Posts by Category

python

ํŠน๋ณ„ํ•œ ํŒŒ์ด์ฌ ํ•จ์ˆ˜

less than 1 minute read

ํŒŒ์ด์ฌ์—๋Š” ๋ณต์žกํ•œ ๊ณผ์ •์„ ๋‹จ์ˆœํ™” ์‹œ์ผœ์ฃผ๋Š” ํŠน๋ณ„ํ•œ ํ•จ์ˆ˜๋“ค์ด ๋งŽ์ด ์กด์žฌํ•œ๋‹ค. ์ด๋Ÿฌํ•œ ํ•จ์ˆ˜๋“ค์„ ์ •๋ฆฌํ•ด๋ณด์•˜๋‹ค.

Masked Language Model vs Causal Language Model

1 minute read

BERT์™€ ๊ฐ™์€ Masked Language Model๊ณผ GPT์™€ ๊ฐ™์€ Causal Language Model์˜ ์ฐจ์ด์ ์„ ์ •๋ฆฌํ•ด๋ณด์•˜๋‹ค. ํ•ด๋‹น ๋งํฌ๋ฅผ ์ฐธ๊ณ ํ•˜์—ฌ ์ž‘์„ฑํ•˜์˜€๋‹ค.

Enriching Word Vectors with Subword Information

less than 1 minute read

์ด๋ฒˆ ๋ฐœํ‘œ ๋…ผ๋ฌธ์˜ ์ œ๋ชฉ์€ Enriching Word Vectors with Subword Information์œผ๋กœ, ์ผ๋ช… FastText ๋ฐฉ๋ฒ•์„ ์ƒˆ๋กญ๊ฒŒ ์ œ์‹œํ•œ ๋…ผ๋ฌธ์ด๋‹ค.

์ œ4ํšŒ AI x Bookathon ์ฐธ์—ฌํ›„๊ธฐ

3 minute read

โ€ƒ ์ง€๋‚œ 3์ฃผ(22.12.27~ 23.1.18) ๋™์•ˆ ์ง„ํ–‰๋˜์—ˆ๋˜ AI x Bookathon ๋Œ€ํšŒ์— ๋Œ€ํ•œ ํ›„๊ธฐ๋ฅผ ๋‚จ๊ธฐ๊ณ ์ž ํ•œ๋‹ค. ํ•ด๋‹น ๋Œ€ํšŒ๋Š” Language Model์„ ์ด์šฉํ•ด ์ˆ˜ํ•„์„ ์ž‘์„ฑํ•˜๋Š” ํ•ด์ปคํ†ค ๋Œ€ํšŒ๋กœ, ์˜ˆ์„ ๊ณผ ๋ณธ์„ ์œผ๋กœ ์ด๋ฃจ์–ด์ ธ ์žˆ์—ˆ๋‹ค. ์˜ˆ์„ ์€ ๋”ฅ๋Ÿฌ๋‹์˜ ๊ธฐ์ดˆ ์ง€์‹๋“ค์— ๋Œ€ํ•œ ํ€ด์ฆˆ...

Lecture 4: Maximum Likelihood Learning

less than 1 minute read

4๋ฒˆ์งธ ๊ฐ•์˜๋Š” Maximum Likelihood Learning์— ๋Œ€ํ•ด ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Lecture 3: Autoregressive Model

less than 1 minute read

3๋ฒˆ์งธ ๊ฐ•์˜๋Š” Autoregressive Model์— ๋Œ€ํ•ด ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Lecture 1&2: Introduction and Background to Deep Generative Model

less than 1 minute read

์ฒซ๋ฒˆ์งธ์™€ ๋‘๋ฒˆ์งธ ๊ฐ•์˜๋Š” Deep Generative Model์— ๋Œ€ํ•œ ์†Œ๊ฐœ์™€ ๋ฐฑ๊ทธ๋ผ์šด๋“œ ์ง€์‹์— ๋Œ€ํ•ด์„œ ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Pytorch tools

less than 1 minute read

view() vs shape() vs permute() vs transpose(): click here

์ •๋ ฌ

5 minute read

๋“ค์–ด๊ฐ€๊ธฐ ์ „์—

ํž™

4 minute read

๋“ค์–ด๊ฐ€๊ธฐ ์ „์—

ํ•ด์‹œ ํ…Œ์ด๋ธ”

5 minute read

๋“ค์–ด๊ฐ€๊ธฐ ์ „์— ํ•ด์‹œ: ํ•ด์‹œ ํ…Œ์ด๋ธ”์—์„œ ์ž…๋ ฅ๊ฐ’์„ ๊ณ ์ • ํฌ๊ธฐ ๊ฐ’์œผ๋กœ ๋งคํ•‘ํ•˜๋Š” ๊ณผ์ •์„ ํ•ด์‹ฑ์ด๋ผ๊ณ  ํ•œ๋‹ค. ํ•ด์‹œ ํ•จ์ˆ˜: ํ•ด์‹œ๋ฅผ ํ•˜๋Š”๋ฐ ์‚ฌ์šฉ๋˜๋Š” ํ•จ์ˆ˜. ์„ฑ๋Šฅ์ด ์ข‹์€ ํ•ด์‹œ ํ•จ์ˆ˜๋“ค์˜ ํŠน์ง•์€ ๋‹ค์Œ๊ณผ ๊ฐ™๋‹ค. ํ•ด์‹œ ํ•จ์ˆ˜ ๊ฐ’ ์ถฉ๋Œ์˜ ์ตœ์†Œํ™” ์‰ฝ๊ณ  ๋น ๋ฅธ ์—ฐ์‚ฐ ...

์„œ์šธ์‹œ ๋น…๋ฐ์ดํ„ฐ์บ ํผ์Šค ๊ณต๋ชจ์ „

4 minute read

2021๋…„ 9์›” 13์ผ๋ถ€ํ„ฐ 10์›” 21์ผ๊นŒ์ง€ ์ง„ํ–‰ํ•œ ์„œ์šธ์‹œ ๋น…๋ฐ์ดํ„ฐ์บ ํผ์Šค ๊ณต๋ชจ์ „์— ์ฐธ๊ฐ€ํ•˜์˜€๋‹ค. ๊ณต๋ชจ์ „์ด ๋๋‚œ์ง€๋Š” ๋ฌด๋ ค ํ•œ๋‹ฌ(โ€ฆ)์ด ์ง€๋‚ฌ์ง€๋งŒ ํ›„๊ธฐ ์ž‘์„ฑ์„ ๋ฏธ๋ฃจ๋‹ค๊ฐ€ ์ง€๊ธˆ์—์„œ์•ผ ํ›„๊ธฐ๋ฅผ ์ž‘์„ฑํ•˜๋ ค ํ•œ๋‹ค.

