소득공제
- 베스트
- 인공지능 top100 2주
이미 소장하고 있다면 판매해 보세요.
- 국내배송만 가능
- 최저가 보상
- 문화비소득공제 신청가능
이 상품의 시리즈 2
이 상품의 시리즈 알림신청
제이펍-I♥A.I 시리즈
이 상품의 시리즈 알림신청
카드뉴스로 보는 책
상세 이미지
책소개
목차
|
PART V 기계학습
CHAPTER 19 견본에서 배우는 학습 3 19.1 학습의 여러 형태 ······················································································· 4 19.2 지도학습 ····································································································· 6 19.3 결정 트리의 학습 ····················································································· 11 19.4 모형 선택과 최적화 ················································································· 21 19.5 학습 이론 ································································································· 30 19.6 선형 회귀와 분류 ····················································································· 35 19.7 비매개변수 모형 ······················································································· 47 19.8 앙상블 학습 ······························································································ 59 19.9 기계학습 시스템 개발 ·············································································· 69 요약 ···································································································· 81 참고문헌 및 역사적 참고사항 ··························································· 82 CHAPTER 20 확률 모형의 학습 89 20.1 통계적 학습 ······························································································ 90 20.2 완전 데이터를 이용한 학습 ····································································· 93 20.3 은닉 변수가 있는 학습: EM 알고리즘 ················································· 109 요약 ·································································································· 119 참고문헌 및 역사적 참고사항 ························································· 120 CHAPTER 21 심층학습 125 21.1 단순 순방향 신경망 ··············································································· 127 21.2 심층학습을 위한 계산 그래프 ······························································· 133 21.3 합성곱 신경망 ························································································ 137 21.4 학습 알고리즘 ························································································ 144 21.5 일반화 ····································································································· 148 21.6 순환 신경망 ···························································································· 153 21.7 비지도학습과 전이학습 ·········································································· 157 21.8 응용 ········································································································ 165 요약 ·································································································· 168 참고문헌 및 역사적 참고사항 ························································· 168 CHAPTER 22 강화학습 173 22.1 보상 기반 학습 ······················································································ 173 22.2 수동 강화학습 ························································································ 176 22.3 능동 강화학습 ························································································ 183 22.4 강화학습의 일반화 ················································································· 191 22.5 정책 검색 ······························································································· 199 22.6 견습 학습과 역강화학습 ········································································ 202 22.7 강화학습의 응용 ····················································································· 206 요약 ·································································································· 209 참고문헌 및 역사적 참고사항 ························································· 211 PART VI 의사소통, 지각, 행동 CHAPTER 23 자연어 처리 217 23.1 언어 모형 ······························································································· 218 23.2 문법 ········································································································ 231 23.3 파싱 ········································································································ 233 23.4 증강 문법 ······························································································· 240 23.5 실제 자연어의 복잡한 사항들 ······························································· 246 23.6 자연어 처리 과제들 ··············································································· 250 요약 ·································································································· 252 참고문헌 및 역사적 참고사항 ························································· 253 CHAPTER 24 자연어 처리를 위한 심층학습 259 24.1 단어 내장 ······························································································· 260 24.2 NLP를 위한 순환 신경망 ······································································ 264 24.3 순차열 대 순차열 모형 ·········································································· 268 24.4 트랜스포머 구조 ····················································································· 274 24.5 사전훈련과 전이학습 ·············································································· 277 24.6 현황 ········································································································ 282 요약 ·································································································· 285 참고문헌 및 역사적 참고사항 ························································· 285 CHAPTER 25 컴퓨터 시각 289 25.1 소개 ········································································································ 289 25.2 이미지 형성 ···························································································· 291 25.3 단순 이미지 특징 ··················································································· 298 25.4 이미지 분류 ···························································································· 306 25.5 물체 검출 ······························································································· 311 25.6 3차원 세계 ····························································································· 314 25.7 컴퓨터 시각의 용도 ··············································································· 319 요약 ·································································································· 334 참고문헌 및 역사적 참고사항 ························································· 335 CHAPTER 26 로봇공학 341 26.1 로봇 ······································································································ 341 26.2 로봇 하드웨어 ······················································································ 342 26.3 로봇공학이 푸는 문제들 ······································································ 347 26.4 로봇 지각 ····························································································· 349 26.5 계획 수립과 제어 ················································································· 357 26.6 불확실한 운동의 계획 ·········································································· 378 26.7 로봇공학의 강화학습 ············································································ 381 26.8 인간과 로봇 ·························································································· 384 26.9 로봇공학의 또 다른 틀 ········································································ 394 26.10 응용 영역 ····························································································· 397 요약 ································································································ 400 참고문헌 및 역사적 참고사항 ······················································· 402 PART VII 결론 CHAPTER 27 인공지능의 철학, 윤리학, 안전 411 27.1 인공지능의 한계 ····················································································· 411 27.2 기계가 정말로 생각할 수 있을까? ······················································· 416 27.3 인공지능의 윤리 ····················································································· 418 요약 ·································································································· 443 참고문헌 및 역사적 참고사항 ························································· 443 CHAPTER 28 인공지능의 미래 451 28.1 인공지능의 구성요소 ·············································································· 452 28.2 인공지능 구조 ························································································ 459 APPENDIX A 수학적 배경 465 A.1 복잡도 분석과 O( ) 표기법 ··································································· 465 A.2 벡터, 행렬, 선형대수 ············································································· 468 A.3 확률분포 ································································································· 470 참고문헌 및 역사적 참고사항 ························································· 473 APPENDIX B 언어와 알고리즘에 관해 475 B.1 BNF를 이용한 언어의 정의 ·································································· 475 B.2 알고리즘 서술에 쓰이는 의사코드 ························································ 476 B.3 온라인 보조 자료 ··················································································· 478 ? 참고문헌 ·················································· 479 ? 찾아보기 ·················································· 537 |
저자 소개3
저스튜어트 러셀
관심작가 알림신청Stuart Russell
버클리에 있는 캘리포니아대학교 컴퓨터과학 교수이자 공학 부문 스미스자데이 석좌교수. 옥스퍼드대학교 웨덤 칼리지에서 물리학을 공부하고 스탠퍼드대학교에서 컴퓨터과학으로 박사학위를 받았다. 기계 학습, 확률론적 추론, 실시간 의사 결정, 계산 생리학 및 철학적 기초를 포함한 인공지능의 광범위한 주제를 놓고 연구했고, 지금은 자율무기의 위협, 인공지능의 장기적 미래 및 인류와의 관계 등에도 관심을 두고 있다. 미국 인공지능협회, 컴퓨터학회, 미국과학진흥협회 회원이며, 세계경제포럼의 AI와 로봇학 위원회 부의장, 유엔 군축 문제 고문도 맡고 있다. 2016 서울디지털포럼, 2020 서울포
버클리에 있는 캘리포니아대학교 컴퓨터과학 교수이자 공학 부문 스미스자데이 석좌교수. 옥스퍼드대학교 웨덤 칼리지에서 물리학을 공부하고 스탠퍼드대학교에서 컴퓨터과학으로 박사학위를 받았다. 기계 학습, 확률론적 추론, 실시간 의사 결정, 계산 생리학 및 철학적 기초를 포함한 인공지능의 광범위한 주제를 놓고 연구했고, 지금은 자율무기의 위협, 인공지능의 장기적 미래 및 인류와의 관계 등에도 관심을 두고 있다. 미국 인공지능협회, 컴퓨터학회, 미국과학진흥협회 회원이며, 세계경제포럼의 AI와 로봇학 위원회 부의장, 유엔 군축 문제 고문도 맡고 있다. 2016 서울디지털포럼, 2020 서울포럼 등에서 강연하기도 했다.
구글 리서치 디렉터 피터 노빅과 함께 『인공지능: 현대적 접근방식』(1995)을 썼다. AI 분야의 결정판 교과서로 널리 인정받고 있는 『인공지능』(현재 4판)은 13개 언어로 번역되어 118개국, 1,500여 대학에서 교재로 사용되고 있다. 2016년에는 UC 버클리를 중심으로 여러 대학과 기관이 협력하는 연구기관 ‘휴먼컴패터블 AI센터’를 설립하여 AI 연구의 일반적인 추진 방향을 증명 가능하게 유익한 AI 시스템 쪽으로 재설정하는 데 필요한 개념적·기술적 도구를 개발해왔고, 그 결과물을 『어떻게 인간과 공존하는 인공지능을 만들 것인가: AI와 통제 문제』에 담았다.
