Machine Learning Flashcards

1
Q

data mining is focused on

A

better understanding patterns among variables in databases using statistical/analytical tools

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

machine learning

A

scientific study of algorithms that computer systems use to perform tasks without explicit instructions, using patterns and inference instead

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

use machine learning when (2)

A

Human expertise does not exist or humans can’t explain their expertise

Models need to be customized or are based on huge amounts of data

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

ML is less useful when rules are ____ and _______, like calculating payroll amounts

A

clear and understood

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

supervised/inductive learning

A

training data and desired outputs are given

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

unsupervised learning

A

where training data is given without the desired output

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

semi-supervised learning

A

where training data is given with a few desired outputs

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

reinforcement learning

A

agent learns to make decisions by interacting with an environment
ex. robot in a maze

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Chatgpt is not directly programmed to give answers. It has functionality similar to an ____-__________, aka predicting the next words. The prediction is based on training data from 2021 and prior

A

auto-completion

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

training data

A

known outcomes

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

validation data

A

used to fine-tune a model

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

test data

A

tests accuracy of the model

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

data in ML can be partitioned into (3)

A

training data
validation data
test data

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

every ML algorithm has 3 components:

A

representation
optimization
evaluation

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

applications of ML in SC (5)

A

Forecasting demand
Optimizing speed
Planning movement of goods based on demand
Ensuring quality
Managing suppliers/documentation

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

prompt design

A

art and science of creating the perfect prompt for a language model like ChatGPT

17
Q

persona pattern

A

strategy setting the AI to function from a certain perspective (“assume the persona of a historian…”)

18
Q

cognitive verifier pattern

A

breaks down complex questions into smaller, manageable sub-questions

19
Q

few shot prompting

A

idea is to feed model with a number of examples before asking the desired question

20
Q

chain of thought prompting

A

using a series of interconnected prompts
Ex. if interested in impact of AI on job markets, asking a string questions related to the subject matter

21
Q

AI exists when a machine has ________ ______ like problem solving

A

cognitive ability

22
Q

narrow AI

A

machine can perform better than a human at a specific task

23
Q

general AI

A

machine performs similar to use in any intellectual task

24
Q

strong AI

A

machine performs better than us in many tasks

25
Q

deep learning

A

AI method where algorithms inspired by human brain/neural network are used to process data and learn patterns

26
Q

AI is →

A

Machine learning → Deep learning

27
Q

issue of ML in SC

A

ML learning depends on reliable, high-quality information, so a lack of access to good data can cause significant issues

28
Q

stock level analysis

A

identify when products are declining in popularity/reaching end of life in marketplace

29
Q

price analysis

A

comparing costs in SC and retail profit margins to est. best combination of pricing +
customer demand

30
Q

supplier analysis

A

upstream delays canbe identified, allowing for contingency planning or alternative sourcing

31
Q

sales analysis

A

retailers can link sales and promotional activities to demand and
supply planning so that stores do not run out of stock