--- title: "AI 가이드 인덱스" id: "797" type: "page" slug: "ai-guide-index" published_at: "2025-12-08T02:52:13+00:00" modified_at: "2026-07-23T03:39:57+00:00" url: "https://doyouknow.kr/ai-guide-index/" markdown_url: "https://doyouknow.kr/ai-guide-index.md" excerpt: "1. 딥러닝 핵심 구성 요소 2. AI 101 · 입문용 핵심 글 3. 학습·최적화·과적합·데이터 증강 4. 딥러닝 아키텍처: CNN·RNN·시계열·추천 4-1. CNN·컴퓨터 비전 4-2. RNN·LSTM·시계열 4-3. 추천 시스템 5. NLP·Transformer·BERT·GPT·LLM·sLLM 6. LLM 파인튜닝·양자화·RAG·Vector DB 6-1. 파인튜닝·LoRA·QLoRA 6-2. 양자화·경량화·Distillation 6-3. RAG·벡터DB·검색 인프라..." --- --- ## 1. 딥러닝 핵심 구성 요소 - **활성화 함수 완전 정복: ReLU·GELU·SwiGLU·TeLU·KAN까지** 👉 [https://doyouknow.kr/activation-function-comparison-guide/](https://doyouknow.kr/activation-function-comparison-guide/) - **손실함수(Loss Function) 완전 정복: MSE·Cross-Entropy·Focal·Dice 등** 👉 [https://doyouknow.kr/loss-function-complete-guide2/](https://doyouknow.kr/loss-function-complete-guide2/) - **성능지표(Evaluation Metrics) 완전 정복** 👉 [https://doyouknow.kr/ai-model-evaluation-metrics-complete-guide2/](https://doyouknow.kr/ai-model-evaluation-metrics-complete-guide2/) - **모델 학습과 최적화 – 손실·경사하강법·역전파·Adam** 👉 [https://doyouknow.kr/ai-model-training-optimization-loss-gradient-descent-backpropagation2/](https://doyouknow.kr/ai-model-training-optimization-loss-gradient-descent-backpropagation2/) - **퍼셉트론에서 딥러닝까지 – 신경망의 모든 것 (뉴런·가중치·활성화)** 👉 [https://doyouknow.kr/perceptron-deep-learning-neural-network-activation-function/](https://doyouknow.kr/perceptron-deep-learning-neural-network-activation-function/) --- ## 2. AI 101 · 입문용 핵심 글 - **AI 개발의 첫걸음 – Python이 AI의 표준 언어가 된 이유** 👉 [https://doyouknow.kr/python-ai-development-tensorflow-pytorch-colab/](https://doyouknow.kr/python-ai-development-tensorflow-pytorch-colab/) - **퍼셉트론에서 딥러닝까지 – 신경망의 모든 것** 👉 [https://doyouknow.kr/perceptron-deep-learning-neural-network-activation-function/](https://doyouknow.kr/perceptron-deep-learning-neural-network-activation-function/) - **[선형회귀와 분류](https://doyouknow.kr/ml-algorithms-linear-logistic-regression-decision-tree-random-forest2/) – 모든 머신러닝 알고리즘의 시작점** 👉 [https://doyouknow.kr/ml-algorithms-linear-logistic-regression-decision-tree-random-forest2/](https://doyouknow.kr/ml-algorithms-linear-logistic-regression-decision-tree-random-forest2/) - **특성 공학(Feature Engineering) – AI 성능을 2배 높이는 데이터 변환의 기술** 👉 [https://doyouknow.kr/feature-engineering-selection-extraction-pca-normalization2/](https://doyouknow.kr/feature-engineering-selection-extraction-pca-normalization2/) - **AI 성능을 좌우하는 데이터 준비의 기술 (Train/Valid/Test·전처리·라벨링)** 👉 [https://doyouknow.kr/ai-dataset-preparation-train-validation-test-labeling2/](https://doyouknow.kr/ai-dataset-preparation-train-validation-test-labeling2/) - **튜링에서 DeepSeek까지, AI 80년의 모든 것** 👉[https://doyouknow.kr/ai-history-complete-guide-turing-to-deepseek/](https://doyouknow.kr/ai-history-complete-guide-turing-to-deepseek/) --- ## 3. 