구글 리서치 디렉터 피터 노빅과 함께 『인공지능: 현대적 접근방식』(1995)을 썼다. AI 분야의 결정판 교과서로 널리 인정받고 있는 『인공지능』(현재 4판)은 13개 언어로 번역되어 118개국, 1,500여 대학에서 교재로 사용되고 있다. 2016년에는 UC 버클리를 중심으로 여러 대학과 기관이 협력하는 연구기관 ‘휴먼컴패터블 AI센터’를 설립하여 AI 연구의 일반적인 추진 방향을 증명 가능하게 유익한 AI 시스템 쪽으로 재설정하는 데 필요한 개념적·기술적 도구를 개발해왔고, 그 결과물을 『어떻게 인간과 공존하는 인공지능을 만들 것인가: AI와 통제 문제』에 담았다.
스튜어트 러셀의 다른 상품
저피터 노빅
관심작가 알림신청Peter Norvig
구글의 연구실장이며, 2002년에서 2005년까지 핵심 웹 검색 엔진 개발을 이끌었다. 전에는 NASA Ames Research Center의 계산 과학 분과장으로서 NASA의 인공지능 및 로봇공학 연구와 개발을 감독했다. 서던 캘리포니아 대학교의 교수였으며, 버클리 대학교와 스탠퍼드 대학교의 연구교수단 일원이었다. 그의 다른 책으로는 《Paradigms of AI Programming: Case Studies in Common Lisp》와 《Verbmobil: A Translation System for Face?to?Face Dialog》, 그리고 《Intelligent He
구글의 연구실장이며, 2002년에서 2005년까지 핵심 웹 검색 엔진 개발을 이끌었다. 전에는 NASA Ames Research Center의 계산 과학 분과장으로서 NASA의 인공지능 및 로봇공학 연구와 개발을 감독했다. 서던 캘리포니아 대학교의 교수였으며, 버클리 대학교와 스탠퍼드 대학교의 연구교수단 일원이었다. 그의 다른 책으로는 《Paradigms of AI Programming: Case Studies in Common Lisp》와 《Verbmobil: A Translation System for Face?to?Face Dialog》, 그리고 《Intelligent Help Systems for UNIX》가 있다.
피터 노빅 의 다른 상품
역류광
관심작가 알림신청
커누스 교수의 『컴퓨터 프로그래밍의 예술』 시리즈를 비롯해 90여 권의 다양한 IT 전문서를 번역한 전문 번역가이다. 이 책과 연관된 번역서로는 『딥러닝을 위한 수학』 『파이썬으로 배우는 자연어 처리 인 액션』 『마스터링 트랜스포머』 등이 있으며, Manning 출판사의 『LLMs in Production』을 번역 중이다. 홈페이지 류광의 번역 이야기(https://occamsrazr.net)와 IT 및 게임 개발 정보 공유 사이트 GpgStudy (https://gpgstudy.com)를 운영한다.
류광의 다른 상품
품목정보
- 발행일
- 2021년 08월 25일
- 쪽수, 무게, 크기
- 628쪽 | 188*245*27mm
- ISBN13
- 9791191600322
책 속으로
|
기계가 지능적으로 행동할 수 있다는 가능성 자체에 반대하면서 약 인공지능을 비판한 사람들이 있었는데, 지금 시점에서 보면 그런 사람들은 선견지명이 부족했다고 할 수 있다. 예를 들어 사이먼 뉴컴(Simon Newcomb)은 1903년 10월에 “공중 비행은 인간이 절대로 감당할 수 없는 어려운 문제 중 하나이다”라고 썼지만, 불과 두 달 후에 라이트 형제가 키티호크의 들판에서 유인 동력 비행에 성공했다. 그러나, 최근 인공지능이 급격히 발전했지만 그렇다고 인공지능의 능력에 한계가 없음이 증명된 것은 아니다.
