profile-img
The merit of an action lies in finishing it to the end.
slide-image
์ „์ฒด ๊ธ€ + 29
list_img
Chapter 2. Uninformed Search
user-img iamnotwhale
2023.05.17
Example Problems Robotic Vacuum Cleaner 8-Puzzle 8-queens Route-finding problem Solving Problems by Searching ์šฉ์–ด ์ •์˜ Problem formulation: ์–ด๋–ค ๊ณ ๋ คํ•  state, action์„ ๊ฒฐ์ •ํ• ์ง€ ์ •ํ•˜๋Š” ํ”„๋กœ์„ธ์Šค → ๋ชฉํ‘œ๋Š” ์ฃผ์–ด์ ธ ์žˆ์Œ ๋ฌธ์ œ๋Š” state space, initial state, goal state, action, transition model, action cost function์œผ๋กœ ๊ตฌ์„ฑ action์˜ ์—ฐ์†์€ path๋ฅผ ๊ตฌ์„ฑ solution: initial state→goal state๋กœ ๊ฐ€๋Š” path optimal solution: path cost๊ฐ€ ๊ฐ€์žฅ ์ ์€ ํ•ด๋‹ต search: ๋ชฉํ‘œ์— ๋„..
list_img
Chapter 12. Big Data: Achieving Scale
user-img iamnotwhale
2023.04.19
Big Data Intro - 2003๋…„ ์ „๊นŒ์ง€ 5 ์—‘์‚ฌ๋ฐ”์ดํŠธ์˜ ์ •๋ณด๋ฅผ ์ƒ์‚ฐํ–ˆ์Œ - ์ด์ œ ์ดํ‹€๋งˆ๋‹ค 5 ์—‘์‚ฌ๋ฐ”์ดํŠธ์˜ ์ •๋ณด๋ฅผ ์ƒ์‚ฐ -> ๋น…๋ฐ์ดํ„ฐ Big Data์˜ 4V - Volume - Variety - Velocity - Veracity Science Paradigms - ๋ช‡์ฒœ ๋…„ ์ „: ๊ณผํ•™์€ empirical ํ–ˆ๋‹ค. -> ์ž์—ฐ ํ˜„์ƒ์„ ์„œ์ˆ  - ๋ช‡๋ฐฑ ๋…„ ์ „: theoretical branch -> ์ด๋ก , ๋ชจ๋ธ๋ง - ๋ช‡์‹ญ ๋…„ ์ „: computational -> ์‹œ๋ฎฌ๋ ˆ์ด์…˜ - ์˜ค๋Š˜๋‚ : data exploration - ๋ฏธ๋ž˜: Data-driven Science -> Data-driven Hypothesis Generation Big Data Challenges - ๋น…๋ฐ์ดํ„ฐ * ํฌ๊ณ  ๋ณต์žกํ•œ ๋ฐ์ดํ„ฐ * ์˜ˆ: s..
list_img
Chapter 11. Machine Learning
user-img iamnotwhale
2023.04.19
๋จธ์‹ ๋Ÿฌ๋‹ ๋ชจ๋ธ์˜ ๋น„๊ต - Power, expressibility: ์–ผ๋งˆ๋‚˜ ๋ณต์žกํ•œ ์ž‘์—…์„ ํ•  ์ˆ˜ ์žˆ๋Š๋ƒ - Interpretability - Ease of Use - Training speed - Prediction speed Linear Regression Nearest Neighbor Deep Learning Power/Expressibility L L H Interpretability H H L Ease of Use H H L Training speed H H L Prediction speed H L H cf) ๋”ฅ๋Ÿฌ๋‹์€ Foward Fast๋ฅผ ์ด์šฉํ•œ๋‹ค. Nearest Neighbor๋Š” ๊ฑฐ๋ฆฌ ๊ณ„์‚ฐ ๋•Œ๋ฌธ์— ์—ฐ์‚ฐ ์‹œ๊ฐ„์ด ๊ธธ๋‹ค. XOR & Linear Classifier - Linear Classifier๋Š”..
list_img
Chapter 10. Distance and Network Methods
user-img iamnotwhale
2023.04.15
Nearest Neighbor Classification - ์–ด๋–ค training example์ด target์— ๊ฐ€์žฅ ๊ฐ€๊นŒ์šด์ง€ ํ™•์ธํ•ด์„œ class label์„ ๋ถ™์ด๋Š” ๊ฒƒ - distance function์„ ์˜ฌ๋ฐ”๋ฅด๊ฒŒ ์ •์˜ํ•˜๋Š” ๊ฒƒ์ด ์ค‘์š” - ์žฅ์ : simplicity, interpretability, non-linearity - k-nearest neighbor Distance Metrics ๋‹ค์Œ๊ณผ ๊ฐ™์€ ์กฐ๊ฑด์„ ๋งŒ์กฑํ•˜๋Š” ๊ฒฝ์šฐ - d(x, y) >= 0 for all x, y (positivity) - d(x, y) = 0 iff x = y (identity) - d(x, y) = d(y, x) (symmetry) - d(x, y) grid indices, kd-trees, Voronoi diagrams, ..
list_img
Chapter 9. Linear and Logistic Regression
user-img iamnotwhale
2023.04.14
Linear Regression - n๊ฐœ์˜ ์ ์ด ์ฃผ์–ด์กŒ์„ ๋•Œ, ๊ฐ€์žฅ ๊ทผ์‚ฌ๋ฅผ ์ž˜ ํ•˜๊ฑฐ๋‚˜ ์ž˜ ๋งž๋Š” ์ง์„ ์„ ์ฐพ๋Š” ๊ฒƒ Error in Linear Regression - residual error: ์˜ˆ์ธก๊ฐ’๊ณผ ์‹ค์ œ๊ฐ’ ์‚ฌ์ด์˜ ์ฐจ์ด - Least squares regression์€ ๋ชจ๋“  ์ ์˜ ์ž”์ฐจ์˜ ํ•ฉ์„ ์ตœ์†Œํ™”ํ•จ. -> nice closed form, ๋ถ€ํ˜ธ ๋ฌด์‹œํ•˜๋ฏ€๋กœ ์„ ํƒ๋จ Contour plots - gradient descent Linear function์„ ์‚ฌ์šฉํ•˜๋Š” ์ด์œ  - ์ดํ•ดํ•˜๊ธฐ ์‰ฌ์›€ - default model์— ์ ํ•ฉ * ์ผํ•œ ์‹œ๊ฐ„์— ๋”ฐ๋ผ ๊ธ‰์—ฌ๊ฐ€ ์ฆ๊ฐ€ / ์ง€์—ญ์ด ์ปค์ง์œผ๋กœ์จ ์„ ํ˜•์ ์œผ๋กœ ์ง‘๊ฐ’์ด ์ƒ์Šน / ๋จน์€ ์Œ์‹์— ๋”ฐ๋ผ ๋ชธ๋ฌด๊ฒŒ๊ฐ€ ์„ ํ˜•์ ์œผ๋กœ ์ฆ๊ฐ€ ๋ณ€์ˆ˜๊ฐ€ ์—ฌ๋Ÿฌ ๊ฐœ์ผ ๋•Œ - ๊ฐ๊ฐ์˜ x_n ๋ณ€์ˆ˜๋“ค๊ณผ y๊ฐ’์„ ํ–‰๋ ฌ๋กœ ๋‚˜..
1 2 3 4 5 6