본문 바로가기

전체 글

[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
[내가 읽은 책]인공지능, 머신러닝, 딥러닝 입문_71 Machine Learning Classification involves creating models to predict or classify data into groups, typically through supervised learning. KNN : Uses neighbor information to classify data; simple and intuitive algorithm.SVM: Commonly used for binary classification by representing data in vector space.Decision Tree: Based on inductive reasoning, splits data iteratively, optimizing tree structure us.. 더보기
[내가 읽은 책]인공지능, 머신러닝, 딥러닝 입문_70 Data Preprocessing :Data must be converted into vectors or matrices for computers to process like humans. Vectors simplify mathematical modeling by quantifying features of the data. Learning and Decision Rules:Learning involves creating decision rules from input data to handle new tasks. Training datasets are used to establish these rules and apply them to new tasks. 머신러닝에 필요한 사전학습머신러닝은 이진수로 표현된.. 더보기
[내가 읽은 책]인공지능, 머신러닝, 딥러닝 입문_69 Machine learning, first coined by IBM researcher Arthur Samuel in his study on checkers, refers to algorithms improving a program's performance based on experience. It developed through three paradigms: neural models (evolving into deep learning), symbolic learning (using logic and graphs), and modern knowledge-based paradigms (recycling prior knowledge). Since the 1990s, practical applications .. 더보기

티스토리툴바