{"id":866,"date":"2023-08-28T10:51:08","date_gmt":"2023-08-28T15:51:08","guid":{"rendered":"http:\/\/infoseeking.org\/iblog\/?p=866"},"modified":"2023-08-28T11:04:50","modified_gmt":"2023-08-28T16:04:50","slug":"%f0%9d%90%80%f0%9d%90%9d%f0%9d%90%9d%f0%9d%90%ab%f0%9d%90%9e%f0%9d%90%ac%f0%9d%90%ac%f0%9d%90%a2%f0%9d%90%a7%f0%9d%90%a0-%f0%9d%90%96%f0%9d%90%9e%f0%9d%90%9a%f0%9d%90%a4-%f0%9d%90%83%f0%9d%90%9e","status":"publish","type":"post","link":"https:\/\/infoseeking.org\/iblog\/2023\/08\/28\/%f0%9d%90%80%f0%9d%90%9d%f0%9d%90%9d%f0%9d%90%ab%f0%9d%90%9e%f0%9d%90%ac%f0%9d%90%ac%f0%9d%90%a2%f0%9d%90%a7%f0%9d%90%a0-%f0%9d%90%96%f0%9d%90%9e%f0%9d%90%9a%f0%9d%90%a4-%f0%9d%90%83%f0%9d%90%9e\/","title":{"rendered":"\ud835\udc00\ud835\udc1d\ud835\udc1d\ud835\udc2b\ud835\udc1e\ud835\udc2c\ud835\udc2c\ud835\udc22\ud835\udc27\ud835\udc20 \ud835\udc16\ud835\udc1e\ud835\udc1a\ud835\udc24 \ud835\udc03\ud835\udc1e\ud835\udc1c\ud835\udc22\ud835\udc2c\ud835\udc22\ud835\udc28\ud835\udc27 \ud835\udc01\ud835\udc28\ud835\udc2e\ud835\udc27\ud835\udc1d\ud835\udc1a\ud835\udc2b\ud835\udc22\ud835\udc1e\ud835\udc2c \ud835\udc22\ud835\udc27 \ud835\udc08\ud835\udc26\ud835\udc1a\ud835\udc20\ud835\udc1e \ud835\udc02\ud835\udc25\ud835\udc1a\ud835\udc2c\ud835\udc2c\ud835\udc22\ud835\udc1f\ud835\udc22\ud835\udc1c\ud835\udc1a\ud835\udc2d\ud835\udc22\ud835\udc28\ud835\udc27 \ud835\udc1b\ud835\udc32 \ud835\udc0b\ud835\udc1e\ud835\udc2f\ud835\udc1e\ud835\udc2b\ud835\udc1a\ud835\udc20\ud835\udc22\ud835\udc27\ud835\udc20 \ud835\udc16\ud835\udc1e\ud835\udc1b \ud835\udc12\ud835\udc1e\ud835\udc1a\ud835\udc2b\ud835\udc1c\ud835\udc21 \ud835\udc1a\ud835\udc27\ud835\udc1d \ud835\udc06\ud835\udc1e\ud835\udc27\ud835\udc1e\ud835\udc2b\ud835\udc1a\ud835\udc2d\ud835\udc22\ud835\udc2f\ud835\udc1e \ud835\udc0c\ud835\udc28\ud835\udc1d\ud835\udc1e\ud835\udc25\ud835\udc2c #IJCAI2023"},"content":{"rendered":"\n

Preetam Dammu, Yunhe Feng<\/a> and Chirag Shah<\/a> <\/p>\n\n\n\n

<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

In an era where machine learning (ML) technologies are becoming more prevalent, the ethical and operational issues surrounding them cannot be ignored. Here’s how we tackled this challenge:<\/p>\n\n\n\n

\ud83d\udca1 The Problem:
ML models often don’t perform equally well for underrepresented groups, placing vulnerable populations at a disadvantage.<\/p>\n\n\n\n

\ud83c\udf10 Our Solution:
We leveraged web search and Generative AI to improve the robustness and reduce bias in discriminative ML models.<\/p>\n\n\n\n

\ud83d\udd0d Methodology:
1. We identified weak decision boundaries for classes representing vulnerable populations (e.g., female doctor of color).
2. We constructed search queries for Google and generated text for creating images with DALL-E 2 and Stable Diffusion.
3. We used these new training samples to reduce population bias.<\/p>\n\n\n\n

