Sign In

Mirror: A Multi-Agent System for AI-Assisted Ethics Review

Created by
  • Haebom
Category
Empty

์ €์ž

Yifan Ding, Yuhui Shi, Zhiyan Li, Zilong Wang, Yifeng Gao, Yajun Yang, Mengjie Yang, Yixiu Liang, Xipeng Qiu, Xuanjing Huang, Xingjun Ma, Yu-Gang Jiang, Guoyu Wang

๐Ÿ’ก ๊ฐœ์š”

๋ณธ ๋…ผ๋ฌธ์€ ๋Œ€๊ทœ๋ชจ ๊ณผํ•™ ์—ฐ๊ตฌ์—์„œ ๋ฐœ์ƒํ•˜๋Š” ์œค๋ฆฌ์  ๋ฌธ์ œ์— ๋Œ€ํ•œ ๊ฒ€ํ† ์˜ ๋ถ€๋‹ด์„ ํ•ด์†Œํ•˜๊ธฐ ์œ„ํ•ด AI ๊ธฐ๋ฐ˜ ๋‹ค์ค‘ ์—์ด์ „ํŠธ ์‹œ์Šคํ…œ์ธ Mirror๋ฅผ ์ œ์•ˆํ•œ๋‹ค. Mirror๋Š” ์œค๋ฆฌ์  ์ถ”๋ก , ๊ทœ์น™ ํ•ด์„, ๋‹ค์ค‘ ์—์ด์ „ํŠธ ํ˜‘์—…์„ ํ†ตํ•ฉํ•˜๋ฉฐ, EthicsQA ๋ฐ์ดํ„ฐ์…‹์œผ๋กœ ํŒŒ์ธํŠœ๋‹๋œ EthicsLLM์„ ๊ธฐ๋ฐ˜์œผ๋กœ ํ•œ๋‹ค. ์ด๋ฅผ ํ†ตํ•ด ๊ฐ„์†Œํ™”๋œ ๊ฒ€ํ† (Mirror-ER)์™€ ์œ„์›ํšŒ ๊ฒ€ํ† (Mirror-CR) ๋ชจ๋“œ๋ฅผ ์ง€์›ํ•˜์—ฌ ๊ธฐ์กด LLM ๋Œ€๋น„ ํ–ฅ์ƒ๋œ ์œค๋ฆฌ ํ‰๊ฐ€์˜ ์งˆ, ์ผ๊ด€์„ฑ, ์ „๋ฌธ์„ฑ์„ ๋ณด์—ฌ์ค€๋‹ค.

๐Ÿ”‘ ์‹œ์‚ฌ์  ๋ฐ ํ•œ๊ณ„

โ€ข
AI๊ฐ€ ์œค๋ฆฌ ๊ฒ€ํ†  ํ”„๋กœ์„ธ์Šค์˜ ํšจ์œจ์„ฑ๊ณผ ์ผ๊ด€์„ฑ์„ ํฌ๊ฒŒ ํ–ฅ์ƒ์‹œํ‚ฌ ์ž ์žฌ๋ ฅ์„ ๊ฐ€์ง€๊ณ  ์žˆ์Œ์„ ์ž…์ฆํ•ฉ๋‹ˆ๋‹ค.
โ€ข
์œค๋ฆฌ์  ๊ทœ์น™๊ณผ ๊ทœ์ œ๋ฅผ ํ•ด์„ํ•˜๊ณ  ์ ์šฉํ•˜๋Š” AI์˜ ๋Šฅ๋ ฅ์ด ๋ฐœ์ „ํ•˜๊ณ  ์žˆ์Œ์„ ๋ณด์—ฌ์ค๋‹ˆ๋‹ค.
โ€ข
์‹ค์ œ ์œค๋ฆฌ ๊ฒ€ํ† ์—์„œ ๋ฏผ๊ฐํ•œ ๋ฐ์ดํ„ฐ๋ฅผ ๋‹ค๋ฃจ๋Š” ํ”„๋ผ์ด๋ฒ„์‹œ ๋ฌธ์ œ์™€ AI์˜ ํŒ๋‹จ์— ๋Œ€ํ•œ ์ฑ…์ž„ ์†Œ์žฌ ๋“ฑ์€ ์—ฌ์ „ํžˆ ์ค‘์š”ํ•œ ํ•œ๊ณ„์ ์œผ๋กœ ๋‚จ์•„ ์žˆ์œผ๋ฉฐ, ํ–ฅํ›„ ์—ฐ๊ตฌ ๊ณผ์ œ์ž…๋‹ˆ๋‹ค.
๐Ÿ‘
Made with Slashpage