학습·최적화·과적합·데이터 증강 - **손실함수 완전 정복** 👉 [https://doyouknow.kr/loss-function-complete-guide2/](https://doyouknow.kr/loss-function-complete-guide2/) - **모델 학습과 최적화 – 손실·GD·역전파·Adam** 👉 [https://doyouknow.kr/ai-model-training-optimization-loss-gradient-descent-backpropagation2/](https://doyouknow.kr/ai-model-training-optimization-loss-gradient-descent-backpropagation2/) - **하이퍼파라미터 튜닝 하나로 AI 성능 2배 – Optuna·Bayesian Optimization** 👉 [https://doyouknow.kr/hyperparameter-tuning-optuna-bayesian-optimization-guide/](https://doyouknow.kr/hyperparameter-tuning-optuna-bayesian-optimization-guide/) - **[과적합과 과소적합](https://doyouknow.kr/overfitting-underfitting-regularization-dropout-early-stopping2/) – Dropout·정규화·Early Stopping** 👉 [https://doyouknow.kr/overfitting-underfitting-regularization-dropout-early-stopping2/](https://doyouknow.kr/overfitting-underfitting-regularization-dropout-early-stopping2/) - **과적합 해결 못하면 AI 프로젝트 70% 실패 – 실전 해결 가이드** 👉 [https://doyouknow.kr/overfitting-underfitting-regularization-dropout-solution/](https://doyouknow.kr/overfitting-underfitting-regularization-dropout-solution/) - **데이터 증강 완벽 가이드 – Mixup·CutMix·AutoAugment·EDA** 👉 [https://doyouknow.kr/data-augmentation-mixup-cutmix-autoaugment-back-translation-eda-guide/](https://doyouknow.kr/data-augmentation-mixup-cutmix-autoaugment-back-translation-eda-guide/) - **AI 프로젝트 85% 실패의 진짜 원인 – 데이터 품질과 정제** 👉 [https://doyouknow.kr/data-quality-labeling-ai-project-failure-prevention/](https://doyouknow.kr/data-quality-labeling-ai-project-failure-prevention/) --- ## 4. 딥러닝 아키텍처: CNN·RNN·시계열·추천 ### 4-1. CNN·컴퓨터 비전 - **CNN – 이미지를 이해하는 AI의 비밀** 👉 [https://doyouknow.kr/cnn-convolutional-neural-network-image-recognition2/](https://doyouknow.kr/cnn-convolutional-neural-network-image-recognition2/) - **컴퓨터 비전 – 이미지 분류·객체 탐지·세그멘테이션·YOLO** 👉 [https://doyouknow.kr/computer-vision-image-classification-object-detection-yolo/](https://doyouknow.kr/computer-vision-image-classification-object-detection-yolo/) - **CNN과 RNN의 결합 – 영상 분류 등 복합 모델** 👉 [https://doyouknow.kr/cnn-lstm-hybrid-model-video-classification/](https://doyouknow.kr/cnn-lstm-hybrid-model-video-classification/) - **Vision Transformer(ViT)·CLIP·SAM – Self-Attention 기반 비전** 👉 [https://doyouknow.kr/vision-transformer-vit-cnn-self-attention-clip-sam-guide/](https://doyouknow.kr/vision-transformer-vit-cnn-self-attention-clip-sam-guide/) ### 4-2. RNN·LSTM·시계열 - **RNN과 LSTM – 시간을 기억하는 AI의 마법** 👉 [https://doyouknow.kr/rnn-lstm-gru-time-series-natural-language-processing/](https://doyouknow.kr/rnn-lstm-gru-time-series-natural-language-processing/) - **시계열 예측: LSTM vs Transformer vs TimesFM** 👉 [https://doyouknow.kr/time-series-forecasting-lstm-transformer-timesfm-stock-demand-prediction/](https://doyouknow.kr/time-series-forecasting-lstm-transformer-timesfm-stock-demand-prediction/) ### 4-3. 