Closed dashaasienga closed 4 months ago
dashaasienga
commented
8 months ago General question about the mechanics of writing a thesis/ research paper:
What is the recommendation on using the 1st person to describe research steps? Should I refrain from that and just focus on writing more objectively about the topic at hand?
katcorr
commented
8 months ago it's fine to use first person to describe research steps
On Wed, Nov 1, 2023 at 11:38 PM dashaasienga @.***> wrote:
General question about the mechanics of writing the thesis:
What is the recommendation on using the 1st person to describe research steps? Should I refrain from that and just focus on writing more objectively about the topic at hand?
— Reply to this email directly, view it on GitHub https://github.com/dashaasienga/Statistics-Senior-Honors-Thesis/issues/15#issuecomment-1790013214, or unsubscribe https://github.com/notifications/unsubscribe-auth/AGMTX5UFOCI5A5JWLCBEC7DYCMIUNAVCNFSM6AAAAAA62HIDPOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTOOJQGAYTGMRRGQ . You are receiving this because you were mentioned.Message ID: @.*** com>
--
Kat Correia, PhD
Assistant Professor of Statistics
Department of Mathematics and Statistics
Amherst College | Seeley Mudd 402
@.*** | (413)-542-5836
dashaasienga
commented
8 months ago Is there any specific citation format that's recommended? APA, MLA, Chicago? Or is any fine? We can chat about this tomorrow as well!
@katcorr
I'm using this issue to think through what I've done so far and craft an outline for Chapter 1 of my thesis. We can talk through this next week as well, but posting it here anyways.
An Introduction to Algorithmic Bias and Statistical Fairness
This section will go into detail about what fairness ML aims to achieve and present the motivation for further research in the area by highlighting which types of issues it aims to solve. I will give examples of contexts where this comes up, including sources of algorithmic bias.
As part of that, I will also formally define algorithmic bias and what we mean when we refer to statistical notions of fairness, at least in the context of the way we have been studying it and will continue to.
Fairness Definitions
As we continue to set the stage for the rest of the thesis, I will spend some time in this section defining the key fairness metrics in the classification context.
I'll give some explanation for the differences between group v individual notions, although preface that this is a continuum and we'll be majorly focusing on group notions of fairness.
I'm thinking of outlining the group fairness definitions under the 3 major subcategories and giving an in-depth explanation of each of them:
I'll conclude this section with a brief mention of some of the existing definitions in the regression context, prefacing that not a lot of extensive research has been done, but it will nevertheless be important for the regression example we will show later on.
Fairness Conflicts
After defining the fairness metrics, I'll spend time in this section highlighting some of the fairness conflicts and showing how some fairness definitions cannot mathematically/ statistically hold at the same time. The equation we re-derived from the paper will be something important to demonstrate in this section.
I could also go into detail about which situations/ scenarios/ settings different fairness definitions may be more suitable to highlight the human value/ judgement that is needed when solving these types of problems.
Addressing Algorithmic Bias
Finally, I'll use this section as a brief overview of some of the existing methods for addressing algorithmic bias. As part of this, I'll preface the Seldonian Algorithm, which will be the focus of the following chapter, so I won't go into too much detail about it here.