5: Technology management, Empathic computing & Machine Learning

Question 1

technology diffusion lens =

+ 8=3+4+1 why’s

Answer 1

= 3 phases of tech adoption lifecycle, w S curve

1. early adopters w specific needs, who don’t mind initial difficulties
2. going mass
3. even laggards

• *Coz:**
• 3 start mental factors:*
• it takes a change of mind
• risk of change (risk appetite is different among public)
• it takes know-how to use new tool

4 start economic issues:

• ppl want to amortize their old tools before changing
• initial price is high
• availability of new good is limited
• complementary goods
• final:*
• at the end, there is market saturation

Question 2

new tech adoption: groups along the S curve

Answer 2

• 2.5% innovators –> rich & educated risk-takers
• 13.5% early adopters
• 34% early majority
• 34% late majority
• 16% laggards –> older, less educated & conservative

Question 3

new VS mature technology:

qualitative & quantitative differences

=> expenses VS performance curve

Answer 3

• w new tech, basic knowledge is missing, while it is present for mature tech
• w new tech the increase of performance is exponential in time, while w mature tech R&D expense has diminishing returns

=> expenses VS performance curve is S-shaped

Question 4

disruptive innovation & S-shaped expense/performance curves:

• when is tech disruptive
• what markets first

Answer 4

• disruptive tech <==> intersecting curves
• the new product performance/expenses curve will reach first the performance demanded by low-end markets, then that of high-end markets

Question 5

persuasion technology

def= w 2 key concepts

3 examples

Answer 5

IT-based implementation of psychological “nudges” based on people’s predictable irrationality

examples:

• the power of defaults
• the power of framing, e.g. adding dominated options
• apps performing just-in-time adaptive interventions based on deduced emotional status

Question 6

Machine Learning:

• basic concept def= in 3 elements
• • 2 schools of thought & their comparison (who’s winning?)
• what makes ML attractive?

Answer 6

• Data / Input => Relationship / Algorithm / Predictor => Response / Output
• • 2 schools of thought:

1. “statistical” school assumes stochastical data model generating the data
2. algorithmic school focuses on the learning algorithms and remains agnostic on the model

=> algorithmic school is gaining ground over ‘wasteful’ statistical school, according to Prof. Fleisch

• ML can (sometimes) construct programs from data, which saves effort !

Question 7

machine learning in 4 phases

Answer 7

FE.TraM.P.F.:

1. Features Extraction from Data
2. (construction of) Trained Model
3. Prediction
4. Feedback

then back to 1.