Assignment 1

1 minute read

Assignment 1์€ ํŒŒ์ด์ฌ ์ฝ”๋“œ๋ฅผ ์ง์ ‘ ์ž‘์„ฑํ•˜์—ฌ ์ด๋ฏธ์ง€ ๋ถ„๋ฅ˜ ๋ชจ๋ธ์„ ์ง์ ‘ ๋งŒ๋“œ๋Š” ๋‚ด์šฉ์ด๋‹ค. ์‚ฌ์šฉํ•˜๋Š” ๋ชจ๋ธ๋กœ๋Š” KNN, SVM, two-layer-net(deep learning network with one hidden layer) ์ด ์žˆ๋‹ค. ์ด ๋ชจ๋ธ๋“ค์— ๋Œ€ํ•ด์„œ๋Š” ์˜ˆ์ „์— ๋ฐฐ์šด...

Lecture 14: Reinforcement Learning

3 minute read

์ด๋ฒˆ ๋‹จ์›์—์„œ๋Š” ๊ฐ•ํ™” ํ•™์Šต(reinforcement learning)์— ๋Œ€ํ•ด์„œ ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Lecture 13: Generative Models

5 minute read

์ด๋ฒˆ ๋‹จ์›์—์„œ๋Š” ํ•™์Šต๋œ ์ด๋ฏธ์ง€๋กœ ์ƒˆ๋กœ์šด ์ด๋ฏธ์ง€๋ฅผ ๋งŒ๋“ค์–ด๋‚ด๋Š”(generate) ์•Œ๊ณ ๋ฆฌ์ฆ˜์— ๋Œ€ํ•ด ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค. ์ด์™€ ๊ฐ™์ด ์ƒˆ๋กœ์šด ์ด๋ฏธ์ง€๋ฅผ ๋งŒ๋“œ๋Š” ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ์ ์šฉํ•œ ๋ชจ๋ธ์„ Generative Model ์ด๋ผ๊ณ  ํ•œ๋‹ค.

Lecture 12: Visualizing and Understanding CNN

5 minute read

์ง€๊ธˆ๊นŒ์ง€ CNN์˜ ์ž‘๋™ ์›๋ฆฌ์™€ CNN์„ ํ†ตํ•ด ์ปดํ“จํ„ฐ๊ฐ€ ์ด๋ฏธ์ง€๋ฅผ ์ฝ๊ณ  ์ฒ˜๋ฆฌํ•˜๋Š” ๊ณผ์ •์„ ์•Œ์•„๋ณด์•˜๋‹ค. ๊ทธ๋ ‡๋‹ค๋ฉด CNN์˜ ๊ฐ ์ธต์—์„œ ์ด๋ฏธ์ง€๊ฐ€ ์–ด๋– ํ•œ ํ˜•์‹์œผ๋กœ ๋‚˜ํƒ€๋‚˜๊ฒŒ ๋ ๊นŒ? CNN์˜ ์ž‘๋™ ๊ณผ์ •์„ ์ข€ ๋” ์ง๊ด€์ ์œผ๋กœ ์•Œ๊ธฐ ์œ„ํ•ด ์‹œ๊ฐํ™” ํ•˜๋Š” ๋ฐฉ๋ฒ•์— ๋Œ€ํ•ด ์ด ๋‹จ์›์€ ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Lecture 11: Detection and Segementation

less than 1 minute read

ํ˜„์žฌ๊นŒ์ง„ ์ด๋ฏธ์ง€์— ๋‚˜ํƒ€๋‚œ ๋ฌผ์ฒด๋ฅผ ํŠน์ • ์นดํ…Œ๊ณ ๋ฆฌ๋กœ ๋ถ„๋ฅ˜ํ•˜๋Š” ์ด๋ฏธ์ง€ ๋ถ„๋ฅ˜(image classification) ์— ๋Œ€ํ•ด ์ง‘์ค‘์ ์œผ๋กœ ๋ฐฐ์›Œ๋ณด์•˜๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์ปดํ“จํ„ฐ ๋น„์ „์€ ๋ถ„ํ•  (segmentation), ํƒ์ง€ (detection)๋„ ์ˆ˜ํ–‰ํ•  ์ˆ˜ ์žˆ๋‹ค. ๊ฐ๊ฐ์˜ ์—ญํ• ์— ๋Œ€ํ•ด ๊ฐ„๋‹จํ•˜๊ฒŒ ์•Œ์•„๋ณด์ž

๋ฐ์ด์ฝ˜ ๊ฒฝ์ง„๋Œ€ํšŒ ์ฐธ์—ฌํ›„๊ธฐ part2

1 minute read

๊ฒฝ์ง„๋Œ€ํšŒ์—์„œ 1๋“ฑ์ด ์‚ฌ์šฉํ•œ ์ฝ”๋“œ๋ฅผ ๋ณด๋ ค๊ณ  ํ–ˆ๋Š”๋ฐ ์•„์‰ฝ๊ฒŒ 1๋“ฑํ•œ ํŒ€์ด ์ฝ”๋“œ๋ฅผ ๊ณต์œ ํ•˜์ง€ ์•Š์•„ 3๋“ฑ ํŒ€์˜ ์ฝ”๋“œ๋ฅผ ์ฐธ๊ณ ํ•˜์˜€๋‹ค.