추천 시스템 - **추천 시스템 – 넷플릭스·유튜브가 취향을 맞추는 방법** 👉 [https://doyouknow.kr/recommendation-system-collaborative-filtering-netflix-youtube/](https://doyouknow.kr/recommendation-system-collaborative-filtering-netflix-youtube/) --- ## 5. NLP·Transformer·BERT·GPT·LLM·sLLM - **자연어 처리(NLP) – 토크나이제이션·감정 분석·챗봇** 👉 [https://doyouknow.kr/nlp-natural-language-processing-tokenization-sentiment-chatbot/](https://doyouknow.kr/nlp-natural-language-processing-tokenization-sentiment-chatbot/) - **Transformer – AI 혁명의 시작 (Self-Attention·Encoder/Decoder)** 👉 [https://doyouknow.kr/transformer-attention-mechanism-bert-gpt-chatgpt/](https://doyouknow.kr/transformer-attention-mechanism-bert-gpt-chatgpt/) - **BERT – 양방향으로 언어를 이해하는 AI** 👉 [https://doyouknow.kr/bert-bidirectional-transformer-masked-language-model/](https://doyouknow.kr/bert-bidirectional-transformer-masked-language-model/) - **GPT – 창의적인 글을 쓰는 AI** 👉 [https://doyouknow.kr/gpt-generative-autoregressive-language-model-chatgpt/](https://doyouknow.kr/gpt-generative-autoregressive-language-model-chatgpt/) - **BERT vs GPT – 두 거인의 차이점** 👉 [https://doyouknow.kr/bert-vs-gpt-comparison-encoder-decoder-guide/](https://doyouknow.kr/bert-vs-gpt-comparison-encoder-decoder-guide/) - **GPT vs Claude vs Gemini – 생성형 AI 완전 해부** 👉 [https://doyouknow.kr/gpt-claude-gemini-transformer-rlhf-diffusion-multimodal-comparison/](https://doyouknow.kr/gpt-claude-gemini-transformer-rlhf-diffusion-multimodal-comparison/) - **GPT-4 쓸까? Llama 쓸까? 오픈소스 vs 클라우드 LLM 완벽 비교** 👉 [https://doyouknow.kr/opensource-vs-cloud-llm-gpt4-llama-claude-comparison-guide/](https://doyouknow.kr/opensource-vs-cloud-llm-gpt4-llama-claude-comparison-guide/) - **sLLM 완벽 가이드 – 온디바이스 AI의 미래** 👉 [https://doyouknow.kr/sllm-guide-small-language-model-on-device-ai/](https://doyouknow.kr/sllm-guide-small-language-model-on-device-ai/) - **트랜스포머의 독주를 깰 것인가? Mamba와 상태 공간 모델(SSM)의 부상 완벽 가이드** 👉 [https://doyouknow.kr/mamba-state-space-model-ssm-transformer-alternative-guide/](https://doyouknow.kr/mamba-state-space-model-ssm-transformer-alternative-guide/) --- ## 6. LLM 파인튜닝·양자화·RAG·Vector DB ### 6-1. 파인튜닝·LoRA·QLoRA - **Fine-Tuning 완벽 가이드 – LoRA·QLoRA·비용·데이터·Colab** 👉 [https://doyouknow.kr/fine-tuning-lora-qlora-cost-data-colab-guide/](https://doyouknow.kr/fine-tuning-lora-qlora-cost-data-colab-guide/) ### 6-2. 양자화·경량화·Distillation - **LLM 양자화·경량화 – GPTQ·AWQ·GGUF·Pruning·Distillation** 👉 [https://doyouknow.kr/llm-quantization-gptq-awq-gguf-pruning-distillation-guide/](https://doyouknow.kr/llm-quantization-gptq-awq-gguf-pruning-distillation-guide/) ### 6-3. RAG·벡터DB·검색 인프라 - **Vector DB 완벽 가이드 – HNSW·IVF·Pinecone·Milvus** 👉 [https://doyouknow.kr/vector-database-guide-embedding-similarity-search-hnsw-ivf-pinecone-milvus/](https://doyouknow.kr/vector-database-guide-embedding-similarity-search-hnsw-ivf-pinecone-milvus/) - **고급 RAG – 청킹·평가·GraphRAG·Agentic RAG** 👉 [https://doyouknow.kr/rag-advanced-chunking-vector-db-ragas-graphrag-agentic-guide/](https://doyouknow.kr/rag-advanced-chunking-vector-db-ragas-graphrag-agentic-guide/) --- ## 7. 