Semantically Far Inspirations Considered Harmful? Accounting For - - PowerPoint PPT Presentation

▶
semantically far inspirations considered harmful
SMART_READER_LITE
LIVE PREVIEW

Semantically Far Inspirations Considered Harmful? Accounting For - - PowerPoint PPT Presentation

Sep 24, 2022 28 likes •474 views

Semantically Far Inspirations Considered Harmful? Accounting For Cognitive States In Collaborative Ideation Joel Chan 1 , Pao Siangliulue 2 , Denisa Qori McDonald 3 , Ruixue Liu 3 , Reza Moradinezhad 3 , Safa Aman 3 , Erin T. Solovey 3 ,

slide-1
SLIDE 1

Semantically Far Inspirations Considered Harmful? Accounting For Cognitive States In Collaborative Ideation

Joel Chan1, Pao Siangliulue2, Denisa Qori McDonald3, Ruixue Liu3, Reza Moradinezhad3, Safa Aman3, Erin T. Solovey3, Krzysztof Z. Gajos2, Steven P. Dow4

1HCI Institute, Carnegie Mellon University 2SEAS, Harvard University 3College of Computing and Informatics, Drexel University 4Cognitive Science Department, UC San Diego

@ C&C 2017 Wednesday, June 28, 2017

slide-2
SLIDE 2

crowd innovation

Introduction Methods Results Discussion

slide-3
SLIDE 3

crowd innovation

Introduction Methods Results Discussion

slide-4
SLIDE 4

crowd innovation

Introduction Methods Results Discussion

slide-5
SLIDE 5

crowd innovation

Introduction Methods Results Discussion

slide-6
SLIDE 6

~46,000 ideas from 150,000 participants

crowd innovation

Introduction Methods Results Discussion

slide-7
SLIDE 7

$1,000,000,000 for most promising idea ~46,000 ideas from 150,000 participants

crowd innovation

Introduction Methods Results Discussion

slide-8
SLIDE 8

creativity (re)combination

(Sawyer 2012; Ward 2001)

Introduction Methods Results Discussion

slide-9
SLIDE 9

creativity (re)combination knowledge diversity crowds

(Sawyer 2012; Ward 2001)

Introduction Methods Results Discussion

slide-10
SLIDE 10

how to design interactions at scale that optimize this pathway?

creativity (re)combination knowledge diversity crowds

(Sawyer 2012; Ward 2001)

Introduction Methods Results Discussion

slide-11
SLIDE 11

how to design interactions at scale that optimize this pathway? in particular: when should you be exposed to ideas that are different from your own?

creativity (re)combination knowledge diversity crowds

(Sawyer 2012; Ward 2001)

Introduction Methods Results Discussion

slide-12
SLIDE 12

answer 1: associationist theory (re)combination knowledge diversity

Introduction Methods Results Discussion

(Gupta et al 2012; Mednick 1962; Koestler, 1964)

slide-13
SLIDE 13

(re)combination

remote associations

knowledge diversity answer 1: associationist theory

Introduction Methods Results Discussion

(Gupta et al 2012; Mednick 1962; Koestler, 1964)

slide-14
SLIDE 14

(re)combination

far stimuli remote associations

knowledge diversity answer 1: associationist theory

Introduction Methods Results Discussion

(Gupta et al 2012; Mednick 1962; Koestler, 1964)

slide-15
SLIDE 15

(re)combination

far stimuli remote associations

+novelty +diversity

knowledge diversity answer 1: associationist theory

Introduction Methods Results Discussion

(Gupta et al 2012; Mednick 1962; Koestler, 1964)

slide-16
SLIDE 16

answer 2: SIAM (search for ideas in associative memory) (re)combination

deep exploration within categories

knowledge diversity

Introduction Methods Results Discussion

(Nijstad & Stroebe 2006; Nijstad et al, 2010)

slide-17
SLIDE 17

answer 2: SIAM (search for ideas in associative memory) (re)combination

if stuck, then far stimuli; else near deep exploration within categories

knowledge diversity

Introduction Methods Results Discussion

(Nijstad & Stroebe 2006; Nijstad et al, 2010)

slide-18
SLIDE 18

answer 2: SIAM (search for ideas in associative memory) (re)combination

if stuck, then far stimuli; else near deep exploration within categories

knowledge diversity

+fluency +iteration

Introduction Methods Results Discussion

(Nijstad & Stroebe 2006; Nijstad et al, 2010)