๋ฐ์ด์ฝ˜ ๊ฒฝ์ง„๋Œ€ํšŒ ์ฐธ์—ฌํ›„๊ธฐ part1

4 minute read

์ง€๊ธˆ๊นŒ์ง€์˜ ๋ฐฐ์šด ๋‚ด์šฉ์„ ์‹ค์ „์— ํ•œ๋ฒˆ ์ ์šฉํ•ด๋ณด๊ณ  ์‹ถ๋‹ค๋Š” ์ƒ๊ฐ์— ๋™์•„๋ฆฌ ์Šคํ„ฐ๋””์›๋“ค๊ณผ ํ•จ๊ป˜ ๋ฐ์ด์ฝ˜ ๊ฒฝ์ง„๋Œ€ํšŒ์— ์ฐธ์—ฌ ํ•˜์˜€๋‹ค. ๋Œ€ํšŒ์˜ ๋‚ด์šฉ์€ ๊ตฌ๋‚ด์‹๋‹น ์‹์ˆ˜ ์ธ์›์„ ์˜ˆ์ธกํ•˜๋Š” ๊ฒƒ์œผ๋กœ, ๋Œ€ํšŒ ๊ธฐ๊ฐ„์€ 6์›” 3์ผ๋ถ€ํ„ฐ 7์›” 23์ผ๊นŒ์ง€์˜€์œผ๋ฉฐ ๋Œ€ํšŒ ์ข…๋ฃŒํ•œ์ง€๋Š” 2๋‹ฌ(โ€ฆ)์ด ๋„˜๊ฒŒ ์ง€๋‚œ ์ง€๊ธˆ ๊ฒฝ์ง„๋Œ€ํšŒ ์ฐธ์—ฌ ํ›„...

Chapter 4: ๋ชจ๋ธ ๊ตฌ์ถ• part2

4 minute read

๋ณธ ์ฑ•ํ„ฐ์˜ ๋‘ ๋ฒˆ์งธ ํŒŒํŠธ๋Š” ์‹ ๊ฒฝ๋ง, ์„ ํ˜• ๋ชจ๋ธ, ๋žœ๋ค ํฌ๋ ˆ์ŠคํŠธ์™€ ๊ธฐํƒ€ ๋ชจ๋ธ๋“ฑ์— ๋Œ€ํ•ด ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Chapter 4: ๋ชจ๋ธ ๊ตฌ์ถ• part1

5 minute read

์ฑ•ํ„ฐ4๋Š” ๋ฐ์ดํ„ฐ์˜ ์ƒ์„ฑ๋œ ํŠน์ง•์„ ๋ฐ”ํƒ•์œผ๋กœ ๋ชจ๋ธ์„ ๊ตฌ์ถ•ํ•˜๋Š” ๋ฐฉ๋ฒ•์— ๋Œ€ํ•ด ์„ค๋ช…ํ•˜๊ณ  ์žˆ๋‹ค. ์ด ์žฅ ๋˜ํ•œ ์–‘์ด ์ƒ๋‹นํžˆ ๋งŽ๊ณ  ์ค‘์š”ํ•œ ๋‚ด์šฉ์„ ๋‹ด๊ณ  ์žˆ๊ธฐ ๋•Œ๋ฌธ์— ๋‘ ๊ฐœ์˜ ํŒŒํŠธ๋กœ ๋‚˜๋ˆ ์„œ ์ •๋ฆฌํ•˜๊ธฐ๋กœ ํ•˜์˜€๋‹ค.

Chapter 3: ํŠน์ง• ์ƒ์„ฑ part2

8 minute read

์ฑ•ํ„ฐ 2์˜ ๋‘๋ฒˆ์งธ ํŒŒํŠธ๋Š” ์‹œ๊ฐ„๊ณผ ๊ด€๋ จ๋œ ๋ฐ์ดํ„ฐ๋ฅผ ์ฒ˜๋ฆฌํ•˜๋Š” ๋ฐฉ๋ฒ•, ๋น„์ง€๋„ ํ•™์Šต ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ์ด์šฉํ•˜์—ฌ ํŠน์ง•์„ ์ƒ์„ฑํ•˜๋Š” ๋ฐฉ๋ฒ•์— ๋Œ€ํ•ด ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Chapter 3: ํŠน์ง• ์ƒ์„ฑ part1

11 minute read

์ฑ•ํ„ฐ 3๋Š” ๋‚ด์šฉ์ด ์ด์ „ ์ฑ•ํ„ฐ๋ณด๋‹ค ๋งŽ์œผ๋ฏ€๋กœ 2๊ฐœ์˜ ํŒŒํŠธ๋กœ ๋‚˜๋ˆ„์–ด์„œ ์ž‘์„ฑํ•˜์˜€๋‹ค. ์ฒซ๋ฒˆ์งธ ํŒŒํŠธ์—์„œ๋Š” ๋ชจ๋ธ์˜ ์„ฑ๋Šฅ์„ ๋†’์ด๊ธฐ ์œ„ํ•ด์„œ ์ƒˆ๋กœ์šด ํŠน์ง•์„ ๋งŒ๋“œ๋Š” ๋ฐฉ๋ฒ•์— ๋Œ€ํ•ด ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Lecture 2: Image Classification pipeline

2 minute read

์›๋ž˜๋Š” Lecture 3๋ฅผ ์ •๋ฆฌํ•˜๊ธฐ ์ „์— Lecture 2๋ฅผ ์ •๋ฆฌํ–ˆ์–ด์•ผ ํ–ˆ๋Š”๋ฐ ๋ฐ€๋ฆฐ Lecture์„ ๋“ฃ๋‹ค๊ฐ€ ์ด์ œ์„œ์•ผ ์ •๋ฆฌ๋ฅผ ํ•œ๋‹ค.

Back to Top ↑

algorithm

ํŠน๋ณ„ํ•œ ํŒŒ์ด์ฌ ํ•จ์ˆ˜

less than 1 minute read

ํŒŒ์ด์ฌ์—๋Š” ๋ณต์žกํ•œ ๊ณผ์ •์„ ๋‹จ์ˆœํ™” ์‹œ์ผœ์ฃผ๋Š” ํŠน๋ณ„ํ•œ ํ•จ์ˆ˜๋“ค์ด ๋งŽ์ด ์กด์žฌํ•œ๋‹ค. ์ด๋Ÿฌํ•œ ํ•จ์ˆ˜๋“ค์„ ์ •๋ฆฌํ•ด๋ณด์•˜๋‹ค.

