Beyond Prompt Engineering: The System Engineering of AI Agents — Deconstructing Memory, Tool-Making, and Autonomy in OpenClaw

Hong-Wei Wu published on 2026-03-16 included in category Essential AI Concept

Deep dive into AI Agent architecture: Learn how to overcome LLM limitations using system engineering. Explore OpenClaw’s design philosophy, context management, autonomous tool-making, and essential security sandboxing for production-grade automated systems.

Why Your RAG Fails at Complex Tables—And How MixRAG Fixes It: A Deep Dive into Heterogeneous Document Retrieval

Hong-Wei Wu published on 2026-03-15 included in category Paper Introduction

Boost your RAG system's performance on complex hierarchical tables and text. Discover the MixRAG framework: featuring H-RCL for precise data retrieval and the RECAP strategy to eliminate LLM calculation hallucinations in heterogeneous documents.

Deep Dive into MRAG: Solving Temporal Reasoning in RAG with Symbolic Logic (EMNLP 2025)

Hong-Wei Wu published on 2026-03-10 included in category Paper Introduction

Discover how the MRAG framework (EMNLP 2025) solves temporal reasoning issues in RAG systems. Learn to combine neural networks with symbolic logic to eliminate time-sensitive AI hallucinations and improve retrieval accuracy.

Selective Clarification: Overcoming LLM Hallucinations via the CLAM Framework

Hong-Wei Wu published on 2026-02-25 included in category Paper Introduction

Explore CLAM (CLarify-if-AMbiguous), a framework reducing LLM hallucinations by enabling meta-cognition. Learn how LLMs detect ambiguity and ask clarifying questions instead of guessing.

[EMNLP 2025] TableRAG Deep Dive: Text is for Reading, Tables are for Querying (via SQL)

Hong-Wei Wu published on 2026-02-13 included in category Paper Introduction

Unlock the power of TableRAG (EMNLP 2025). Learn how this hybrid RAG framework combines SQL execution with vector search to master complex reasoning in text-table heterogeneous documents.

Stop LLMs From Blindly Guessing: A Deep Dive into "Clarify When Necessary" (NAACL 2025)

Hong-Wei Wu published on 2026-02-08 included in category Paper Introduction

Stop LLMs from blindly guessing! Explore the "Clarify When Necessary" (NAACL 2025) paper and the INTENT-SIM algorithm. Learn how to teach AI to ask clarifying questions instead of hallucinating, reducing ambiguity and boosting accuracy in conversational agents.

Beyond HyDE: How ReDE-RF Makes RAG 10x Faster by "Judging" Instead of "Writing"

Hong-Wei Wu published on 2026-02-03 included in category Paper Introduction

Discover ReDE-RF, a breakthrough RAG approach from MIT that outperforms HyDE by shifting LLMs from "writers" to "judges". Learn how using Output Logits and real document feedback can eliminate hallucinations and boost retrieval speed by up to 10x in zero-shot domains. Perfect for engineers looking to optimize Semantic Search.