良い講義ありがとうございました！

jjhgwx

生成AIまたはLLMをベースにRAG（検索拡張生成：Retrieval-Augmented Generation）システムを構築しても、望む性能が得られず、適切な解決策が見つからない場合はどうすればよいでしょうか？
本講義では、認知負荷（Cognitive Load）理論に基づき、RAG性能を改善する方策を提示します。この講義を通じて、LLMコンテキストウィンドウの限界を理解し、RAGシステムにおける認知負荷の効果的な管理方法を習得できます。チャンク（Chunk）のサイズと構造設計、高品質チャンク生成手法、動的最適化、性能評価、実践技法などを網羅する実務レベルの理論講義です。

arigaram

RAG性能の限界を突破する認知負荷管理技術

有益な講義となったようで嬉しいです。

컨텍스트창관리

chunk전략설계

rag시스템구축

llm성능평가및튜닝

동적최적화

What should be done when building a generative AI or LLM-based RAG (Retrieval-Augmented Generation) system, but the desired performance isn't achieved and there's no suitable solution?

This lecture presents methods to improve RAG performance based on Cognitive Load theory. Through this lecture, you will understand the limitations of LLM context windows and learn how to effectively manage cognitive load in RAG systems. It is a practical-level theoretical lecture covering Chunk size and structure design, high-quality Chunk generation techniques, dynamic optimization, performance evaluation, and practical techniques.

입문

초급

중급이상

챗지피티

ChatGPT

대규모 언어 모델

검색증강생성

Generative AI

Cognitive Load Management Technology Breaking Through the Limits of RAG Performance

[Basic] Course Introduction and Basic Concepts

[Basic] Context Window and Token Limitations

[Basic] Chunking Strategy: Chunk Size and Structure

[Basic] High-Quality Chunk Generation Techniques

[Basic] RAG Pipeline Integration Step 1: Preprocessing

[Basic] RAG Pipeline Integration Stage 2: Retrieval (Similarity Search and Filtering)

[Basic] RAG Pipeline Integration Step 3: Chunk Reconstruction and Prompt Redesign

[Basic] RAG Pipeline Integration Step 4: LLM Answer Generation and Post-processing

[Basic] RAG Pipeline Integration Step 5: Hallucination Detection and Re-injection Strategy

[Basic] RAG Pipeline Integration Practice: Building a RAG System that Responds to Simple Questions

[Advanced] Dynamic Optimization: Question Complexity Assessment

[Advanced] Dynamic Optimization: Dynamic Chunk Size Adjustment Algorithm

[Advanced] Dynamic Optimization: Adaptive Summary Parameter Adjustment

[Advanced] Dynamic Optimization: Context Accumulation Management in Multi-turn Conversations

[Advanced] Dynamic Optimization: System Monitoring and Feedback Loop Design

[Advanced] Performance Evaluation: Measuring Recall and Accuracy

[Advanced] Performance Evaluation: Latency Analysis and Cost Analysis

[Advanced] Performance Evaluation: Token Usage Efficiency Metrics

[Advanced] Performance Evaluation: User Satisfaction Assessment Methodology

[Advanced] Performance Evaluation: Strategy Validation Through A/B Testing

[Advanced] Technology Outlook: Fully Automated Chunk Optimization Research Project

[Advanced] Technology Outlook: Long-term Memory Integration

[Advanced] Technology Outlook: Large-Scale Multi-Document Based RAG Scenarios

[Advanced] Technology Outlook: Multimodal Information Processing

[Advanced] RAG System Construction Practical Implementation Methods: 1. Project Topic Selection and Proposal Writing

[Advanced] RAG System Implementation Practical Methods: 2. Building Pipelines for Data Collection and Preprocessing

[Advanced] RAG System Implementation Practical Methods: 3. Chunking Strategy Design and Implementation

[Advanced] RAG System Implementation Practical Methods: 4. RAG System Integration and Coordination

[Advanced] RAG System Implementation Practical Methods: 5. Performance Evaluation and Assessment Report Writing

[Advanced] RAG System Implementation Practical Methods: 6. Final Presentation and Code Review

(Long-term plan: Schedule TBD) [Supplement] Supplement to "Chunking Strategy: Size and Structure of Chunks"