09_AI [Mark Down]

09_AI

0901_AI 개요🚩

AI의 정의와 개념🚩

• AI(Artificial Intelligence)
• 귀납적사고(Inductive Reasoning)🚩128관2_(6)
• 술어논리(Predicate Logic)🚩122관1_(12)
• 특이점(Singularity)
• 엑소브레인(Exobrain)

AI 데이터와 품질 관리🚩

• AI학습용데이터특성🚩126관3_(6)
• 데이터획득정제방법과기준🚩126관3_(6)
• 데이터라벨링(Data Labeling)🚩126관3_(6)
• 어노테이션(Annotation)🚩134관1_(6),126관3_(6)
• 벤치마크 데이터셋
• AI학습용데이터품질관리가이드라인🚩131관3_(1),128관2_(1)

AI의 윤리와 법적 고려사항🚩

• AI윤리🚩134관4_(3),131관2_(3),130관2_(5),124관4_(6)
• AI신뢰성🚩133관1_(8)
• AI거버넌스모형🚩131관2_(3)
• 정부의인공지능윤리기준🚩129관1_(2)
• 딥페이크(Deepfake)🚩133관1_(10)
• 디지털 카르텔

0902_머신러닝🚩

머신러닝의 개념과 유형🚩

• 머신러닝(Machin Learning)🚩131관1_(7),128관2_(6)
• 기계학습모델링(Machine Learning Modeling)🚩128관1_(4)
• 지도학습(Supervised Learning)
• 비지도학습(Unsupervised Learning)
• 강화학습(Reinforcement Learning)
• 자기지도학습(Self Supervised Learning)
• 반지도학습(Semi Supervised Learning)

머신러닝의 알고리즘🚩

• 퍼셉트론
• SVM(Support Vector Machine)🚩132관2_(2),130관2_(1),127관1_(10)
• KNN(K-Nearest Neighbors)
• Q-learning
• 군집(Clustering)
• K-평균알고리즘(K-Means)🚩130관2_(1),129관1_(5)
• DBSCAN(Density Based Spatial Clustering of Applications with Noise)🚩130관2_(1),129관1_(5)
• SOM(Self Organizing Map)🚩134관4_(6)
• 데이터차원축소(Data Dimensionality)🚩131관1_(5)

머신러닝 발생 문제

• 차원의저주(Curse of Dimensionality)
• 선형성문제(Linearity Problem)
• 편향분산트레이드오프(Bias Variance Tradeoff)

머신러닝의 최적화🚩

• 머신러닝최적화알고리즘(Optimization Algorithm)🚩130관1_(9)
• 결정 트리 깊이 (Depth of Decision Tree)
• K-값 (K for K-Nearest Neighbors)
• 그리드탐색(Grid Search)
• 랜덤탐색(Random Search)
• 교차 검증 (Cross Validation)
• Quadratic Programming (QP)
• Coordinate Descent
• L-BFGS (Limited-memory BFGS)

0903_딥러닝🚩

딥러닝의 개념🚩

• 딥러닝(Deep Learning)🚩131관1_(7)

딥러닝 구조🚩

• 신경망분석(Neural Network Analysis)🚩134관4_(6)
• RNN(Recurrent Neural Network)
• LSTM(Long Short Term Memory)
• GRU(Gated Recurrent Unit)
• CNN(Convolutional Neural Network)
• SNN(Spiking Neural Network)🚩126관1_(4)

딥러닝 모델 유형

• GAN(Generative Adversarial Networks)
• DCGAN(Deep Convolutional GAN)
• Auto Encoder
• VAE(Variational Autoencoder)
• GNN(Graph Neural Network)
• GCN(Graph Convolutional Network)
• R-CNN(Region-based Convolutional Neural Network)
• Fast R-CNN(Fast Region-based Convolutional Neural Networks)
• Faster R-CNN
• YOLO(You Only Look Once)
• Vision Transformer(ViT)

딥러닝 발생 문제

• 기울기소실문제(Gradient Vanishing)
• 기울기폭발문제(Gradient Exploding)
• 딥러닝모델의수렴불안정성(Unstable Convergence)
• 비선형성문제(Nonlinearity Problem)
• 소프트맥스함수의출력불안정성(Softmax Output Instability)

