본문 바로가기

전체 글

[소설보다 재미있는]사기 열전_79 하늘은 과연 착한 사람의 편인가?백이 숙제 : 수양산에서 굶어 죽은 백이와 숙제적어도 이 땅의 팔팔한 사람이라면 '백이佰夷·숙제叔齊 이름 정도는 어디선가 들어보았으리라. 그렇다. 백이·숙제는 아주 까마득한 옛날, 중국에 있던 고죽국孤竹國이란 나라의 왕자였다. 이제 그 옛날의 이야 기로 잠시 돌아가 보자. 고죽국의 임금은 이제 나이가 들어 왕위를 물려주어야 할 때가 왔다 고 생각했다. 그래서 누구에게 왕위를 물려줄까 생각하다가 셋째인 숙제에게 넘겨주기로 했다. 아마 맏이인 백이보다 셋째인 숙제가 더 마음 에 들었던 모양이다. 하지만 고죽국의 임금은 숙제가 왕위에 오르는 것 도 보지 못한 채 세상을 떠나고 말았다.“보위는 형님의 것이오. 제가 어찌 그 자리에 오를 수 있겠습니까?" 숙제는 형인 백이에게 보위.. 더보기
[내가 읽은 책]인공지능, 머신러닝, 딥러닝 입문_78 In 1979, Kunihiko Fukushima introduced the Neocognitron based on neurophysiological theories, which became the foundation of CNN. CNN mimics the human visual perception process and is used in image processing and speech recognition. It extracts and classifies features by repeatedly performing filtering, activation, and pooling. Pooling methods include max pooling, average pooling, and stochastic.. 더보기
[내가 읽은 책]인공지능, 머신러닝, 딥러닝 입문_77 Human and artificial neural networks have intriguing parallels and differences. The human brain is divided into the brainstem, cerebellum, and cerebrum. Its cerebral cortex contains approximately 100 billion neurons, which process and transmit signals through synapses.In 1943, McCulloch and Pitts introduced the concept of the Threshold Logic Unit (TLU), inspired by biological neurons and resembl.. 더보기
[내가 읽은 책]인공지능, 머신러닝, 딥러닝 입문_76 Apple lags behind competitors like Google, Microsoft, and Facebook in AI research but has recently taken steps to strengthen its AI capabilities through startup acquisitions and expert recruitment. Deep learning technology is spreading across industries, with startups innovating in areas such as computer vision, natural language processing, and healthcare. Major companies like Google and IBM lev.. 더보기
[내가 읽은 책]인공지능, 머신러닝, 딥러닝 입문_75 The deep learning race is driven by cutting-edge algorithms, with major players like Google, Facebook, Amazon, and Apple leading the charge. Collaboration between startups and corporations is essential, while Korea needs to strengthen its ecosystem. Netflix uses deep learning for personalized recommendations, accounting for 75% of its consumed content. Google, a leader in deep learning, acquired.. 더보기
[내가 읽은 책]인공지능, 머신러닝, 딥러닝 입문_74 Deep learning, a subfield of machine learning, utilizes multi-layer neural networks to automatically learn features from data and solve complex problems. Initially hindered by computational and algorithmic limitations, it advanced significantly after 2000 due to innovations like gradient descent, long short-term memory (LSTM), ReLU functions, and dropout algorithms. Deep learning excels in solvi.. 더보기
[내가 읽은 책]인공지능, 머신러닝, 딥러닝 입문_73 Reinforcement learning operates based on the Markov Decision Process (MDP) framework, which consists of states (S), actions (A), state transition probabilities (P), rewards (R), and a discount factor (γ). Adding actions to the Markov Reward Process (MRP) forms the MDP. Key algorithms for reinforcement learning include dynamic programming, Monte Carlo methods, and temporal difference methods (Q-L.. 더보기
[내가 읽은 책]인공지능, 머신러닝, 딥러닝 입문_72 Difference between Supervised and Unsupervised Learning.Supervised Learning: Predicts labels for new data using labeled datasets..Unsupervised Learning: Groups unlabeled data into clusters. Clustering is a representative model. Clustering Models.K-Means: Predetermines the number of clusters and iteratively adjusts centroids for optimal clustering..DBSCAN: Density-based clustering, identifying cl.. 더보기

티스토리툴바