Loading video...
Video Failed to Load
1/2 😼 Your Own Private #DESCI Assistant in action - built using . Talk to data like you talk to your academic supervisor! 😍AI Engineer Nikita explains how we make interaction with our governance platform easy and fun for scientists and non-scientists alike. Features ☑️Article Overview ☑️ Research Q&A... show more
187,407 views • 2 years ago •via X (Twitter)
10 Comments

2/2 🫡 @Langchain deep-dive We used recursive splitting to divide the text into manageable blocks. Texts were embedded using "Huggingface embeddings," which convert the text into numerical vectors. AI assistant processes requests by calculating L2 distance, selecting the top7 text blocks with the lowest L2 distance for further analysis. A semantic check ensures that selected text blocks are closely related, allowing the AI assistant to generate an appropriate and relevant answer based on the processed information. #AI #DEV

@LangChainAI KGs will make things easier for us 🙏

@LangChainAI Especially graphs derived via data crowdsourcing

@LangChainAI Great project best wish @NavidM1 @Lukadn8686 @alivedrop2 @Aniarz89 @13751375ff

@LangChainAI Anticipating the big launch🤭

@LangChainAI @jokoz_xyz @Asquared_2112 @BigGbotex #PrivateAI #PGPT #PrivateAITGE #June10th

@LangChainAI Great team definitely 👀👀 @Tripxxzz @Ara2spicy @custommade_ng

@LangChainAI Keep cooking🫡

@LangChainAI This is going to bring big positive change

@LangChainAI i am sooo ready