딥러닝 최적화🚩

• 배치크기(Batch Size)
• 에포크 수(Number of Epochs)
• 드롭아웃 비율(Dropout Rate)
• 드롭아웃(Dropout)
• ReLU(Rectified Linear Unit)
• 배치 정규화(Batch Normalization)
• 가중치 정규화(Weight Regularization)
• EfficientNet🚩123관1_(3)
• EfficientDet🚩123관1_(3)
• 텐서플로(TensorFlow)🚩122관1_(6)

0904_머신러닝 딥러닝 공통🚩

공통 학습 기법🚩

• 인공신경망(Neural Networks)🚩133관3_(6)
• 피드포워드뉴럴네트워크(Feedforward Neural Network)🚩133관3_(6)
• 오류역전파(Backpropagation)🚩133관3_(6)
• 활성화함수(Activation Function)🚩133관3_(6)

공통 발생 문제🚩

• 과적합(Overfitting)🚩127관1_(11)
• 학습속도저하(Slow Convergence)
• 비대칭가중치초기화문제(Weight Initialization Problem)
• 메모리문제(Memory Bottleneck)

공통 최적화 항목🚩

• 파라미터(Parameter)🚩120관4_(3)
• 하이퍼파라미터(Hyperparameter)🚩120관4_(3)
• 학습률(Learning Rate)
• 정규화 계수(Regularization Coefficient)

공통 최적화 기법

• 경사하강법
• 학습률 조정(Learning Rate Scheduling)

공통 최적화 알고리즘 유형🚩

• 베이지안최적화(Bayesian Optimization)🚩132관1_(3),130관1_(9)
• 모멘텀(Momentum)🚩130관1_(9)
• Adam(Adaptive Moment Estimation)🚩130관1_(9)
• RMSProp(Root Mean Square Propagation)🚩130관1_(9)
• EM 알고리즘(Expectation Maximization Algorithm)
• 메타휴리스틱스(Metaheuristics)🚩126관1_(2)

0905_자연어 처리(NLP)🚩

자연어 처리 프레임워크🚩

• 랭체인(LangChain)🚩132관4_(3)

자연어 처리 모델🚩

• NLP(Natural Language Processing)🚩
• PLM(Pre-trained Language Model)🚩133관2_(4)
• LLM(Large Language Model)🚩132관4_(3)
• BERT(Bidirectional Encoder Representations from Transformers)🚩123관4_(4)
• GPT(Generative Pre-trained Transformer)
• GPT-3🚩123관4_(4)
• Word Embedding🚩123관4_(4)
• 자연어처리임베딩(Embedding)🚩124관4_(1)
• Instruct GPT
• ChatGPT

자연어 처리 응용 기술🚩

• NER(Named Entity Recognition)🚩123관4_(4)
• TF-IDF(Term Frequency-Inverse Document Frequency)🚩132관3_(3)
• RAG(Retrieval Augmented Generation)🚩134관1_(13)
• 트랜스포머(Transformer)
• Attention Machanism

0906_AI 모델 평가와 최적화🚩

모델 성능 평가 지표🚩

• AI모델성능평가지표🚩134관1_(3)
• 재현율(Recall)🚩134관1_(3)
• 정밀도(Precision)🚩134관1_(3)
• F1-Score🚩134관1_(3)
• 혼동행렬(Confusion Matrix)🚩134관1_(3)

모델 최적화와 경량화

• 오버피팅(Overfitting)
• 언더피팅(Underfitting)
• 편향(Bias)
• 클래스 불균형
• 딥러닝모델 경량화기술

0907_AI 실무 적용과 발전🚩

AI의 실무 적용과 운영🚩

• MLOps🚩130관2_(2)
• DSML(Data Science & Machine Learning)프로젝트🚩130관2_(2)
• 머신러닝파이프라인(Machine Learning Pipeline) 🚩121관1_(13)

AI 도구와 자동화

• AutoML
• Apache Mahout
• H2O.ai
• Microsoft Azure Machine Learning
• AIaaS

AI 모델 운영과 특수 기술🚩

• AI파운데이션모델🚩131관4_(4)
• XAI
• 온디바이스AI(Artificial Intelligence)🚩127관2_(4)
• 멀티모달 AI
• 연합학습(Federated Learning)🚩128관3_(6)
• 딥뷰(DeepView)🚩129관2_(3)
• Adaptive AI

생성형 AI 활용🚩

• 생성형AI🚩133관3_(2)
• Machine Reading Comprehesion
• Speech to Text
• DALL-E
• Codex