프롬프트 엔지니어링·인컨텍스트 학습·AI 에이전트 - **ChatGPT 10배 활용법 – Prompt Engineering 완벽 마스터** 👉 [https://doyouknow.kr/prompt-engineering-chatgpt-cot-few-shot-tot-guide/](https://doyouknow.kr/prompt-engineering-chatgpt-cot-few-shot-tot-guide/) - **In-Context Learning – Zero-shot·Few-shot·CoT·ToT** 👉 [https://doyouknow.kr/in-context-learning-cot-tot-zero-shot-few-shot-prompting-guide/](https://doyouknow.kr/in-context-learning-cot-tot-zero-shot-few-shot-prompting-guide/) - **AI Agent 완벽 가이드 – ReAct·AutoGPT·Function Calling·Multi-Agent** 👉 [https://doyouknow.kr/ai-agent-react-autogpt-function-calling-multi-agent-langchain-guide/](https://doyouknow.kr/ai-agent-react-autogpt-function-calling-multi-agent-langchain-guide/) - **AI 에이전트 프레임워크 비교 – LangChain vs LlamaIndex vs CrewAI vs AutoGen** 👉 [https://doyouknow.kr/ai-agent-framework-langchain-llamaindex-crewai-autogen-comparison/](https://doyouknow.kr/ai-agent-framework-langchain-llamaindex-crewai-autogen-comparison/) --- ## 8. 코딩 어시스턴트·AI 개발 생산성 - **AI 코딩 어시스턴트 – GitHub Copilot vs Cursor vs Claude** 👉 [https://doyouknow.kr/ai-coding-assistant-github-copilot-cursor-claude-comparison/](https://doyouknow.kr/ai-coding-assistant-github-copilot-cursor-claude-comparison/) --- ## 9. 데이터·합성 데이터·MLOps - **AI가 AI를 위해 데이터를 만든다 – Synthetic Data 완벽 가이드 (GAN·VAE·Diffusion)** 👉 [https://doyouknow.kr/synthetic-data-guide-gan-vae-diffusion-ai-training/](https://doyouknow.kr/synthetic-data-guide-gan-vae-diffusion-ai-training/) - **MLOps 완벽 가이드 – 파이프라인·서빙·모니터링** 👉 [https://doyouknow.kr/mlops-complete-guide-pipeline-serving-monitoring/](https://doyouknow.kr/mlops-complete-guide-pipeline-serving-monitoring/) --- ## 10. 평가·환각·보안·윤리 - **AI 벤치마크 평가 – MMLU·HumanEval·MT-Bench·LLM Judge** 👉 [https://doyouknow.kr/ai-benchmark-evaluation-mmlu-humaneval-mt-bench-llm-judge/](https://doyouknow.kr/ai-benchmark-evaluation-mmlu-humaneval-mt-bench-llm-judge/) - **AI 환각(Hallucination) – 원인·탐지·RAG·Grounding** 👉 [https://doyouknow.kr/ai-hallucination-causes-detection-solutions-rag-grounding/](https://doyouknow.kr/ai-hallucination-causes-detection-solutions-rag-grounding/) - **AI 보안 – Adversarial Attack·Defense** 👉 [https://doyouknow.kr/ai-security-guide-adversarial-attack-defense/](https://doyouknow.kr/ai-security-guide-adversarial-attack-defense/) - **Prompt Injection·Jailbreak – 공격 기법 7가지와 방어** 👉 [https://doyouknow.kr/prompt-injection-jailbreak-llm-attack-defense-guide/](https://doyouknow.kr/prompt-injection-jailbreak-llm-attack-defense-guide/) - **생성형 AI 윤리·저작권·EU AI Act·Deepfake** 👉 [https://doyouknow.kr/generative-ai-ethics-copyright-eu-ai-act-deepfake-hallucination-guide/](https://doyouknow.kr/generative-ai-ethics-copyright-eu-ai-act-deepfake-hallucination-guide/) - **AI가 차별하는 충격적 이유 – 데이터 편향과 공정성** 👉 [https://doyouknow.kr/ai-bias-fairness-ethics-aif360-fairlearn-eu-ai-act/](https://doyouknow.kr/ai-bias-fairness-ethics-aif360-fairlearn-eu-ai-act/) - **Constitutional AI & RLHF 심화 완벽 가이드** 👉[https://doyouknow.kr/constitutional-ai-rlhf-alignment-claude-anthropic-guide/](https://doyouknow.kr/constitutional-ai-rlhf-alignment-claude-anthropic-guide/) - **성능지표(Evaluation Metrics) 완전 정복** 👉 [https://doyouknow.kr/ai-model-evaluation-metrics-complete-guide2/](https://doyouknow.kr/ai-model-evaluation-metrics-complete-guide2/) --- ## 11. 