์ •๋ ฌ

5 minute read

๋“ค์–ด๊ฐ€๊ธฐ ์ „์—

ํž™

4 minute read

๋“ค์–ด๊ฐ€๊ธฐ ์ „์—

ํ•ด์‹œ ํ…Œ์ด๋ธ”

5 minute read

๋“ค์–ด๊ฐ€๊ธฐ ์ „์— ํ•ด์‹œ: ํ•ด์‹œ ํ…Œ์ด๋ธ”์—์„œ ์ž…๋ ฅ๊ฐ’์„ ๊ณ ์ • ํฌ๊ธฐ ๊ฐ’์œผ๋กœ ๋งคํ•‘ํ•˜๋Š” ๊ณผ์ •์„ ํ•ด์‹ฑ์ด๋ผ๊ณ  ํ•œ๋‹ค. ํ•ด์‹œ ํ•จ์ˆ˜: ํ•ด์‹œ๋ฅผ ํ•˜๋Š”๋ฐ ์‚ฌ์šฉ๋˜๋Š” ํ•จ์ˆ˜. ์„ฑ๋Šฅ์ด ์ข‹์€ ํ•ด์‹œ ํ•จ์ˆ˜๋“ค์˜ ํŠน์ง•์€ ๋‹ค์Œ๊ณผ ๊ฐ™๋‹ค. ํ•ด์‹œ ํ•จ์ˆ˜ ๊ฐ’ ์ถฉ๋Œ์˜ ์ตœ์†Œํ™” ์‰ฝ๊ณ  ๋น ๋ฅธ ์—ฐ์‚ฐ ...

Back to Top ↑

computer vision

Lecture 4: Maximum Likelihood Learning

less than 1 minute read

4๋ฒˆ์งธ ๊ฐ•์˜๋Š” Maximum Likelihood Learning์— ๋Œ€ํ•ด ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Lecture 3: Autoregressive Model

less than 1 minute read

3๋ฒˆ์งธ ๊ฐ•์˜๋Š” Autoregressive Model์— ๋Œ€ํ•ด ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Lecture 1&2: Introduction and Background to Deep Generative Model

less than 1 minute read

์ฒซ๋ฒˆ์งธ์™€ ๋‘๋ฒˆ์งธ ๊ฐ•์˜๋Š” Deep Generative Model์— ๋Œ€ํ•œ ์†Œ๊ฐœ์™€ ๋ฐฑ๊ทธ๋ผ์šด๋“œ ์ง€์‹์— ๋Œ€ํ•ด์„œ ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Pytorch tools

less than 1 minute read

view() vs shape() vs permute() vs transpose(): click here

On Adaptive Attacks to Adversarial Example Defenses

2 minute read

ํ•ด๋‹น ๋…ผ๋ฌธ ๋ฆฌ๋ทฐ๋Š” arxiv์— ๊ฐœ์ œ๋œ โ€˜On Adaptive Attacks to Adversarial Example Defensesโ€™๋…ผ๋ฌธ๊ณผ ๋…ผ๋ฌธ์— ๋Œ€ํ•œ ๋ฆฌ๋ทฐ ์˜์ƒ์„ ์ฐธ๊ณ ํ•˜์—ฌ ์ž‘์„ฑํ•˜์˜€๋‹ค.

Assignment 1

1 minute read

Assignment 1์€ ํŒŒ์ด์ฌ ์ฝ”๋“œ๋ฅผ ์ง์ ‘ ์ž‘์„ฑํ•˜์—ฌ ์ด๋ฏธ์ง€ ๋ถ„๋ฅ˜ ๋ชจ๋ธ์„ ์ง์ ‘ ๋งŒ๋“œ๋Š” ๋‚ด์šฉ์ด๋‹ค. ์‚ฌ์šฉํ•˜๋Š” ๋ชจ๋ธ๋กœ๋Š” KNN, SVM, two-layer-net(deep learning network with one hidden layer) ์ด ์žˆ๋‹ค. ์ด ๋ชจ๋ธ๋“ค์— ๋Œ€ํ•ด์„œ๋Š” ์˜ˆ์ „์— ๋ฐฐ์šด...

Lecture 14: Reinforcement Learning

3 minute read

์ด๋ฒˆ ๋‹จ์›์—์„œ๋Š” ๊ฐ•ํ™” ํ•™์Šต(reinforcement learning)์— ๋Œ€ํ•ด์„œ ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Lecture 13: Generative Models

5 minute read

์ด๋ฒˆ ๋‹จ์›์—์„œ๋Š” ํ•™์Šต๋œ ์ด๋ฏธ์ง€๋กœ ์ƒˆ๋กœ์šด ์ด๋ฏธ์ง€๋ฅผ ๋งŒ๋“ค์–ด๋‚ด๋Š”(generate) ์•Œ๊ณ ๋ฆฌ์ฆ˜์— ๋Œ€ํ•ด ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค. ์ด์™€ ๊ฐ™์ด ์ƒˆ๋กœ์šด ์ด๋ฏธ์ง€๋ฅผ ๋งŒ๋“œ๋Š” ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ์ ์šฉํ•œ ๋ชจ๋ธ์„ Generative Model ์ด๋ผ๊ณ  ํ•œ๋‹ค.

Lecture 12: Visualizing and Understanding CNN

5 minute read

์ง€๊ธˆ๊นŒ์ง€ CNN์˜ ์ž‘๋™ ์›๋ฆฌ์™€ CNN์„ ํ†ตํ•ด ์ปดํ“จํ„ฐ๊ฐ€ ์ด๋ฏธ์ง€๋ฅผ ์ฝ๊ณ  ์ฒ˜๋ฆฌํ•˜๋Š” ๊ณผ์ •์„ ์•Œ์•„๋ณด์•˜๋‹ค. ๊ทธ๋ ‡๋‹ค๋ฉด CNN์˜ ๊ฐ ์ธต์—์„œ ์ด๋ฏธ์ง€๊ฐ€ ์–ด๋– ํ•œ ํ˜•์‹์œผ๋กœ ๋‚˜ํƒ€๋‚˜๊ฒŒ ๋ ๊นŒ? CNN์˜ ์ž‘๋™ ๊ณผ์ •์„ ์ข€ ๋” ์ง๊ด€์ ์œผ๋กœ ์•Œ๊ธฐ ์œ„ํ•ด ์‹œ๊ฐํ™” ํ•˜๋Š” ๋ฐฉ๋ฒ•์— ๋Œ€ํ•ด ์ด ๋‹จ์›์€ ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Lecture 11: Detection and Segementation