--- p.412 이러한 신뢰성 문제가 두드러지게 노출된 사례가 있다. 1986년 9월 26일 소비에트 미사일 장교 스타니슬라프 페트로프의 컴퓨터 디스플레이에 미사일 공격 경보가 떴다. 프로토콜에 따라 페트로프는 핵 무기 반격 절차를 시작해야 했지만, 그는 그 경보가 시스템의 버그 때문이라고 의심하고는 조사해 보았다. 그가 옳았으며, 덕분에 인류는 제3차 세계대전을 (가까스로) 피할 수 있었다. 그 과정에서 인간의 개입이 없었다면 어떤 일이 일어났을지 우리는 알지 못한다. --- p.422 효용 함수가 외부효과(externality)들 때문에 잘못될 수도 있다. 외부효과는 경제학에서 측정 및 지불 대상 바깥의 요인을 가리키는 용어이다. 온실 가스를 외부효과로 간주하면 기업과 국가는 온실 가스를 배출해도 벌칙을 받지 않으며, 결과적으로 지구의 모든 사람이 고통을 받는다. 생태학자 개릿 하딘은 공유 자원의 남용을 가리켜 공유지의 비극(tragedy of the commons)이라고 불렀다. 외부효과를 내부화하면, 즉 외부효과를 효용 함수의 일부로 두면 비극을 완화할 수 있다. 탄소세의 도입이 그런 내부화의 예이다. --- p.439 굿의 ‘지능 폭발’을 수학 교수이자 과학소설 작가 버너 빈지(Vernor Vinge)는 기술적 특이점(technological singularity)이라고 불렀다. 1993년에 그는 “30년 내로 인류는 초인적 지능을 만들어 낼 기술적 수단들을 갖출 것이다. 얼마 후에는 인류의 시대가 끝날 것이다”라고 썼다(Vinge, 1993). 2017년에 발명가이자 미래학자인 레이 커즈와일(Ray Kurzweil)은 2045년이면 특이점이 나타날 것이라고 예측했다. 24년 만에 미래 시점이 30년에서 28년으로 2년 앞당겨진 것이다. (이 속도로 간다면 336년밖에 남지 않았다!) 기술적 진보의 여러 측정치가 현재 지수적으로 증가한다는 빈지와 커즈와일의 지적은 사실이다. --- p.441 |
출판사 리뷰
|
2010년대 인공신경망의 부활과 심층학습의 눈부신 성과를 반영한 인공지능 연구의 결정판!
2016년에 나온 제3판 번역서(2009년에 출간된 원서 3판을 번역)는 최신 연구 반영에 한계가 있었으나, 이번 제4판 번역서는 최근 성과(자연어 이해, 로봇공학, 컴퓨터 시각에 심층학습이 끼친 영향, 강화학습을 로봇공학에 적용하는 방법, 기계학습, 인공지능 윤리 등)를 충실하게 반영한, 2020년에 출간된 원서를 옮긴 것이라 이 분야의 ‘좀 더 통합된 상’을 원하는 여러 독자의 갈증을 해소하는 데 큰 도움이 될 것입니다. 전체적으로, 책의 약 25%가 완전히 새로운 내용이고 나머지 75%도 이 분야의 좀 더 통합된 상을 제시하기 위해 크게 변경되었으며, 이번 판에서 인용한 문헌의 22%는 2010년 이후에 출판된 것들입니다. 제4판에서 새로운 점들 ■ 사람이 손으로 짜는 지식 공학보다는 기계학습에 좀 더 무게를 실었다. 기계학습은 가용 데이터와 컴퓨팅 자원이 증가하고 새로운 알고리즘들이 등장한 덕분에 큰 성공을 거두고 있다. ■ 심층학습, 확률적 프로그래밍, 다중 에이전트 시스템을 각각 개별적인 장(챕터)으로 두어서 좀 더 자세히 다룬다. ■ 자연어 이해, 로봇공학, 컴퓨터 시각에 관한 장들을 심층학습이 끼친 영향을 반영해서 수정했다. ■ 로봇공학 장에 사람과 상호작용하는 로봇에 관한 내용과 강화학습을 로봇공학에 적용하는 방법에 관한 내용이 추가되었다. ■ 이전에는 인공지능의 목표를 사람이 구체적인 효용 정보(목적함수)를 제공한다는 가정하에서 기대 효용을 최대화하려는 시스템을 만드는 것이라고 정의했다. 그러나 이번 판에서는 목적함수가 고정되어 있으며 인공지능 시스템이 목적함수를 알고 있다고 가정하지 않는다. 대신, 시스템은 자신이 봉사하는 인간의 진짜 목적이 무엇인지 확실하게 알지 못할 수 있다고 가정한다. 시스템은 반드시 자신이 무엇을 최대화할 것인지를 배워야 하며, 목적에 관해 불확실성이 존재하더라도 적절히 작동해야 한다. ■ 인공지능이 사회에 미치는 영향을 좀 더 자세하게 다루었다. 여기에는 윤리, 공정성, 신뢰, 안정성에 관한 핵심적인 문제들을 고찰한다. ■ 각 장 끝의 연습문제들을 온라인 사이트로 옮겼다. 덕분에 강사들의 요구와 이 분야 및 인공지능 관련 소프트웨어 도구의 발전에 맞게 연습문제들을 계속 추가, 갱신, 개선할 수 있게 되었다. |