slide-19
SLIDE 19

answer 2: SIAM (search for ideas in associative memory) (re)combination

if stuck, then far stimuli; else near deep exploration within categories

+novelty

knowledge diversity

+fluency +iteration

Introduction Methods Results Discussion

(Nijstad & Stroebe 2006; Nijstad et al, 2010)

slide-20
SLIDE 20

hypotheses to test

Predicted 
 best Predicted 
 worst Associationist Always-Far:

maximize novelty+diversity w/ remote associations

Always-Near SIAM

Introduction Methods Results Discussion

slide-21
SLIDE 21

hypotheses to test

Predicted 
 best Predicted 
 worst Associationist Always-Far:

maximize novelty+diversity w/ remote associations

Always-Near SIAM Match-State:

maximize “roll” exploration within categories: far when stuck, else near

Mismatch-State

Introduction Methods Results Discussion

slide-22
SLIDE 22

hypotheses to test no direct data yet: let’s find out!

Predicted 
 best Predicted 
 worst Associationist Always-Far:

maximize novelty+diversity w/ remote associations

Always-Near SIAM Match-State:

maximize “roll” exploration within categories: far when stuck, else near

Mismatch-State

Introduction Methods Results Discussion

slide-23
SLIDE 23

Introduction Methods Results Discussion

slide-24
SLIDE 24

(1) task:

brainstorm ideas for themed weddings

Introduction Methods Results Discussion

slide-25
SLIDE 25

(2) inspirations

  • themes + props sampled from
  • ther brainstormers
  • near/far tailored to last idea,

using GloVe (Pennington et al 2014)

Introduction Methods Results Discussion

slide-26
SLIDE 26

(2) inspirations

  • themes + props sampled from
  • ther brainstormers
  • near/far tailored to last idea,

using GloVe (Pennington et al 2014)

  • ther examples - for

“football” theme:

  • Near: [season, fun and

games, fourth of July]

  • Far: [toga, hula, prom].

Introduction Methods Results Discussion

slide-27
SLIDE 27

(2) inferring participants’ cognitive states

  • user-driven approach
  • button click = “stuck”; else, “roll”

Introduction Methods Results Discussion

slide-28
SLIDE 28

(2) inferring participants’ cognitive states

  • user-driven approach
  • button click = “stuck”; else, “roll”

Introduction Methods Results Discussion

slide-29
SLIDE 29

more details

  • 245 participants from Amazon

Mechanical Turk

  • 5 conditions:
  • No-stimuli (baseline)
  • Always-Far
  • Always-Near
  • Match-State (far if stuck; else near)
  • Mismatch-State (near if stuck; else far)
  • 8 minutes for brainstorming

Introduction Methods Results Discussion

slide-30
SLIDE 30
  • verview

Inter-idea interval Transition similarity Fluency Diversity Novelty No-stimuli Always-Far Always- Near Match- State Mismatch- State

Introduction Methods Results Discussion

slide-31
SLIDE 31

Inter-idea interval Transition similarity Novelty No-stimuli Always-Far Always-Near Match-State Mismatch-State

measured by:

median # seconds between ideas

Introduction Methods Results Discussion

lower is better

slide-32
SLIDE 32

Inter-idea interval Transition similarity Novelty No-stimuli

64.2 (5.3) 0.19 (0.01) 0.88 (0.07)

Always-Far

86.2 (5.7) * 0.12 (0.02) ** 0.64 (0.07) m

Always-Near

74.3 (5.6) 0.20 (0.02) 0.67 (0.07)

Match-State

76.6 (5.5) 0.19 (0.01) 0.88 (0.07)

Mismatch-State

88.7 (5.8) ** 0.14 (0.02) 0.79 (0.07)

measured by:

median # seconds between ideas

slower ideation if far when not stuck

Introduction Methods Results Discussion

F(4,233)=3.2, p=.01

slide-33
SLIDE 33

Transition similarity Transition similarity Novelty No-stimuli Always-Far Always-Near Match-State Mismatch-State

measured by:

mean GloVe similarity between temporally adjacent ideas

Introduction Methods Results Discussion

higher is better

slide-34
SLIDE 34

Transition similarity Transition similarity Novelty No-stimuli

0.19 (0.01) 0.19 (0.01) 0.88 (0.07)

Always-Far

0.12 (0.02) ** 0.12 (0.02) ** 0.64 (0.07) m

Always-Near

0.20 (0.02) 0.20 (0.02) 0.67 (0.07)

Match-State

0.19 (0.01) 0.19 (0.01) 0.88 (0.07)

Mismatch-State

0.14 (0.02) 0.14 (0.02) 0.79 (0.07)

Introduction Methods Results Discussion

always-far reduces iteration

measured by:

mean GloVe similarity between temporally adjacent ideas F(4,218)=4.9, p<.01

slide-35
SLIDE 35

Novelty Transition similarity Novelty No-stimuli Always-Far Always-Near Match-State Mismatch-State

Introduction Methods Results Discussion

measured by:

max (highest) z-scored subjective (1-7) rating by workers (r = .64)

  • ex. high: “[Chemistry] [Lab

experiment] (z-score=1.61).

