MiniRAG Introduces Near-LLM Accurate RAG for Small Language Models with Just 25% of the Storage
šš£MiniRAG Introduces Near-LLM Accurate RAG for Small Language Models with Just 25% of the Storage.
Achieving that by Semantic-Aware Heterogeneous Graphā¦
MiniRAG Introduces Near-LLM Accurate RAG for Small Language Models with Just 25% of the Storage
Agentic Deep Graph Reasoning Yields Self-Organizing Knowledge Networks
I love Markus J. Buehler's work, and his latest paper "Agentic Deep Graph Reasoning Yields Self-Organizing Knowledge Networks" does not disappoint, revealingā¦ | 19 comments on LinkedIn
Agentic Deep Graph Reasoning Yields Self-Organizing Knowledge Networks
MiniRAG Introduces Near-LLM Accurate RAG for Small Language Models with Just 25% of the Storage
šš£MiniRAG Introduces Near-LLM Accurate RAG for Small Language Models with Just 25% of the Storage.
Achieving that by Semantic-Aware Heterogeneous Graphā¦
MiniRAG Introduces Near-LLM Accurate RAG for Small Language Models with Just 25% of the Storage
KnowPath: Knowledge-enhanced Reasoning via LLM-generated Inference Paths over Knowledge Graphs
Breaking LLM Hallucinations in a Smarter Way!
(Itās not about feeding more data)
Large Language Models (LLMs) still struggle with factual inaccuracies, butā¦
KET-RAG: Turbocharging AI Agents with 10x Cheaper, Smarter Knowledge Retrieval
KET-RAG: Turbocharging AI Agents with 10x Cheaper, Smarter Knowledge Retrieval
This Multi-Granular Graph Framework uses PageRank and Keyword-Chunk Graph to have the Best Cost-Quality Tradeoff
ļ¹ļ¹ļ¹ļ¹ļ¹ļ¹ļ¹ļ¹ļ¹
ćThe Problem: Knowledge Graphs Are Expensive (and Clunky)
AI agents needĀ contextĀ to answer complex questionsālike connecting āCOVID vaccinesā to āmyocarditis risksā across research papers. But todayās solutions face two nightmares:
āøĀ Cost: Building detailed knowledge graphs with LLMs can costĀ $33,000 for a 5GB legal case.
āøĀ Quality: Cheap methods (like KNN graphs) miss key relationships, leading toĀ 32% worse answers.
āĀ Imagine training an AI doctor that either bankrupts you or misdiagnoses patients. Ouch.
ļ¹ļ¹ļ¹ļ¹ļ¹ļ¹ļ¹ļ¹ļ¹
ćThe Fix: KET-RAGās Two-Layer Brain
KET-RAG mergesĀ precisionĀ (knowledge graphs) andĀ efficiencyĀ (keyword-text maps) into one system:
āøĀ Layer 1: Knowledge Graph Skeleton
ā Uses PageRank to findĀ core text chunksĀ (like āvaccine side effectsā in medical docs).
ā Builds a sparse graphĀ onlyĀ on these chunks with LLMsāsaving 80% of indexing costs.
āøĀ Layer 2: Keyword-Chunk Bipartite Graph
ā Links keywords (e.g., āmyocarditisā) to all related text snippetsāno LLM needed.
ā Acts as a āfast laneā for retrieving context without expensive entity extraction.
ļ¹ļ¹ļ¹ļ¹ļ¹ļ¹ļ¹ļ¹ļ¹
ćResults: Beating Microsoftās Graph-RAG with Pennies
On HotpotQA and MuSiQue benchmarks, KET-RAG:
āøĀ Retrieves 81.6%Ā of critical info vs. Microsoftās 74.6%āwith 10x lower cost.
āø Boosts answer accuracy (F1 score) byĀ 32.4%Ā while cutting indexing bills byĀ 20%.
āø Scales to terabytes of data without melting budgets.
āĀ Think of it as a Tesla Model 3 outperforming a Lamborghini at 1/10th the price.
ļ¹ļ¹ļ¹ļ¹ļ¹ļ¹ļ¹ļ¹ļ¹
ćWhy AI Agents Need This
AI agents arenāt just chatbotsātheyāreĀ problem solversĀ for medicine, law, and customer service. KET-RAG gives them:
āøĀ Real-time, multi-hop reasoning: Connecting ādrug A ā gene B ā side effect Cā in milliseconds.
āøĀ Cost-effective scalability: Deploying agents across millions of documents without going broke.
āøĀ Adaptability: Mixing precise knowledge graphs (for critical data) with keyword maps (for speed).
Paper in comments
ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£ā£
ćBuild Your Own Supercharged AI Agent?