less than 1 minute read

ํ˜„์žฌ๊นŒ์ง„ ์ด๋ฏธ์ง€์— ๋‚˜ํƒ€๋‚œ ๋ฌผ์ฒด๋ฅผ ํŠน์ • ์นดํ…Œ๊ณ ๋ฆฌ๋กœ ๋ถ„๋ฅ˜ํ•˜๋Š” ์ด๋ฏธ์ง€ ๋ถ„๋ฅ˜(image classification) ์— ๋Œ€ํ•ด ์ง‘์ค‘์ ์œผ๋กœ ๋ฐฐ์›Œ๋ณด์•˜๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์ปดํ“จํ„ฐ ๋น„์ „์€ ๋ถ„ํ•  (segmentation), ํƒ์ง€ (detection)๋„ ์ˆ˜ํ–‰ํ•  ์ˆ˜ ์žˆ๋‹ค. ๊ฐ๊ฐ์˜ ์—ญํ• ์— ๋Œ€ํ•ด ๊ฐ„๋‹จํ•˜๊ฒŒ ์•Œ์•„๋ณด์ž

Lecture 2: Image Classification pipeline

2 minute read

์›๋ž˜๋Š” Lecture 3๋ฅผ ์ •๋ฆฌํ•˜๊ธฐ ์ „์— Lecture 2๋ฅผ ์ •๋ฆฌํ–ˆ์–ด์•ผ ํ–ˆ๋Š”๋ฐ ๋ฐ€๋ฆฐ Lecture์„ ๋“ฃ๋‹ค๊ฐ€ ์ด์ œ์„œ์•ผ ์ •๋ฆฌ๋ฅผ ํ•œ๋‹ค.

Back to Top ↑

Paper Review

Robust Speech Recognition via Large-Scale Weak Supervision

less than 1 minute read

์ด๋ฒˆ์— ๋ฐœํ‘œํ•œ ๋…ผ๋ฌธ์˜ ์ œ๋ชฉ์€ Robust Speech Recognition via Large-Scale Weak Supervision ์œผ๋กœ, ์ผ๋ช… Whisper์— ๋Œ€ํ•œ ๋…ผ๋ฌธ์ด๋‹ค.

An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale

less than 1 minute read

์ด๋ฒˆ ๋ฐœํ‘œ ๋…ผ๋ฌธ์˜ ์ œ๋ชฉ์€ An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale์œผ๋กœ, ์ผ๋ช… Visual Transformer ๋ฐฉ๋ฒ•์„ ์ƒˆ๋กญ๊ฒŒ ์ œ์‹œํ•œ ๋…ผ๋ฌธ์ด๋‹ค.

Enriching Word Vectors with Subword Information

less than 1 minute read

์ด๋ฒˆ ๋ฐœํ‘œ ๋…ผ๋ฌธ์˜ ์ œ๋ชฉ์€ Enriching Word Vectors with Subword Information์œผ๋กœ, ์ผ๋ช… FastText ๋ฐฉ๋ฒ•์„ ์ƒˆ๋กญ๊ฒŒ ์ œ์‹œํ•œ ๋…ผ๋ฌธ์ด๋‹ค.

On Adaptive Attacks to Adversarial Example Defenses

2 minute read

ํ•ด๋‹น ๋…ผ๋ฌธ ๋ฆฌ๋ทฐ๋Š” arxiv์— ๊ฐœ์ œ๋œ โ€˜On Adaptive Attacks to Adversarial Example Defensesโ€™๋…ผ๋ฌธ๊ณผ ๋…ผ๋ฌธ์— ๋Œ€ํ•œ ๋ฆฌ๋ทฐ ์˜์ƒ์„ ์ฐธ๊ณ ํ•˜์—ฌ ์ž‘์„ฑํ•˜์˜€๋‹ค.

Back to Top ↑

cs231n

Assignment 1

1 minute read

Assignment 1์€ ํŒŒ์ด์ฌ ์ฝ”๋“œ๋ฅผ ์ง์ ‘ ์ž‘์„ฑํ•˜์—ฌ ์ด๋ฏธ์ง€ ๋ถ„๋ฅ˜ ๋ชจ๋ธ์„ ์ง์ ‘ ๋งŒ๋“œ๋Š” ๋‚ด์šฉ์ด๋‹ค. ์‚ฌ์šฉํ•˜๋Š” ๋ชจ๋ธ๋กœ๋Š” KNN, SVM, two-layer-net(deep learning network with one hidden layer) ์ด ์žˆ๋‹ค. ์ด ๋ชจ๋ธ๋“ค์— ๋Œ€ํ•ด์„œ๋Š” ์˜ˆ์ „์— ๋ฐฐ์šด...

Lecture 14: Reinforcement Learning

3 minute read

์ด๋ฒˆ ๋‹จ์›์—์„œ๋Š” ๊ฐ•ํ™” ํ•™์Šต(reinforcement learning)์— ๋Œ€ํ•ด์„œ ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Lecture 13: Generative Models

5 minute read

์ด๋ฒˆ ๋‹จ์›์—์„œ๋Š” ํ•™์Šต๋œ ์ด๋ฏธ์ง€๋กœ ์ƒˆ๋กœ์šด ์ด๋ฏธ์ง€๋ฅผ ๋งŒ๋“ค์–ด๋‚ด๋Š”(generate) ์•Œ๊ณ ๋ฆฌ์ฆ˜์— ๋Œ€ํ•ด ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค. ์ด์™€ ๊ฐ™์ด ์ƒˆ๋กœ์šด ์ด๋ฏธ์ง€๋ฅผ ๋งŒ๋“œ๋Š” ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ์ ์šฉํ•œ ๋ชจ๋ธ์„ Generative Model ์ด๋ผ๊ณ  ํ•œ๋‹ค.