  • ex. low: “[formal] [gift]”

(z-score=–1.94)

higher is better

slide-36
SLIDE 36

Novelty Transition similarity Novelty No-stimuli

0.88 (0.07) 0.19 (0.01) 0.88 (0.07)

Always-Far

0.64 (0.07) m 0.12 (0.02) ** 0.64 (0.07) m

Always-Near

0.67 (0.07) 0.20 (0.02) 0.67 (0.07)

Match-State

0.88 (0.07) 0.19 (0.01) 0.88 (0.07)

Mismatch-State

0.79 (0.07) 0.14 (0.02) 0.79 (0.07)

Introduction Methods Results Discussion

always-far reduces novelty

measured by:

max (highest) z-scored subjective (1-7) rating by workers (r = .64)

  • ex. high: “[Chemistry] [Lab

experiment] (z-score=1.61).

  • ex. low: “[formal] [gift]”

(z-score=–1.94)

F(4,239)=2.5, p=.04

slide-37
SLIDE 37

summary: slower, less iteration, lower novelty if far stimuli when not stuck

Introduction Methods Results Discussion

Inter-idea interval Transition similarity Novelty No-stimuli 64.2 (5.3) ** 0.19 (0.01) ** 0.88 (0.07) m Always-Far 86.2 (5.7) ** 0.12 (0.02) ** 0.64 (0.07) m Always-Near 74.3 (5.6) ** 0.20 (0.02) ** 0.67 (0.07) m Match-State 76.6 (5.5) ** 0.19 (0.01) ** 0.88 (0.07) m Mismatch-State 88.7 (5.8) ** 0.14 (0.02) ** 0.79 (0.07) m

slide-38
SLIDE 38

implications

Introduction Methods Results Discussion

slide-39
SLIDE 39

implications

  • be careful with far inspirations
  • complementary to other work on distance

from problem (Fu et al, 2013; Goncalves et al 2013; Chan et al

2015)

  • better strategies/scaffolding?
  • better mindset?
  • respect constraints (Yu et al 2016)?

Introduction Methods Results Discussion

slide-40
SLIDE 40

implications

  • be careful with far inspirations
  • complementary to other work on distance

from problem (Fu et al, 2013; Goncalves et al 2013; Chan et al

2015)

  • better strategies/scaffolding?
  • better mindset?
  • respect constraints (Yu et al 2016)?
  • need better theories (SIAM only slightly less bad)

Introduction Methods Results Discussion

slide-41
SLIDE 41

implications

  • be careful with far inspirations
  • complementary to other work on distance

from problem (Fu et al, 2013; Goncalves et al 2013; Chan et al

2015)

  • better strategies/scaffolding?
  • better mindset?
  • respect constraints (Yu et al 2016)?
  • need better theories (SIAM only slightly less bad)
  • dual paths of creativity (Nijstad et al 2010)

Introduction Methods Results Discussion

slide-42
SLIDE 42

implications

  • be careful with far inspirations
  • complementary to other work on distance

from problem (Fu et al, 2013; Goncalves et al 2013; Chan et al

2015)

  • better strategies/scaffolding?
  • better mindset?
  • respect constraints (Yu et al 2016)?
  • need better theories (SIAM only slightly less bad)
  • dual paths of creativity (Nijstad et al 2010)
  • Introduction

Methods Results Discussion

slide-43
SLIDE 43

looking ahead

  • how can we create context-aware

creativity support tools?

  • how can we best design both

sampling (IR) and interactions with inspirational stimuli?

Introduction Methods Results Discussion

can [physiological computing, BCI] give us real “thinking caps”?

slide-44
SLIDE 44

THANK YOU!

Funding Participants, PC, Reviewers, and YOU!

QUESTIONS?

joelchan.me // joelchuc@cs.cmu.edu

#1122206 #1122320