\ud83d\udcc8 Results:
1. Achieved a significant reduction (77.30%) in the model’s gender accuracy disparity.
2. Enhanced the classifier’s decision boundary, resulting in fewer weak spots and better class separation.<\/p>\n\n\n\n

\ud83c\udf0d Applicability:
Although demonstrated on vulnerable populations, this approach is extendable to a wide range of problems and domains.<\/p>\n\n\n\n

<\/p>\n\n\n\n

https:\/\/media.licdn.com\/dms\/document\/media\/D561FAQHC2lEiwd7tXg\/feedshare-document-pdf-analyzed\/0\/1692451349751?e=1694044800&v=beta&t=f5YjwyHd_KCSHH8ckfZcOrbyLCIlFG4u-nECZSFTX1o<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"

\ud835\udc00\ud835\udc1d\ud835\udc1d\ud835\udc2b\ud835\udc1e\ud835\udc2c\ud835\udc2c\ud835\udc22\ud835\udc27\ud835\udc20 \ud835\udc16\ud835\udc1e\ud835\udc1a\ud835\udc24 \ud835\udc03\ud835\udc1e\ud835\udc1c\ud835\udc22\ud835\udc2c\ud835\udc22\ud835\udc28\ud835\udc27 \ud835\udc01\ud835\udc28\ud835\udc2e\ud835\udc27\ud835\udc1d\ud835\udc1a\ud835\udc2b\ud835\udc22\ud835\udc1e\ud835\udc2c \ud835\udc22\ud835\udc27 \ud835\udc08\ud835\udc26\ud835\udc1a\ud835\udc20\ud835\udc1e \ud835\udc02\ud835\udc25\ud835\udc1a\ud835\udc2c\ud835\udc2c\ud835\udc22\ud835\udc1f\ud835\udc22\ud835\udc1c\ud835\udc1a\ud835\udc2d\ud835\udc22\ud835\udc28\ud835\udc27 \ud835\udc1b\ud835\udc32 \ud835\udc0b\ud835\udc1e\ud835\udc2f\ud835\udc1e\ud835\udc2b\ud835\udc1a\ud835\udc20\ud835\udc22\ud835\udc27\ud835\udc20 \ud835\udc16\ud835\udc1e\ud835\udc1b \ud835\udc12\ud835\udc1e\ud835\udc1a\ud835\udc2b\ud835\udc1c\ud835\udc21 \ud835\udc1a\ud835\udc27\ud835\udc1d \ud835\udc06\ud835\udc1e\ud835\udc27\ud835\udc1e\ud835\udc2b\ud835\udc1a\ud835\udc2d\ud835\udc22\ud835\udc2f\ud835\udc1e \ud835\udc0c\ud835\udc28\ud835\udc1d\ud835\udc1e\ud835\udc25\ud835\udc2c #IJCAI2023<\/p>\n","protected":false},"author":1,"featured_media":867,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[1],"tags":[],"jetpack_featured_media_url":"https:\/\/infoseeking.org\/iblog\/wp-content\/uploads\/2023\/08\/Screenshot-2023-08-28-at-8.50.29-AM.png","_links":{"self":[{"href":"https:\/\/infoseeking.org\/iblog\/wp-json\/wp\/v2\/posts\/866"}],"collection":[{"href":"https:\/\/infoseeking.org\/iblog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/infoseeking.org\/iblog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/infoseeking.org\/iblog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/infoseeking.org\/iblog\/wp-json\/wp\/v2\/comments?post=866"}],"version-history":[{"count":2,"href":"https:\/\/infoseeking.org\/iblog\/wp-json\/wp\/v2\/posts\/866\/revisions"}],"predecessor-version":[{"id":869,"href":"https:\/\/infoseeking.org\/iblog\/wp-json\/wp\/v2\/posts\/866\/revisions\/869"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/infoseeking.org\/iblog\/wp-json\/wp\/v2\/media\/867"}],"wp:attachment":[{"href":"https:\/\/infoseeking.org\/iblog\/wp-json\/wp\/v2\/media?parent=866"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/infoseeking.org\/iblog\/wp-json\/wp\/v2\/categories?post=866"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/infoseeking.org\/iblog\/wp-json\/wp\/v2\/tags?post=866"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}