Lecture 12: Visualizing and Understanding CNN

5 minute read

์ง€๊ธˆ๊นŒ์ง€ CNN์˜ ์ž‘๋™ ์›๋ฆฌ์™€ CNN์„ ํ†ตํ•ด ์ปดํ“จํ„ฐ๊ฐ€ ์ด๋ฏธ์ง€๋ฅผ ์ฝ๊ณ  ์ฒ˜๋ฆฌํ•˜๋Š” ๊ณผ์ •์„ ์•Œ์•„๋ณด์•˜๋‹ค. ๊ทธ๋ ‡๋‹ค๋ฉด CNN์˜ ๊ฐ ์ธต์—์„œ ์ด๋ฏธ์ง€๊ฐ€ ์–ด๋– ํ•œ ํ˜•์‹์œผ๋กœ ๋‚˜ํƒ€๋‚˜๊ฒŒ ๋ ๊นŒ? CNN์˜ ์ž‘๋™ ๊ณผ์ •์„ ์ข€ ๋” ์ง๊ด€์ ์œผ๋กœ ์•Œ๊ธฐ ์œ„ํ•ด ์‹œ๊ฐํ™” ํ•˜๋Š” ๋ฐฉ๋ฒ•์— ๋Œ€ํ•ด ์ด ๋‹จ์›์€ ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Lecture 11: Detection and Segementation

less than 1 minute read

ํ˜„์žฌ๊นŒ์ง„ ์ด๋ฏธ์ง€์— ๋‚˜ํƒ€๋‚œ ๋ฌผ์ฒด๋ฅผ ํŠน์ • ์นดํ…Œ๊ณ ๋ฆฌ๋กœ ๋ถ„๋ฅ˜ํ•˜๋Š” ์ด๋ฏธ์ง€ ๋ถ„๋ฅ˜(image classification) ์— ๋Œ€ํ•ด ์ง‘์ค‘์ ์œผ๋กœ ๋ฐฐ์›Œ๋ณด์•˜๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์ปดํ“จํ„ฐ ๋น„์ „์€ ๋ถ„ํ•  (segmentation), ํƒ์ง€ (detection)๋„ ์ˆ˜ํ–‰ํ•  ์ˆ˜ ์žˆ๋‹ค. ๊ฐ๊ฐ์˜ ์—ญํ• ์— ๋Œ€ํ•ด ๊ฐ„๋‹จํ•˜๊ฒŒ ์•Œ์•„๋ณด์ž

Lecture 2: Image Classification pipeline

2 minute read

์›๋ž˜๋Š” Lecture 3๋ฅผ ์ •๋ฆฌํ•˜๊ธฐ ์ „์— Lecture 2๋ฅผ ์ •๋ฆฌํ–ˆ์–ด์•ผ ํ–ˆ๋Š”๋ฐ ๋ฐ€๋ฆฐ Lecture์„ ๋“ฃ๋‹ค๊ฐ€ ์ด์ œ์„œ์•ผ ์ •๋ฆฌ๋ฅผ ํ•œ๋‹ค.

Back to Top ↑

data analysis

์„œ์šธ์‹œ ๋น…๋ฐ์ดํ„ฐ์บ ํผ์Šค ๊ณต๋ชจ์ „

4 minute read

2021๋…„ 9์›” 13์ผ๋ถ€ํ„ฐ 10์›” 21์ผ๊นŒ์ง€ ์ง„ํ–‰ํ•œ ์„œ์šธ์‹œ ๋น…๋ฐ์ดํ„ฐ์บ ํผ์Šค ๊ณต๋ชจ์ „์— ์ฐธ๊ฐ€ํ•˜์˜€๋‹ค. ๊ณต๋ชจ์ „์ด ๋๋‚œ์ง€๋Š” ๋ฌด๋ ค ํ•œ๋‹ฌ(โ€ฆ)์ด ์ง€๋‚ฌ์ง€๋งŒ ํ›„๊ธฐ ์ž‘์„ฑ์„ ๋ฏธ๋ฃจ๋‹ค๊ฐ€ ์ง€๊ธˆ์—์„œ์•ผ ํ›„๊ธฐ๋ฅผ ์ž‘์„ฑํ•˜๋ ค ํ•œ๋‹ค.

๋ฐ์ด์ฝ˜ ๊ฒฝ์ง„๋Œ€ํšŒ ์ฐธ์—ฌํ›„๊ธฐ part2

1 minute read

๊ฒฝ์ง„๋Œ€ํšŒ์—์„œ 1๋“ฑ์ด ์‚ฌ์šฉํ•œ ์ฝ”๋“œ๋ฅผ ๋ณด๋ ค๊ณ  ํ–ˆ๋Š”๋ฐ ์•„์‰ฝ๊ฒŒ 1๋“ฑํ•œ ํŒ€์ด ์ฝ”๋“œ๋ฅผ ๊ณต์œ ํ•˜์ง€ ์•Š์•„ 3๋“ฑ ํŒ€์˜ ์ฝ”๋“œ๋ฅผ ์ฐธ๊ณ ํ•˜์˜€๋‹ค.

๋ฐ์ด์ฝ˜ ๊ฒฝ์ง„๋Œ€ํšŒ ์ฐธ์—ฌํ›„๊ธฐ part1

4 minute read

์ง€๊ธˆ๊นŒ์ง€์˜ ๋ฐฐ์šด ๋‚ด์šฉ์„ ์‹ค์ „์— ํ•œ๋ฒˆ ์ ์šฉํ•ด๋ณด๊ณ  ์‹ถ๋‹ค๋Š” ์ƒ๊ฐ์— ๋™์•„๋ฆฌ ์Šคํ„ฐ๋””์›๋“ค๊ณผ ํ•จ๊ป˜ ๋ฐ์ด์ฝ˜ ๊ฒฝ์ง„๋Œ€ํšŒ์— ์ฐธ์—ฌ ํ•˜์˜€๋‹ค. ๋Œ€ํšŒ์˜ ๋‚ด์šฉ์€ ๊ตฌ๋‚ด์‹๋‹น ์‹์ˆ˜ ์ธ์›์„ ์˜ˆ์ธกํ•˜๋Š” ๊ฒƒ์œผ๋กœ, ๋Œ€ํšŒ ๊ธฐ๊ฐ„์€ 6์›” 3์ผ๋ถ€ํ„ฐ 7์›” 23์ผ๊นŒ์ง€์˜€์œผ๋ฉฐ ๋Œ€ํšŒ ์ข…๋ฃŒํ•œ์ง€๋Š” 2๋‹ฌ(โ€ฆ)์ด ๋„˜๊ฒŒ ์ง€๋‚œ ์ง€๊ธˆ ๊ฒฝ์ง„๋Œ€ํšŒ ์ฐธ์—ฌ ํ›„...

