Lecture 13 Fairness, LLM addendum

0.0(0)
Studied by 0 people
call kaiCall Kai
learnLearn
examPractice Test
spaced repetitionSpaced Repetition
heart puzzleMatch
flashcardsFlashcards
GameKnowt Play
Card Sorting

1/22

encourage image

There's no tags or description

Looks like no tags are added yet.

23 Terms

1
New cards

what does ‘fair’ even mean

“similar treatment of similar people”

2
New cards

2 dimensies van fairness met confsion matrix

equal representation : proportional selection across groups, independent of true class

equal opportunity : equally accurate in identifying true positives across groups

3
New cards

fairness metrics : demographic parity

(TP + FP) /N1 = (TP2 + FR2) / N2

equal representation across groups → as many (proportionally) men as women selected, independently of their ‘real’ class membership

<p>(TP + FP) /N1 = (TP2 + FR2) / N2 </p><p>equal representation across groups → as many (proportionally) men as women selected, independently of their ‘real’ class membership </p>
4
New cards

metrics : equal opportunity

TP1 /( TP1 + FN1) = TP2/ (TP2+FN2)

equally accurate in identifying true positives across groups (percentage of identified true positives)

equal opportunity for qualified individuals, regardless of the group

<p>TP1 /( TP1 + FN1) = TP2/ (TP2+FN2) </p><p>equally accurate in identifying true positives across groups (percentage of identified true positives) </p><p>equal opportunity for qualified individuals, regardless of the group </p>
5
New cards

external impact : mitigation strategies : fairness through unawareness

just remove the protected features A from the data and everything is fine right ?

BUT : the feature X still contain a lot of information about A

6
New cards

when can we intervene to engineer fairness

  • pre-processing

  • in-procesing (optimization)

  • post-processing


7
New cards

external impact : mitigation strategies : pre-processing

from sepcific transformatioons on the features X and/ or outcome Y with the aim to remove bias from the training data

the most flexible part of the data science, as it makes no assumptions with respect to the choice of subsequently applied modeling techniques

→ eg reweighting, re sampling, relabelling and perturbations, adversial learning

8
New cards

what does (external impact) : mitigation strategies : pre-processing requires

access to the training data

9
New cards

external impact : mitigation strategies : in-processing

incorporating one or more fairness notions into the model optimization functions

the joint optimization of performance and fairness may lead to the best result on multiple metrics

eg : constraint optimization, regularization, adversial learning, bandits

10
New cards

what does (external impact ): mitigation strategies : in-processing requires

access to the training data and model

11
New cards

external impact : mitigation strategies : post-processing

sometimes you deal wit black-boc scenarios and only have access to the predictions y^and your own input X, including the sensitive attribute A information

apply transformtaions to model output Y^ to improve prediction fairness with resoect to A

But it may worsen the performance of the original model


eg thresholding, calibrations, transormation, constraint optimization

12
New cards

what types of bias might occur during modeling

historical bias

representation bias

measurement bias

learning bias

aggregation bias

evaluation bias

deployment bias

13
New cards

Why are LLMs stochastic

bc the base model predicts word by word autoregressively until it hits a < stop> token

the model predicts each token with a certain probability therefore LLMS are stochastic

14
New cards

How LLMs adapt : what is adaptation

the process of specializing a general language model for a specific task

15
New cards

two ways of adapting LLMs

  • finetuning : updating the model’s parameters with a small, focused dataset

  • prompting (in context learning) : guiding the model by providing examples within the input prompt itself


16
New cards

LLM adaptation : by fine-tuning

starting with the langauge model and (partly) updating its parameters based on a limited set of training instances

17
New cards

LLMs concepts : adaptation : prompting

in context learning : varying the input prompt to achieve better modleing results

eg giving examples of a task in the prompt

18
New cards

defining dangers

hallucinantions

alignment

19
New cards

defining dangers : hallucinations

a confidence response by an AI that does not seem to be justified by its training data

20
New cards

defining dangers : alignment

behavior in accordance with what a human wants

alignment strategies :

  • supervised finetuning

  • RLHF

  • DPO : direct preferences optimization


21
New cards

dangers of LLMs

  • reliability + unexplainable

  • social bias

  • Locked-in/ monopoly forming

  • security

  • disingformation

  • legal considerations

  • cost and environmental impact


22
New cards

LLMs in businesses of the futiure

LLms will be used for various tasks, extent and impact highly uncertain

made available through API-based subscriptions, owned and sustained by large corporations

end-users : fine-tuning, prompt engineering, chaining of tools, agentic AI, human in the loop

the majority of projects will fail

23
New cards

agentic AI

is a system typically prowered by a language model, that can use tools to complete multi-step tasks