- Видео 252
- Просмотров 384 937
Weaviate • Vector Database
США
Добавлен 17 сен 2018
This channel is about all topics around vector search, and about the open-source vector database Weaviate.
New embedding model: Contextual Document Embeddings
Traditional document embeddings have a significant limitation: they encode documents independently, without considering their context or neighboring documents.
This means they have to choose a single global weighting for terms, potentially missing important contextual nuances, or overweighting terms that might occur a lot in the dataset. This can be problematic when embedding in different domains or contexts.
✨ The Solution: Contextual Document Embeddings (CDE) ✨
CDE operates in two stages:
1️⃣ Adversarial contrastive learning: batch and embed related context from neighboring documents
2️⃣ Embed the target document while considering the contextual embeddings of the related document batch
CDE ...
This means they have to choose a single global weighting for terms, potentially missing important contextual nuances, or overweighting terms that might occur a lot in the dataset. This can be problematic when embedding in different domains or contexts.
✨ The Solution: Contextual Document Embeddings (CDE) ✨
CDE operates in two stages:
1️⃣ Adversarial contrastive learning: batch and embed related context from neighboring documents
2️⃣ Embed the target document while considering the contextual embeddings of the related document batch
CDE ...
Просмотров: 705
Видео
Agentic RAG with Erika Cardenas - Weaviate Podcast #109!
Просмотров 65214 дней назад
Hey everyone! Thank you so much for watching the 109th episode of the Weaviate Podcast with Erika Cardenas! Erika, in collaboration with Leonie Monigatti, have recently published "What is Agentic RAG". This blog post that was even covered in VentureBeat with additional quotes from Weaviate Co-Founder and CEO Bob van Luijt! This podcast continues the discussion on all things Agentic RAG, coverin...
Let Me Speak Freely? with Zhi Rui Tam - Weaviate Podcast #108!
Просмотров 26121 день назад
JSON mode has been one of the biggest enablers for working with Large Language Models! JSON mode is even expanding into Multimodal Foundation models! But how exactly is JSON mode achieved? There are generally 3 paths to JSON mode: (1) constrained generation (such as Outlines), (2) begging the model for a JSON response in the prompt, and (3) A two stage process of generate-then-format. I am BEYO...
Optimize your vector database's search speed, accuracy, and costs
Просмотров 14421 день назад
Weaviate's new hot, warm, and cold storage tiers offer flexible options for managing resources to optimize search speed, accuracy, and costs 🚀 There are three main levers to adjust: • Choosing the vector index type (HNSW, flat, or dynamic) • Using compression techniques (binary, product, or scalar quantization) • Managing flexible tenant states (active, inactive, or offloaded) Learn when you sh...
SWE-bench with John Yang and Carlos E. Jimenez - Weaviate Podcast #107!
Просмотров 263Месяц назад
Hey everyone! Thank you so much for watching the 107th episode of the Weaviate Podcast! This one dives into SWE-bench, SWE-agent, and most recently SWE-bench Multimodal with John Yang from Stanford University and Carlos E. Jimenez from Princeton University! One of the most impactful applications of AI we have seen so far is in programming and software engineering! John, Carlos, and team are at ...
AI in Education with Rose E. Wang - Weaviate Podcast #106!
Просмотров 341Месяц назад
Hey everyone! I am SUPER excited to publish the 106th episode of the Weaviate Podcast featuring Rose E. Wang!! Rose is a Ph.D. student at Stanford University where she has lead incredible research at the cutting-edge of AI applications in Education. The podcast heavily discusses her recent work on Tutor CoPilot! Tutor CoPilot is one of the world's largest randomized control trials on the impact...
Compound AI Systems with Philip Kiely - Weaviate Podcast #105!
Просмотров 440Месяц назад
Hey everyone! Thanks so much for watching the 105th episode of the Weaviate Podcast with Philip Kiely! This one dives into all sorts of apsects related to Compound AI Systems! We are now seeing far better results with AI models by breaking up tasks into multiple stages and inferences. Philip explains the work they are doing at Baseten to optimize and scale deployments of these emerging systems ...
Hack Night at GitHub with Weaviate
Просмотров 256Месяц назад
Beyond hacking and writing code, there’s something incredibly fun about creating environments for likeminded and smart people to get together to learn and hack on new tech. It takes a lot of work, but the reward is great and it's pure vibes. It creates the perfect synergy for incredible things to happen, from rad demos by magically talented people like Leann Chen from Diffbot, Ben A. at Telepor...
Late chunking improves context recall in RAG pipelines
Просмотров 1,1 тыс.Месяц назад
Optimizing your chunking techniques is one of the top places to improve performance in your RAG pipelines, but what’s the best one? Jina AI just released a new method called late chunking that takes the same amount of storage space as naive chunking, but solves the problem of lost context, similarly to ColBERT. You can implement it super easily with just a few extra lines in your embedding step...
Matryoshka Representation Learning (MRL) for ML tasks and vector compression
Просмотров 4982 месяца назад
Matryoshka Representation Learning (MRL) for ML tasks and vector compression
AI Agents That Matter with Sayash Kapoor and Benedikt Stroebl - Weaviate Podcast #104!
Просмотров 6272 месяца назад
AI Agents That Matter with Sayash Kapoor and Benedikt Stroebl - Weaviate Podcast #104!
MIPRO and DSPy with Krista Opsahl-Ong! - Weaviate Podcast #103
Просмотров 2,1 тыс.3 месяца назад
MIPRO and DSPy with Krista Opsahl-Ong! - Weaviate Podcast #103
AI-Native Development with Guy Podjarny and Bob van Luijt - Weaviate Podcast #102!