Chapter 4: ๋ชจ๋ธ ๊ตฌ์ถ• part2

4 minute read

๋ณธ ์ฑ•ํ„ฐ์˜ ๋‘ ๋ฒˆ์งธ ํŒŒํŠธ๋Š” ์‹ ๊ฒฝ๋ง, ์„ ํ˜• ๋ชจ๋ธ, ๋žœ๋ค ํฌ๋ ˆ์ŠคํŠธ์™€ ๊ธฐํƒ€ ๋ชจ๋ธ๋“ฑ์— ๋Œ€ํ•ด ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Chapter 4: ๋ชจ๋ธ ๊ตฌ์ถ• part1

5 minute read

์ฑ•ํ„ฐ4๋Š” ๋ฐ์ดํ„ฐ์˜ ์ƒ์„ฑ๋œ ํŠน์ง•์„ ๋ฐ”ํƒ•์œผ๋กœ ๋ชจ๋ธ์„ ๊ตฌ์ถ•ํ•˜๋Š” ๋ฐฉ๋ฒ•์— ๋Œ€ํ•ด ์„ค๋ช…ํ•˜๊ณ  ์žˆ๋‹ค. ์ด ์žฅ ๋˜ํ•œ ์–‘์ด ์ƒ๋‹นํžˆ ๋งŽ๊ณ  ์ค‘์š”ํ•œ ๋‚ด์šฉ์„ ๋‹ด๊ณ  ์žˆ๊ธฐ ๋•Œ๋ฌธ์— ๋‘ ๊ฐœ์˜ ํŒŒํŠธ๋กœ ๋‚˜๋ˆ ์„œ ์ •๋ฆฌํ•˜๊ธฐ๋กœ ํ•˜์˜€๋‹ค.

Chapter 3: ํŠน์ง• ์ƒ์„ฑ part2

8 minute read

์ฑ•ํ„ฐ 2์˜ ๋‘๋ฒˆ์งธ ํŒŒํŠธ๋Š” ์‹œ๊ฐ„๊ณผ ๊ด€๋ จ๋œ ๋ฐ์ดํ„ฐ๋ฅผ ์ฒ˜๋ฆฌํ•˜๋Š” ๋ฐฉ๋ฒ•, ๋น„์ง€๋„ ํ•™์Šต ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ์ด์šฉํ•˜์—ฌ ํŠน์ง•์„ ์ƒ์„ฑํ•˜๋Š” ๋ฐฉ๋ฒ•์— ๋Œ€ํ•ด ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Chapter 3: ํŠน์ง• ์ƒ์„ฑ part1

11 minute read

์ฑ•ํ„ฐ 3๋Š” ๋‚ด์šฉ์ด ์ด์ „ ์ฑ•ํ„ฐ๋ณด๋‹ค ๋งŽ์œผ๋ฏ€๋กœ 2๊ฐœ์˜ ํŒŒํŠธ๋กœ ๋‚˜๋ˆ„์–ด์„œ ์ž‘์„ฑํ•˜์˜€๋‹ค. ์ฒซ๋ฒˆ์งธ ํŒŒํŠธ์—์„œ๋Š” ๋ชจ๋ธ์˜ ์„ฑ๋Šฅ์„ ๋†’์ด๊ธฐ ์œ„ํ•ด์„œ ์ƒˆ๋กœ์šด ํŠน์ง•์„ ๋งŒ๋“œ๋Š” ๋ฐฉ๋ฒ•์— ๋Œ€ํ•ด ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Back to Top ↑

Natural Language Processing

Robust Speech Recognition via Large-Scale Weak Supervision

less than 1 minute read

์ด๋ฒˆ์— ๋ฐœํ‘œํ•œ ๋…ผ๋ฌธ์˜ ์ œ๋ชฉ์€ Robust Speech Recognition via Large-Scale Weak Supervision ์œผ๋กœ, ์ผ๋ช… Whisper์— ๋Œ€ํ•œ ๋…ผ๋ฌธ์ด๋‹ค.

An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale

less than 1 minute read

์ด๋ฒˆ ๋ฐœํ‘œ ๋…ผ๋ฌธ์˜ ์ œ๋ชฉ์€ An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale์œผ๋กœ, ์ผ๋ช… Visual Transformer ๋ฐฉ๋ฒ•์„ ์ƒˆ๋กญ๊ฒŒ ์ œ์‹œํ•œ ๋…ผ๋ฌธ์ด๋‹ค.

Masked Language Model vs Causal Language Model

1 minute read

BERT์™€ ๊ฐ™์€ Masked Language Model๊ณผ GPT์™€ ๊ฐ™์€ Causal Language Model์˜ ์ฐจ์ด์ ์„ ์ •๋ฆฌํ•ด๋ณด์•˜๋‹ค. ํ•ด๋‹น ๋งํฌ๋ฅผ ์ฐธ๊ณ ํ•˜์—ฌ ์ž‘์„ฑํ•˜์˜€๋‹ค.

Enriching Word Vectors with Subword Information

less than 1 minute read

์ด๋ฒˆ ๋ฐœํ‘œ ๋…ผ๋ฌธ์˜ ์ œ๋ชฉ์€ Enriching Word Vectors with Subword Information์œผ๋กœ, ์ผ๋ช… FastText ๋ฐฉ๋ฒ•์„ ์ƒˆ๋กญ๊ฒŒ ์ œ์‹œํ•œ ๋…ผ๋ฌธ์ด๋‹ค.

Back to Top ↑

project

์ œ4ํšŒ AI x Bookathon ์ฐธ์—ฌํ›„๊ธฐ

3 minute read

โ€ƒ ์ง€๋‚œ 3์ฃผ(22.12.27~ 23.1.18) ๋™์•ˆ ์ง„ํ–‰๋˜์—ˆ๋˜ AI x Bookathon ๋Œ€ํšŒ์— ๋Œ€ํ•œ ํ›„๊ธฐ๋ฅผ ๋‚จ๊ธฐ๊ณ ์ž ํ•œ๋‹ค. ํ•ด๋‹น ๋Œ€ํšŒ๋Š” Language Model์„ ์ด์šฉํ•ด ์ˆ˜ํ•„์„ ์ž‘์„ฑํ•˜๋Š” ํ•ด์ปคํ†ค ๋Œ€ํšŒ๋กœ, ์˜ˆ์„ ๊ณผ ๋ณธ์„ ์œผ๋กœ ์ด๋ฃจ์–ด์ ธ ์žˆ์—ˆ๋‹ค. ์˜ˆ์„ ์€ ๋”ฅ๋Ÿฌ๋‹์˜ ๊ธฐ์ดˆ ์ง€์‹๋“ค์— ๋Œ€ํ•œ ํ€ด์ฆˆ...

