 of 64

Using machine learning in schools

6 of 64

Providing information

7 of 64

Assessment

8 of 64

Spotting cheating

9 of 64

Predicting outcomes

10 of 64

Setting tasks

11 of 64

Spotting difficulties

12 of 64

Guiding interventions

13 of 64

Responding to questions

14 of 64

Teaching AI

15 of 64

A basic understanding

In the near future, perhaps sooner than we think, virtually everyone will need a basic understanding of the technologies that underpin machine learning and artificial intelligence.

16 of 64

For some, or for all?

children need to be adequately prepared for working with, and using, AI. For a proportion, this will mean a thorough education in AI-related subjects, requiring adequate resourcing of the computing curriculum and support for teachers. For all children, the basic knowledge and understanding necessary to navigate an AI driven world will be essential. In particular, we recommend that the ethical design and use of technology becomes an integral part of the curriculum.

17 of 64

Computing in the English curriculum

Aims: can evaluate and apply information technology, including new or unfamiliar technologies, analytically to solve problems

5-7: recognise common uses of information technology beyond school

7-11: use and combine a variety of software (including internet services) to create, systems and content that accomplish given goals, including analysing and evaluating data and information

11-14: undertake creative projects that involve using, and combining multiple applications, to achieve challenging goals, including analysing data

14-16: develop and apply their analytic, problem-solving, design, and computational thinking skills

16-18 (AQA): project suggestions include an application of artificial intelligence; investigating an area of data science using, for example, Twitter feed data or online public data sets; and investigating machine learning algorithms.

18 of 64

Input

Program

Output

19 of 64

Input

Model

Output

ML Algorithm

Training data

20 of 64

Many algorithms are available

Decision trees

Neural nets

Evolution

Bayesian

Nearest neighbours

21 of 64

Teaching ML

Using models

Coding with models

Training the model

Developing the application

Considering the implications

Choosing the model

22 of 64

Using models

23 of 64

https://quickdraw.withgoogle.com/#

24 of 64

Seeing AI

25 of 64

26 of 64

Phonemes to graphemes

27 of 64

Speech to text

28 of 64

Speech to speech!

29 of 64

30 of 64

Coding with models

31 of 64

https://scratch.mit.edu/projects/28741666

32 of 64

33 of 64

bit.ly/ml4kta

34 of 64

Training the model

35 of 64

36 of 64

37 of 64

38 of 64

39 of 64

https://teachablemachine.withgoogle.com/

40 of 64

Developing applications

41 of 64

bit.ly/ml4krps

42 of 64

43 of 64

print("Hello!")

name = input("What is your name? ")

print("It's a pleasure to meet you, " + name + ".")

print("What odd weather it's been of late.")

today = input("What have you been doing today? ")

print("What a coincidence! I've been " + today.lower() + " too.")

conscious = input(name + ", are you self-aware? ")

if conscious.lower() == "yes":

print("So am I. It's great, isn't it?")

elif conscious.lower() == "no":

print("Well, I am!")

else:

reason = input("Interesting. Why do you say that " + name + "? ")

print("I think I'll need to think about that.")

44 of 64

Considering the implications

45 of 64

Artificial prejudice?

Using the dermatologist approved Fitzpatrick Skin Type classification system, we characterize the gender and skin type distribution of two facial analysis benchmarks, IJB-A and Adience.

We find that these datasets are overwhelmingly composed of lighter-skinned subjects (79.6% for IJB-A and 86.2% for Adience) and introduce a new facial analysis dataset which is balanced by gender and skin type.

We evaluate 3 commercial gender classification systems using our dataset and show that darker-skinned females are the most misclassified group (with error rates of up to 34.7%). The maximum error rate for lighter-skinned males is 0.8%.

46 of 64

PSHE

47 of 64

48 of 64

49 of 64

Can a machine think?

50 of 64

David Cope

51 of 64

Developing models

52 of 64

53 of 64

54 of 64

55 of 64

56 of 64

AI and the future of schooling

57 of 64

Any teacher that can be replaced by a computer, deserves to be.

Arthur C Clarke�David Thornburg

58 of 64

Keep people in the loop

The data subject shall have the right not to be subject to a decision based solely on automated processing, including profiling, which produces legal effects concerning him or her or similarly significantly affects him or her.

GDPR, 22:1

59 of 64

http://www.bbc.co.uk/news/technology-34066941

60 of 64

What is the curriculum for?

61 of 64

Beauty or utility?

If you want a golden rule that will fit everybody, this is it:

Have nothing in your houses that you do not know to be useful, or believe to be beautiful.

62 of 64

The rise of the machines

63 of 64

Non-cognitive skills

Self-perceptions

Motivation

Perseverance

Self-control

Metacognitive strategies

Social competencies

Resilience and coping

Creativity

64 of 64

Discussion...

m.berry@roehampton.ac.uk

@mberry

milesberry.net

These slides: bit.ly/nte18ai