Просмотров 2913 месяца назад
AI-Native Development with Guy Podjarny and Bob van Luijt - Weaviate Podcast #102!
Chat with your code: RAG with Weaviate and LlamaIndex
Просмотров 4344 месяца назад
Chat with your code: RAG with Weaviate and LlamaIndex
Scaling Pandas with Devin Petersohn - Weaviate Podcast #101!
Просмотров 3014 месяца назад
Scaling Pandas with Devin Petersohn - Weaviate Podcast #101!
Generative UIs with Lucas Negritto and Bob van Luijt - Weaviate Podcast #100!
Просмотров 7074 месяца назад
Generative UIs with Lucas Negritto and Bob van Luijt - Weaviate Podcast #100!
ACORN with Liana Patel and Abdel Rodriguez - Weaviate Podcast #99!
Просмотров 8325 месяцев назад
ACORN with Liana Patel and Abdel Rodriguez - Weaviate Podcast #99!
Window Search Tree with Josh Engels - Weaviate Podcast #98!
Просмотров 4095 месяцев назад
Window Search Tree with Josh Engels - Weaviate Podcast #98!
Vector Quantization: The Vector Clubhouse Episode 2
Просмотров 2715 месяцев назад
Vector Quantization: The Vector Clubhouse Episode 2
AI Renaissance Berlin - AI Buzzwords
Просмотров 2015 месяцев назад
AI Renaissance Berlin - AI Buzzwords
The Future of Search with Nils Reimers and Erika Cardenas - Weaviate Podcast #97!
Просмотров 1,3 тыс.5 месяцев назад
The Future of Search with Nils Reimers and Erika Cardenas - Weaviate Podcast #97!
Deep Learning with Letitia Parcalabescu - Weaviate Podcast #96!
Просмотров 4535 месяцев назад
Deep Learning with Letitia Parcalabescu - Weaviate Podcast #96!
All Your Vector Embeddings Are Belong To You
Просмотров 8366 месяцев назад
All Your Vector Embeddings Are Belong To You
Open Source RAG running LLMs locally with Ollama
Просмотров 29 тыс.6 месяцев назад
Open Source RAG running LLMs locally with Ollama
Guest Lecture: Vector Quantization Techniques with Etienne | Brown University CSCI
Просмотров 5806 месяцев назад
Guest Lecture: Vector Quantization Techniques with Etienne | Brown University CSCI
DSPy End-to-End: Meetup in San Francisco
Просмотров 6 тыс.6 месяцев назад
DSPy End-to-End: Meetup in San Francisco
Google Cloud Marketplace with Dai Vu and Bob van Luijt - Weaviate Podcast #95!
Просмотров 3466 месяцев назад
Google Cloud Marketplace with Dai Vu and Bob van Luijt - Weaviate Podcast #95!
ParlayANN with Magdalen Dobson Manohar - Weaviate Podcast #94!
Просмотров 3917 месяцев назад
ParlayANN with Magdalen Dobson Manohar - Weaviate Podcast #94!
Please avoid using background music for technical videos
If interested, Erika's talk from Google Pier 57 is live on Arize AI RUclips!
Great
very excited to learn from you, i remain your thankful !
Wow! Someone put the head out of the box!
Intro music goes incredibly hard
Nice explanation
This seems to be a detailed method of RAPTOR RAG applied to context and implied extension...can we now go further filling the gaps of basic language mechanics that should have been applied over 40yrs ago, and finally put this primitive stuff behind us getting to the good stuff🤔🧐😁
You guys look like you're having so much fun haha
Oh looking forward to this!
Fantastic episode, great explanation!
In-house episode! On a great topic! Exciting
oh thank you !
Very helpful. Thank you.
Can i use ollama?
🤪
Structured Outputs!!
❤ weaviate
WEAVIATE FTW, yaaaaasss 🫰
I can think of many practical use cases where long context can't replace vector retrieval, even if context window size explodes
Suppose we are using 1536 dimensions for chunks and now we moved to create embedding for entire article to implement late chunking. Are not we diluting the information by this way. Because same vector dimension is now representing my entire article. Any thought on this. ?
Thanks for the tutorial!
I mean she explains it beautifully and that's a kind of a explanation I was looking for but what really got me into a state of flow is this ambient Techno track in the background :D
Nice video - very informative.
Thanks so much for joining Philip!
Connor at 50:10 you mention a paper Alto but didnt link it? Or am I mistaken?
Just updated! Sorry about that! Super interesting paper!
@@connor-shorten Thanks
Very informative
Heck yeah!!
Great video, Connor!
Where is the file or path of data that is stored locally?
This is great! Thank You!
what's the difference between collections and schemas ?
amazing idea.
This playlist is very helpful. 👍👍
why don't you use rerank attribute. that is most important attribute
Is that mean we don't need to covert unstructured data into structure for RAG applications but can use vector indexing instead?
Thank you!
Hartmann Loaf
The ads between each video is almost longer than each video. Can't you turn ads off?
Cool.
Thank you so much for joining the podcast Benedikt and Sayash! Learned so much from our chat and really excited about where this research is heading!
Keep up the great work Connor
Thanks so much Karl! Means a lot!
excellent video, and impressed insights.
I am not able to view the Overview
But again sharing private info to this company can be dangerous, so uploading documents is a doubt
Great video about an awesome open source contribution. Thanks!
Adding multimodal support would just make this amazing. Weaviate does come with the multimodal operators
Great video !!!!
Can it use external vector databases?
Weaviate only
Does this get around the issue of inaccuracies in the function call or in the translation from natural to structured language?