์„œ์šธ์‹œ ๋น…๋ฐ์ดํ„ฐ์บ ํผ์Šค ๊ณต๋ชจ์ „

4 minute read

2021๋…„ 9์›” 13์ผ๋ถ€ํ„ฐ 10์›” 21์ผ๊นŒ์ง€ ์ง„ํ–‰ํ•œ ์„œ์šธ์‹œ ๋น…๋ฐ์ดํ„ฐ์บ ํผ์Šค ๊ณต๋ชจ์ „์— ์ฐธ๊ฐ€ํ•˜์˜€๋‹ค. ๊ณต๋ชจ์ „์ด ๋๋‚œ์ง€๋Š” ๋ฌด๋ ค ํ•œ๋‹ฌ(โ€ฆ)์ด ์ง€๋‚ฌ์ง€๋งŒ ํ›„๊ธฐ ์ž‘์„ฑ์„ ๋ฏธ๋ฃจ๋‹ค๊ฐ€ ์ง€๊ธˆ์—์„œ์•ผ ํ›„๊ธฐ๋ฅผ ์ž‘์„ฑํ•˜๋ ค ํ•œ๋‹ค.

๋ฐ์ด์ฝ˜ ๊ฒฝ์ง„๋Œ€ํšŒ ์ฐธ์—ฌํ›„๊ธฐ part2

1 minute read

๊ฒฝ์ง„๋Œ€ํšŒ์—์„œ 1๋“ฑ์ด ์‚ฌ์šฉํ•œ ์ฝ”๋“œ๋ฅผ ๋ณด๋ ค๊ณ  ํ–ˆ๋Š”๋ฐ ์•„์‰ฝ๊ฒŒ 1๋“ฑํ•œ ํŒ€์ด ์ฝ”๋“œ๋ฅผ ๊ณต์œ ํ•˜์ง€ ์•Š์•„ 3๋“ฑ ํŒ€์˜ ์ฝ”๋“œ๋ฅผ ์ฐธ๊ณ ํ•˜์˜€๋‹ค.

๋ฐ์ด์ฝ˜ ๊ฒฝ์ง„๋Œ€ํšŒ ์ฐธ์—ฌํ›„๊ธฐ part1

4 minute read

์ง€๊ธˆ๊นŒ์ง€์˜ ๋ฐฐ์šด ๋‚ด์šฉ์„ ์‹ค์ „์— ํ•œ๋ฒˆ ์ ์šฉํ•ด๋ณด๊ณ  ์‹ถ๋‹ค๋Š” ์ƒ๊ฐ์— ๋™์•„๋ฆฌ ์Šคํ„ฐ๋””์›๋“ค๊ณผ ํ•จ๊ป˜ ๋ฐ์ด์ฝ˜ ๊ฒฝ์ง„๋Œ€ํšŒ์— ์ฐธ์—ฌ ํ•˜์˜€๋‹ค. ๋Œ€ํšŒ์˜ ๋‚ด์šฉ์€ ๊ตฌ๋‚ด์‹๋‹น ์‹์ˆ˜ ์ธ์›์„ ์˜ˆ์ธกํ•˜๋Š” ๊ฒƒ์œผ๋กœ, ๋Œ€ํšŒ ๊ธฐ๊ฐ„์€ 6์›” 3์ผ๋ถ€ํ„ฐ 7์›” 23์ผ๊นŒ์ง€์˜€์œผ๋ฉฐ ๋Œ€ํšŒ ์ข…๋ฃŒํ•œ์ง€๋Š” 2๋‹ฌ(โ€ฆ)์ด ๋„˜๊ฒŒ ์ง€๋‚œ ์ง€๊ธˆ ๊ฒฝ์ง„๋Œ€ํšŒ ์ฐธ์—ฌ ํ›„...

Back to Top ↑

cs236

Lecture 4: Maximum Likelihood Learning

less than 1 minute read

4๋ฒˆ์งธ ๊ฐ•์˜๋Š” Maximum Likelihood Learning์— ๋Œ€ํ•ด ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Lecture 3: Autoregressive Model

less than 1 minute read

3๋ฒˆ์งธ ๊ฐ•์˜๋Š” Autoregressive Model์— ๋Œ€ํ•ด ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Lecture 1&2: Introduction and Background to Deep Generative Model

less than 1 minute read

์ฒซ๋ฒˆ์งธ์™€ ๋‘๋ฒˆ์งธ ๊ฐ•์˜๋Š” Deep Generative Model์— ๋Œ€ํ•œ ์†Œ๊ฐœ์™€ ๋ฐฑ๊ทธ๋ผ์šด๋“œ ์ง€์‹์— ๋Œ€ํ•ด์„œ ๋‹ค๋ฃจ๊ณ  ์žˆ๋‹ค.

Back to Top ↑

markdown

๋งˆํฌ๋‹ค์šด(Markdown) ๋ฌธ๋ฒ•

2 minute read

ํ‹ฐ์Šคํ† ๋ฆฌ์—์„œ ๊นƒํ—ˆ๋ธŒ ๋ธ”๋กœ๊ทธ๋กœ ๋ฐ”๊พธ๋ฉด์„œ ๋งˆํฌ๋‹ค์šด์— ๋Œ€ํ•ด ์ฒ˜์Œ ์ ‘ํ•˜๊ฒŒ ๋˜์—ˆ๋‹ค. ์›ํ• ํ•œ ๋ธ”๋กœ๊ทธ ์ •๋ฆฌ๋ฅผ ์œ„ํ•ด ๋งˆํฌ๋‹ค์šด ๋ฌธ๋ฒ•์— ๋Œ€ํ•ด ๋ฐฐ์šธ ํ•„์š”์„ฑ์ด ์žˆ๋‹ค๊ณ  ์ƒ๊ฐํ•˜์—ฌ ์ด๋ ‡๊ฒŒ ์ •๋ฆฌ๋ฅผ ์ง„ํ–‰ ์ค‘์ด๋‹ค.

Back to Top ↑

Computer Vision

Back to Top ↑

data engineering

Back to Top ↑