๋กœ๊ทธ์ธ

Arcane: An Assertion Reduction Framework through Semantic Clustering and MCTS-Guided Rule Exploring

์ž‘์„ฑ์ž
  • Haebom
์นดํ…Œ๊ณ ๋ฆฌ
๋น„์–ด ์žˆ์Œ

์ €์ž

Hongqin Lyu, Yonghao Wang, Zhiteng Chao, Tiancheng Wang, Huawei Li

๐Ÿ’ก ๊ฐœ์š”

๊ธฐ์กด ์ž๋™ํ™”๋œ ๋ช…์ œ ์ƒ์„ฑ ๋ฐฉ์‹์€ ๋ฐฉ๋Œ€ํ•œ ์–‘์˜ ์ค‘๋ณต ๋ช…์ œ๋ฅผ ์ƒ์„ฑํ•˜์—ฌ ์‹œ๋ฎฌ๋ ˆ์ด์…˜ ํšจ์œจ์„ฑ์„ ์ €ํ•˜์‹œํ‚ค๋Š” ๋ฌธ์ œ๋ฅผ ๊ฐ€์ง€๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ๋ณธ ๋…ผ๋ฌธ์—์„œ๋Š” ์ด๋Ÿฌํ•œ ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•˜๊ธฐ ์œ„ํ•ด, ์˜๋ฏธ๋ก ์  ํด๋Ÿฌ์Šคํ„ฐ๋ง๊ณผ MCTS ๊ธฐ๋ฐ˜ ๊ทœ์น™ ํƒ์ƒ‰์„ ํ†ตํ•ฉํ•œ ๋ช…์ œ ์ถ•์†Œ ํ”„๋ ˆ์ž„์›Œํฌ์ธ Arcane์„ ์ œ์•ˆํ•ฉ๋‹ˆ๋‹ค. Arcane์€ ์‹คํ—˜ ๊ฒฐ๊ณผ, ๋ช…์ œ ์ˆ˜๋ฅผ ์ตœ๋Œ€ 76.2%๊นŒ์ง€ ์ค„์ด๋ฉด์„œ๋„ ํ˜•์‹์  ์ปค๋ฒ„๋ฆฌ์ง€์™€ ๋Œ์—ฐ๋ณ€์ด ํƒ์ง€ ๋Šฅ๋ ฅ์„ ์™„๋ฒฝํ•˜๊ฒŒ ๋ณด์กดํ•˜๋ฉฐ, ์‹œ๋ฎฌ๋ ˆ์ด์…˜ ์‹œ๊ฐ„์„ 2.6๋ฐฐ์—์„œ 6.1๋ฐฐ๊นŒ์ง€ ๋‹จ์ถ•์‹œ์ผฐ์Šต๋‹ˆ๋‹ค.

๐Ÿ”‘ ์‹œ์‚ฌ์  ๋ฐ ํ•œ๊ณ„

โ€ข
๋Œ€๊ทœ๋ชจ ๋ช…์ œ ์ง‘ํ•ฉ์—์„œ ์˜๋ฏธ๋ก ์  ํด๋Ÿฌ์Šคํ„ฐ๋ง์„ ํ†ตํ•ด ์ค‘๋ณต์„ ํšจ๊ณผ์ ์œผ๋กœ ์‹๋ณ„ํ•˜๊ณ  ์ œ๊ฑฐํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
โ€ข
MCTS๋ฅผ ํ™œ์šฉํ•˜์—ฌ ๋ช…์ œ ์ถ•์†Œ ๊ณผ์ •์—์„œ ์ตœ์ ์˜ ๊ทœ์น™ ์ ์šฉ ์ˆœ์„œ๋ฅผ ํƒ์ƒ‰ํ•จ์œผ๋กœ์จ ํšจ์œจ์„ฑ์„ ๊ทน๋Œ€ํ™”ํ•ฉ๋‹ˆ๋‹ค.
โ€ข
์ œ์•ˆ๋œ Arcane ํ”„๋ ˆ์ž„์›Œํฌ๋Š” ์‹ค์ œ ํ•˜๋“œ์›จ์–ด ๊ฒ€์ฆ์—์„œ ์‹œ๋ฎฌ๋ ˆ์ด์…˜ ์‹œ๊ฐ„ ๋‹จ์ถ•๊ณผ ํšจ์œจ์„ฑ ํ–ฅ์ƒ์— ํฌ๊ฒŒ ๊ธฐ์—ฌํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
โ€ข
์ œ์•ˆ๋œ ํ”„๋ ˆ์ž„์›Œํฌ์˜ ์‹ค์ œ ์ ์šฉ ๊ฐ€๋Šฅ์„ฑ๊ณผ ๋‹ค์–‘ํ•œ ๊ฒ€์ฆ ์‹œ๋‚˜๋ฆฌ์˜ค์—์„œ์˜ ์ผ๋ฐ˜ํ™” ์„ฑ๋Šฅ์— ๋Œ€ํ•œ ์ถ”๊ฐ€ ์—ฐ๊ตฌ๊ฐ€ ํ•„์š”ํ•ฉ๋‹ˆ๋‹ค.
๐Ÿ‘
Slashpage๋กœ ์ œ์ž‘๋จ