After the topic of the CS degree comes up, later on in the conversation I will try to bring up things relevant to my background that make me stand out from the typical CS grad she may be used to. Here's some of the things I may bring up:
- Apollo Agriculture:
In short, this is a company that uses technology to aid farmers in emerging economies. From the website: "Our first product is a customized package of seed and fertilizer, farming advice, and credit delivered to maize farmers in Kenya.", this is genuinely a really cool company so it's not hard for me to be enthusiastic when talking about what they're trying to do.
- Machine Learning biases:
I'll talk about how biases in training data can manifest themselves in the learning outcomes. For example, if the training data suggests that a particular socioeconomic demographic is more likely to default on a loan then the neural network will learn those biases and they'll be apparent in the resulting analysis. This might perhaps result in a higher interest rate being offered on a financial product to that demographic resulting in them having to make higher monthly payments and keeping them disadvantaged.
- Word2Vec biases:
I'll introduce them to the idea that given some corpus of text (and i'll elaborate on that) and some word, you can produce a vector for it. Similarly given a vector and you can take the closest word that matches it you can also add and subtract these vectors and then find a word to match it. Finally, I give them some of the common examples that are used to showcase the results you might achieve:
- “Paris : France :: Tokyo : x” and it will give you the answer x = Japan.
- “father : doctor :: mother : x” and it will say x = nurse.
- “man : computer programmer :: woman : x” gives x = homemaker.
You can learn more about it here if you'd like to learn how to better explain it:
https://www.technologyreview.com/s/60202...-language/
Little things like that help me convey quickly and in a way she can understand the ways in which computers can be used to change the world for better and for worse. Demonstrating awareness of this seems to be engaging to left-leaning types which appears to be most women to some extent.
Have any of you ran anything of the sort? I'm after more ideas of similar nature, hopefully ones that paint CS in a positive light like Apollo agro.
- Apollo Agriculture:
In short, this is a company that uses technology to aid farmers in emerging economies. From the website: "Our first product is a customized package of seed and fertilizer, farming advice, and credit delivered to maize farmers in Kenya.", this is genuinely a really cool company so it's not hard for me to be enthusiastic when talking about what they're trying to do.
- Machine Learning biases:
I'll talk about how biases in training data can manifest themselves in the learning outcomes. For example, if the training data suggests that a particular socioeconomic demographic is more likely to default on a loan then the neural network will learn those biases and they'll be apparent in the resulting analysis. This might perhaps result in a higher interest rate being offered on a financial product to that demographic resulting in them having to make higher monthly payments and keeping them disadvantaged.
- Word2Vec biases:
I'll introduce them to the idea that given some corpus of text (and i'll elaborate on that) and some word, you can produce a vector for it. Similarly given a vector and you can take the closest word that matches it you can also add and subtract these vectors and then find a word to match it. Finally, I give them some of the common examples that are used to showcase the results you might achieve:
- “Paris : France :: Tokyo : x” and it will give you the answer x = Japan.
- “father : doctor :: mother : x” and it will say x = nurse.
- “man : computer programmer :: woman : x” gives x = homemaker.
You can learn more about it here if you'd like to learn how to better explain it:
https://www.technologyreview.com/s/60202...-language/
Little things like that help me convey quickly and in a way she can understand the ways in which computers can be used to change the world for better and for worse. Demonstrating awareness of this seems to be engaging to left-leaning types which appears to be most women to some extent.
Have any of you ran anything of the sort? I'm after more ideas of similar nature, hopefully ones that paint CS in a positive light like Apollo agro.