š® Join My ššš§šš¬-šš§ šš šš šš§šš¬ šš«šš¢š§š¢š§š TODAY!
and Learn Building AI Agent with Langgraph/Langchain, CrewAI and OpenAI Swarm + RAG Pipelines
šš§š«šØš„š„ ššš [34% discount]:
š https://lnkd.in/eGuWr4CH | 10 comments on LinkedIn
KET-RAG: Turbocharging AI Agents with 10x Cheaper, Smarter Knowledge Retrieval
Adaptive Graph of Thoughts (AGoT), a test-time framework that replaces rigid prompting strategies (like Chain/Tree of Thought) with dynamic directed acyclic graphs
Dynamic Reasoning Graphs + LLMs = š¤
Large Language Models (LLMs) often stumble on complex tasks when confined to linear reasoning.
What if they could dynamically restructure their thought process like humans?
A new paper introduces Adaptive Graph of Thoughts (AGoT), a test-time framework that replaces rigid prompting strategies (like Chain/Tree of Thought) with dynamic directed acyclic graphs (DAGs).
Instead of forcing fixed reasoning steps, AGoT recursively decomposes problems into sub-tasks, selectively expanding only the most critical pathways.
This is crucial for industries like scientific research or legal analysis, where problems demand non-linear, nested reasoning.
The key innovation lies in complexity checks: AGoT assesses each reasoning node, spawning sub-graphs for intricate subtasks while resolving simpler ones directly.
This mirrors how experts allocate mental effortādrilling into uncertainties while streamlining obvious steps.
The framework achieved 46.2% improvement on GPQA (a notoriously hard science QA benchmark), rivaling gains from compute-heavy fine-tuning.
By unifying chain, tree, and graph paradigms, AGoT retains CoTās clarity, ToTās exploration, and GoTās flexibility without manual tuning.
The result? LLMs that self-adapt their reasoning depth based on problem complexityāno architectural changes needed.
For AI practitioners, AGoTās DAG structure offers a principled interface to scale reasoning modularly.
ā
ššš§š§š š¤š§šØš° š°š”šš š²šØš® š¦š¢š¬š¬šš? Join my newsletter with 50k+ readers that breaks down all you need to know about the latest LLM research: llmwatch.com š”
Adaptive Graph of Thoughts (AGoT), a test-time framework that replaces rigid prompting strategies (like Chain/Tree of Thought) with dynamic directed acyclic graphs
GFM-RAG: The First Graph Foundation Model for Retrieval-Augmented Generation
š Introducing GFM-RAG: The First Graph Foundation Model for Retrieval-Augmented Generation!
Weāre excited to share our latest research: GFM-RAG: Graphā¦ | 20 comments on LinkedIn
GFM-RAG: The First Graph Foundation Model for Retrieval-Augmented Generation
Dynamic Reasoning Graphs + LLMs = š¤
Large Language Models (LLMs) often stumble on complex tasks when confined to linear reasoning.
What if they couldā¦ | 10 comments on LinkedIn
š Pathway to Artificial General Intelligence (AGI) š This is my view on the evolutionary steps toward AGI: 1ļøā£ Large Language Models (LLMs): Language modelsā¦
KAG: Boosting LLMs in Professional Domains via Knowledge Augmented...
The recently developed retrieval-augmented generation (RAG) technology has enabled the efficient construction of domain-specific applications. However, it also has limitations, including the gap...
A comparison between ChatGPT and DeepSeek capabilities writing a valid Cypher query
Today, I conducted a comparison between ChatGPT and DeepSeek chat capabilities by providing them with a schema and a natural language question. I tasked themā¦
a comparison between ChatGPT and DeepSeek chat capabilities by providing them with a schema and a natural language question. I tasked them with writing a valid Cypher query to answer the question.
Terminology Augmented Generation (TAG)? Recently some fellow terminologists have proposed the new term "Terminology-Augmented Generation (TAG)" to refer toā¦ | 29 comments on LinkedIn
What is really Graph RAG? Inspired by "From Local to Global: A Graph RAG Approach to Query-Focused Summarization" paper from Microsoft! How do you combineā¦ | 12 comments on LinkedIn
Knowledge Graphs as a source of trust for LLM-powered enterprise question answering
Knowledge Graphs as a source of trust for LLM-powered enterprise question answering That has been our position from the beginning when we started our researchā¦ | 29 comments on LinkedIn
Knowledge Graphs as a source of trust for LLM-powered enterprise question answering
A zero-hallucination AI chatbot that answered over 10000 questions of students at the University of Chicago using GraphRAG
UChicago Genie is now open source! How we built a zero-hallucination AI chatbot that answered over 10000 questions of students at the University ofā¦ | 25 comments on LinkedIn
a zero-hallucination AI chatbot that answered over 10000 questions of students at the University of Chicago
Enhancing RAG-based apps by constructing and leveraging knowledge graphs with open-source LLMs
Graph Retrieval Augmented Generation (Graph RAG) is emerging as a powerful addition to traditional vector search retrieval methods. Graphs are great at repre...
