R to @noduslabs: I would like to store it, like in one of TimBL’s Solid Pods, able to give access to a doctor as needs be

R to @philgribbon: What do you mean? You wouldn't like your HRV data to be stored?

R to @noduslabs: My heart could do with an AI learning its language. Needs a Solid State Protocol for Privacy Sharing

R to @MovesenseSensor: What's the difference between this and HR+ and HR2?

R to @noduslabs: Thanks so much! Working!

R to @DrevZiga: @noduslabs I saw your video of knowledge graph on YouTube. Maybe it will be interesting to you

Hello good morning @noduslabs, I have registered after giving my details to test your tool. I try to access, but it does not let me, I try to recover the password, but it does not let me. I would like to try the tool if possible. Thank you!

R to @tantaman: Damn. That thing went deep for sure. What a powerful tool. Minority report is on the way.

R to @jessmartin: Feels like we’re all pulling at the same thread huh? Thanks for the link. Definitely useful

R to @ATMOS_PR: these references might be helpful as well! https://x.com/tantaman/status/1710832184230125910?s=20

R to @heyai_pro: You can find the video here 👉 https://youtu.be/XMwa9q6wWkw?si=HJW_FAGLtqLA5L0q

It's no secret that I'm a huge fan of Knowledge Graphs as a more effective memory model for LLMs. 🤓 Just watched @noduslabs's recent video (link in comments 👇) on how they have combined KGs and LLM inference to create dynamic memory maps—a #GenerationAI human cognitive tool. 🤯 (@yoheinakajima, might interest you!) Why am I so exited about #KnowledgeGraphs as a memory model for LLMs? Contextual Understanding: Unlike flat databases, KGs store information in a network of nodes and edges that capture semantic relationships. This structure allows for a more nuanced understanding of context, making the retrieved information more relevant. Richer Representations: A KG captures not just facts but also their interrelationships. This allows the system to pull in a cluster of related facts, offering the LLM a richer context from which it can generate responses. Dynamic Reasoning: KGs inherently encode multi-dimensional reasoning. When coupled with RAG, the LLM can unpack this latent reasoning from minimal context and well as potentially availing pathways to recognize higher-order relationships. Query Flexibility: The query mechanisms for KGs are generally more flexible than text-based retrieval, supporting complex queries that can return highly specific information. This helps overcome the "lost in the middle" weakness of current LLMs, as the KG query language does the heavy lifting of relevant information selection and data-dependent branching. Anyone who has built a real-world #GenAI application knows this is the bane of our existence. Applications Sales & Customer Service Agents: In customer service or sales, agents equipped with KG-RAG could pull in real-time data, grasp a deeper context of a user query, and provide highly relevant and personalized responses. Sales agents can utilize KG-RAG to generate a semantically rich network of products, thereby simplifying recommender systems like those used in e-commerce. Decision Support Systems: In sectors like healthcare or finance—where decision-making often involves synthesizing complex and interrelated data—KG-RAG could serve multiple functions. It could 1) predict market trends by analyzing a plethora of economic indicators that may exceed LLM context limitations, and 2) develop patient treatment plans that take into account a wide array of factors such as medical history, current pathologies, and potential pharmacological interactions with existing treatments. In these ways, KG-RAG stands as a transformative technology. Complex Infrastructure Co-pilots & Agents: Modern enterprises are complex webs of interconnected data and systems. Despite the abundance of data, many organizations struggle to utilize it effectively. Agents, empowered by a KG-based understanding of these systems, have the potential to significantly enhance employee capabilities. These AI agents could undertake new tasks, even initiating multi-step processes, based on straightforward natural language directives. This is the stuff the keeps me up coding until 2AM!!!

wow ! @noduslabs https://www.youtube.com/watch?v=XMwa9q6wWkw

R to @jessmartin: @noduslabs

R to @raymmar_: inspiration for solution: @noduslabs

RT by @Britopian: How to enhance your #chatgpt conversations using #knowledgegraphs (with the help of #infranodus): https://youtu.be/Wy6NKGrVWZM