강화학습·RLHF - **강화학습 심화 – RLHF·DPO·GRPO·PPO·DeepSeek 사례** 👉 [https://doyouknow.kr/reinforcement-learning-rlhf-dpo-ppo-grpo-deepseek-guide/](https://doyouknow.kr/reinforcement-learning-rlhf-dpo-ppo-grpo-deepseek-guide/) - **Constitutional AI & RLHF 심화 완벽 가이드** 👉[https://doyouknow.kr/constitutional-ai-rlhf-alignment-claude-anthropic-guide/](https://doyouknow.kr/constitutional-ai-rlhf-alignment-claude-anthropic-guide/) --- ## 12. AI 하드웨어·온디바이스·시스템 - **AI 하드웨어 전쟁 – GPU vs TPU vs NPU** 👉 [https://doyouknow.kr/nvidia-gpu-google-tpu-npu-ai-hardware-war-b200-trillium-trainium/](https://doyouknow.kr/nvidia-gpu-google-tpu-npu-ai-hardware-war-b200-trillium-trainium/) - **역설의 칩들 – 왜 어떤 칩은 배우지 못하고 추론만 할까?** 👉 [https://doyouknow.kr/why-ai-chips-cannot-learn-inference-only-paradox/](https://doyouknow.kr/why-ai-chips-cannot-learn-inference-only-paradox/) - **sLLM·온디바이스 AI** 👉 [https://doyouknow.kr/sllm-guide-small-language-model-on-device-ai/](https://doyouknow.kr/sllm-guide-small-language-model-on-device-ai/) --- ## 13. 논문 읽기 시리즈 - **YOLO v1 완벽 분석 – 실시간 객체 탐지의 시작** 👉 [http://doyouknow.kr/yolo-v1-object-detection-revolution/](https://doyouknow.kr/yolo-v1-object-detection-revolution/) - **Deep Belief Network(DBN) – AI 겨울을 끝낸 기계** 👉 [https://doyouknow.kr/deep-belief-network-dbn-history-breakthrough/](https://doyouknow.kr/deep-belief-network-dbn-history-breakthrough/) - **AlexNet – ImageNet 우승으로 시작된 딥러닝 혁명** 👉 [https://doyouknow.kr/alexnet-deep-learning-revolution-imagenet-cnn/](https://doyouknow.kr/alexnet-deep-learning-revolution-imagenet-cnn/) - **VGGNet – 3×3 컨볼루션만으로 만든 딥러닝 레전드** 👉 [https://doyouknow.kr/vggnet-complete-guide/](https://doyouknow.kr/vggnet-complete-guide/) --- ## 14. AI의 미래 - **Mechanistic Interpretability와 희소 오토인코더(Sparse Autoencoder) 완벽 가이드** 👉[https://doyouknow.kr/mechanistic-interpretability-sparse-autoencoder-sae-claude-guide/](https://doyouknow.kr/mechanistic-interpretability-sparse-autoencoder-sae-claude-guide/) - **Kolmogorov-Arnold Networks (KAN) 완벽 해부! 학습 가능한 활성화 함수의 혁명** 👉[https://doyouknow.kr/kolmogorov-arnold-networks-kan-learnable-activation-functions/](https://doyouknow.kr/kolmogorov-arnold-networks-kan-learnable-activation-functions/) - **DPO (Direct Preference Optimization) 완벽 해부! 강화학습 없이 최적화하다** 👉[https://doyouknow.kr/direct-preference-optimization-dpo-rlhf-alternative-guide/](https://doyouknow.kr/direct-preference-optimization-dpo-rlhf-alternative-guide/) --- ## 15. 실전편 - **내일의 태양광 발전량, 파이썬으로 1초 만에 예측하기 (feat. 동서발전 API)** 👉 [https://doyouknow.kr/solar-forecast-api-python-guide/](https://doyouknow.kr/solar-forecast-api-python-guide/) - **기상청 산업특화(태양광) API로 PyTorch LSTM 날씨 예측 모델 만들기: 지상관측 일통계 완전 정복** 👉 [https://doyouknow.kr/kma-solar-energy-api-pytorch-lstm/](https://doyouknow.kr/kma-solar-energy-api-pytorch-lstm/) - **기상청 평년값(일조, 구름, 기온, 습도) API 완벽 활용 가이드** 👉 [https://doyouknow.kr/kma-normal-value-api-guide/](https://doyouknow.kr/kma-normal-value-api-guide/) - **태양광 데이터, 아직도 노가다? 기상청 무료 API로 1년치 일사량 10분 만에 자동 수집!** 👉 [https://doyouknow.kr/kma-solar-radiation-sunshine-api-guide/](https://doyouknow.kr/kma-solar-radiation-sunshine-api-guide/)