Talk at ECS, Southampton University, 9th February, 2009. I will talk about two different ways in which things that appear bad, may actually do you good. The first part is about BadIdeas, a technique used over a number of years to encourage technical creativity and innovation. As well as the technique itself, I will talk about some of the reasons why it works and about tentative steps to understand the nature of technical creativity and at a crude level model it. The second part will focus on regret, this time starting with a tentative model of regret and then seeing how this was used to inspire a computational model to speed up simple machine learning. However, analysing the results of the computational model in turn suggests insights into the nature of human regret. Creativity and regret are linked in that they both involve a rich interplay between analytic/rational thinking and more emotional/imaginative insight.

1. Southampton 9 Feb. 2009
Bad Things May Be Good for You:
creativity and regret
Alan Dix
Lancaster University
www.hcibook.com/alan/
www.alandix.com

2. Southampton 9 Feb. 2009
today I am not talking about …
• intelligent internet interfaces
fuzzy personal ontologies and
structure from folksonomies
• visualisation and sampling
• situated displays, eCampus,
small device – large display interactions
• fun and games, virtual crackers,
artistic performance, slow time
• physicality and product design
but have before

3. Southampton 9 Feb. 2009
… or even lots of lights
http:/www.hcibook.com/alan/projects/firefly/

4. Southampton 9 Feb. 2009
... but I will talk about
bad ideas for creativity and design
understanding regret
linked by imagination and rationality
for innovation
in computing
for innovation
in computing
using computational
modeling
using computational
modeling

5. Southampton 9 Feb. 2009
bad ideas
for creativity and design
origins ... nearly 15 years ago, UG research methods ...

6. Southampton 9 Feb. 2009
design exercise (recent example)
Collaborative or Social Networking Thing* for
babies and/or parents of babies …
… but … design a bad one / silly one
* at least some physical token or device, not purely
web/digital

7. Southampton 9 Feb. 2009
prompts …
THE BAD
1 what is bad about this idea?
2 why is this a bad thing?
3 are there any other things that
share this feature but are not
bad?
4 if so what is the difference?
try different contexts
used car salesman – how would
you sell it to someone?
THE GOOD
1 what is good about this idea?
2 why is this a good thing?
3 anything that shares this feature
but is not good?
4 if so what is the difference?

8. Southampton 9 Feb. 2009
make it a good idea
• What is good - keep it
• What is bad - change it
• Change context
• Learn from aspects

9. Southampton 9 Feb. 2009
Good Ideas
why bad ideas?
training:
– low commitment => easier to critique
design:
– large jumps through the design space
?
?

10. Southampton 9 Feb. 2009
Bad Ideas
why bad ideas?
training:
– low commitment => easier to critique
design:
– large jumps through the design space

11. Southampton 9 Feb. 2009
why bad ideas?
training:
– low commitment => easier to critique
design:
– large jumps through the design space
– understanding of the design space
Bad Ideas
Meta-level
dimensions
criteria
properties

12. Southampton 9 Feb. 2009
plus ...
• other divergent techniques:
– random metaphors, putting ideas together
• arbitrary constraints:
– time, materials, etc.
• externalisation
• personality prostheses

13. Southampton 9 Feb. 2009
bad ideas ... related things ...
critical transitions
examples

14. Southampton 9 Feb. 2009
critical transitions
• construct a boundary case …
– example A in category B not in category
– make ‘path of small changes from A to B
– where does it ‘cross’ the boundary
– good for ‘felt’ categories
in category not in category
A B
critical
transition

15. Southampton 9 Feb. 2009
but how to find examples?
• generating examples – hard
• examples from experience easy ??? or is it ???
past now
old
concept
experience
need

16. Southampton 9 Feb. 2009
but how to find examples?
• generating examples – hard
• examples from experience easy ??? or is it ???
past now
new
concept
experience
need
generate
examplesimilar surface
characteristics

17. Southampton 9 Feb. 2009
but how to find examples?
• generating examples – hard
• examples from experience ... actually harder!
but .. generating examples ...
• take arbitrary concrete example
• morph to new concept
• constant concrete – abstract movement

18. Southampton 9 Feb. 2009
modelling regret
W
ARNING!
speculative

19. Southampton 9 Feb. 2009
why regret?
it seems such a negative emotion
is there some adaptive reason for it?
... or just an accident

20. Southampton 9 Feb. 2009
features of regret
• modal/counterfactual “what if” analysis
• worst when you ‘nearly’ averted disaster
• seems to be about learning
so how do we learn ....

21. Southampton 9 Feb. 2009
sensesaction
emotion
(3) evaluation
ow! it hurts!
(4) learnt association
touching thorn
is bad
(1) touch thorn (2) thorn pricks
finger
basic reactions - learning

22. Southampton 9 Feb. 2009
sensesaction
emotion
(4) veto
(2) learnt association
‘fires’
No action!
(1) about to
touch thorn
(3) bad feeling
basic reactions – moderating action

23. Southampton 9 Feb. 2009
sensesaction
(3) learnt
association fires
(1) imagination of
planned action
(2) causes similar
brain activity to
actually doing it!
emotion
(4) veto
basic reactions – moderating intention

24. Southampton 9 Feb. 2009
only works for instant effects
so what about delayed effects?
(e.g. poisonous plant)
need imagination!

25. Southampton 9 Feb. 2009
sensesaction
emotion
(3) evaluation
“that hurts”
(1) touch plant (2) some time
later your finger
is sore
why?
(4) desire to
make sense
delayed effect – the gap

26. Southampton 9 Feb. 2009
sensesaction
(7) learnt association
don’t touch that plant
why?
(5) recent salient events
brought to mind
(6) causes simultaneous
activation in
relevant areas
emotion
delayed effect – bringing to mind

27. Southampton 9 Feb. 2009
sensesaction
(3) evaluation
yuck :-(
(7) learnt association
drinking beer is yucky
(1) drink beer (2) next morning
feel sick
(4) desire to
make sense
why?
(5) recent salient events
brought to mind
(6) causes simultaneous
activation in
relevant areas
emotion
delayed effect – put it together

28. Southampton 9 Feb. 2009
and now regret ...
similar but also:
causal connections
moderating emotions

29. Southampton 9 Feb. 2009
sensesaction
emotion
(3) evaluation
yuck :-(
(1) drink beer (2) next morning
feel sick
why?
(4) desire to
make sense
regret – the gap

30. Southampton 9 Feb. 2009
sensesaction
(7) learnt association
even though action
not obviously linked
or most salient
“drinking beer is yucky”
(5) imagination
causes simultaneous
activation in
relevant areas
emotion
(4) logical deduction of
what mattered
determines what is
brought to mind
(6) causes negative
emotion
“if only I hadn’t”
… regret
regret – casual thinking

31. Southampton 9 Feb. 2009
sensesaction(7) learnt association
stronger or weaker
depending on
strength of emotion
(5) imagination
causes simultaneous
activation in
relevant areas
emotion
(4) logical deduction of
what mattered
determines what is
brought to mind
(6) logical deduction
of how much
it matters influences
strength of emotion
regret – modifying emotion

32. Southampton 9 Feb. 2009
but is it true?
if I were a psychologist
I would run an experiment
if I were a brain scientist
I would do a scan
but as a computer scientist ...
... build a computer model

33. Southampton 9 Feb. 2009
model architecture
game
mechanics
stimulus
cards dealt
response
stick/twist
effect
win/lose
SRE
assoclookup and
choose
emotion
update plug-in
regret
module
post-hoc info.
further cards dealt
modify
basic ML module

34. Southampton 9 Feb. 2009
it works!
faster (not better) learning 

35. Southampton 9 Feb. 2009
the data
no regret
iteration %best
50 87.47
100 94.43
500 97.27
1000 97.94
with regret
iteration %best
50 90.05
100 97.31
150 97.94
1000 98.60

36. Southampton 9 Feb. 2009
and the twist ... positive regret
the code:
if ( effect negative ) do Regret
positive regret?
the grass is greener ...
in code has greatest effect
– ameliorates winner takes all local minima
for people too?
coder thinks,
“do we need the condition?”

37. Southampton 9 Feb. 2009
bad things really may be good
Bad Ideas make us creative
– with the right prompts
Regret helps us learn
– maybe machines too
understanding how we think helps us:
– develop practical techniques
– maybe even tools
both need imagination
and rationality