Weaviate Podcast

Weaviate Podcast

By WeaviateTechnology
Download on the App Store

Weaviate Podcast episodes

  • Charles Frye on Full Stack Deep Learning - Weaviate Podcast #57!

    Hey everyone! Thank you so much for watching the 57th Weaviate podcast with Charles Frye! Charles is an educator at Full Stack Deep Learning, one of the world's top courses on Deep Learning with lectures available on YouTube (link below)! This was one of the most thorough Weaviate podcasts published so far, covering all sorts of topics around the evolution of Deep Learning! Particularly we discussed the Retrieval-Augmented Generation stack with Vector Databases and Zero-Shot Large Language Models and how that compares to more conventional machine learning workflows and the MLOPs stack! I really enjoyed chatting with Charles and am more than happy to answer any questions or discuss any ideas you have about the content in the podcast! Thank you so much for listening!

    Check out Full Stack Deep Learning! https://fullstackdeeplearning.com/
    Full Stack Deep Learning on YouTube! https://www.youtube.com/@The_Full_Stack
    Chapters
    0:00 Welcome Charles Frye!
    0:52 Charles’ journey into Deep Learning
    3:00 Weights & Biases and MLOps
    5:30 Retrieval-Augmented Generation Stack
    8:58 Data Engines and AI Products
    13:50 Fine-Tuning
    16:35 Information Retrieval Techniques
    20:10 RAG as Tool Use and RETRO
    23:33 Gorilla and Fine-Tuned Tool Use
    27:36 Text-to-SQL Tool Use
    30:46 Generative Data Augmentation
    33:05 LLM generated queries for embeddings
    38:04 Long-Tail and Data Imbalance
    41:45 LoRA LLM Fine-Tuning
    44:50 Eigenvectors and Disentaglement
    50:00 LLM for Each User
    55:00 Embedding Visualization and ML Observability
    58:40 GPU Utilization
    1:05:05 Discord Q&A Bot App
    1:16:10 Data Schema Design
    1:21:25 Graph and Vector Databases
    1:28:35 Future Directions in AI

    1 hr 40 min
  • Aleksa Gordcic - Weaviate Podcast #55!

    Hey everyone! Thank you so much for watching the 55th episode of the Weaviate Podcast with Aleksa Gordcic! This episodes dives into Aleksa's incredible story from Deep Learning YouTube to DeepMind and now creating Ortus! We dived into all sorts of topics, I loved hearing about the latest updates on Ortus and how Aleksa is sees the current state of AI development! We are more than happy to answer any questions or discuss any ideas you might have about the content in the podcast! Thanks so much for watching!

    Check out Ortus here! - https://www.ortusbuddy.ai/welcome
    Chapters
    0:00 Introduction
    1:08 Deep Learning YouTube
    5:40 DeepMind
    9:40 Ortus
    19:50 LangChain and LlamaIndex
    23:10 Software 2.0 and Full Stack DL
    29:20 Training Embedding Models
    32:23 Text Chunking for Vector DBs
    34:35 Visual Information in YouTube
    38:15 Simulating Conversations
    42:46 Aidan Gomez Quote on Synthetic Data
    44:40 Tree of Thoughts
    47:40 New Ortus Features
    49:00 Embedding Marketplace
    54:00 Personal Organization

    1 hr 8 min
  • Stephanie Horbaczewski and Gunjan Bhattarai on Vody - Weaviate Podcast #53!

    Chapters

    0:00 Introduction
    0:38 Founding Story of Vody
    8:15 Custom Embedding Models
    12:42 Movie Genre Vectors
    13:42 Classification and Contrastive Learning
    15:45 Foundation Model Tuning
    21:13 Multimodal Generative Models
    25:08 Training Embedding Models
    33:20 Tabular Data Ranking Models
    36:00 RoomGPT
    41:36 Diversity in Recommendations
    48:25 Future Directions in Multimodal AI
    51:15 Open-Source
    55:45 Keeping up with Vody!

    57 min
  • Yana Welinder on Kraftful - Weaviate Podcast #52!

    Hey everyone, thank you so much for watching the 52nd episode of the Weaviate Podcast with Yana Welinder! Yana is the Founder and CEO of Kratful (https://www.kraftful.com/). Kratful is an incredibly interesting "ChatGPT but for Product Research" -- curating specific skills for Product Managers into a collection of prompts. We discussed all sorts of things from the latest innovations in LLMs to the ChatGPT marketplace and product management, I really hope you enjoy the podcast!

    42 min
  • Greg Kamradt and Colin Harmon on LLM Agents - Weaviate Podcast #51

    Hey everyone, thank you so much for watching the 51st episode of the Weaviate Podcast with Greg Kamradt and Colin Harmon! Greg and Colin are both entrepreneurs in the space of new AI tools powered by LLMs! This podcast is about keeping up with the evolution of LLM Agents from AutoGPT to connecting LLMs with Vector Databases or Wolfram Alpha, as well as the ChatGPT Marketplace, Personalized LLMs, Private LLMs, and many more! I think there are so many interesting nuggets from this podcast, thank you so much to Greg and Colin for joining, really enjoyed this one!

    Data Independent: https://www.youtube.com/@DataIndependent
    Greg Kamradt on Twitter: https://twitter.com/GregKamradt
    Nesh: https://hellonesh.io/
    Colin Harmon on LinkedIn: https://www.linkedin.com/in/coluha/
    Colin Harmon Blog: https://colinharman.substack.com/
    Colin Harmon at Haystack US 2023: https://www.youtube.com/watch?v=LO3U5iqnTpk
    Chapters
    0:00 Introduction
    0:42 Backgrounds
    2:43 Defining “LLM Agents”
    6:12 Data-Aware LLMs
    13:04 Tool Use
    13:38 ChatGPT API vs. Marketplace
    17:40 Personalized LLMs, LLM for Greg
    19:20 PrivateGPT
    25:14 AutoGPT and Chain-of-Thought Prompting
    32:30 Few-Shot Examples
    35:30 Early AI Signals and Open-Source
    43:10 Multi-Agent LLMs
    47:14 Fine-Tuning and Long Input Lengths
    52:20 Greg’s LLM Wishlist Hierarchy
    53:15 Keeping up with Greg and Colin!

    55 min
  • Retrieving Texts based on Abstract Descriptions Explained!

    This video explores a new paper exploring the use of summarization chains to represent long texts and use (original text, summary) pairs for optimizing text embeddings models! Here are 3 main takeaways I think everyone working with Weaviate may get value from:

    1. Understanding of Summary Indexing and the Prompts (as well as Prompt Chains) used to build them.
    2. Continued development of LLM-generated data for search -- creating (full text, summary) pairs gives you (1) data to build a summary index with as mentioned, (2) data to compare different embedding models with, and (3) data to train your own embedding model.
    3. Tournament style evaluation with human annotators -- the top 5 retrieved texts from one model are concatenated with the top 5 from another model, these 10 are given to human annotators to pick 5 and this is how the authors are reporting the performance of their models rather than traditional benchmarks. This m ay be a more productive evaluation technique for most real world search applications.
    Thank you so much for watching, here are some links mentioned in the video!
    Retrieving Texts based on Abstract Descriptions: https://arxiv.org/abs/2305.12517
    Weaviate Blog - Combining LangChain and Weaviate: https://weaviate.io/blog/combining-langchain-and-weaviate
    Weaviate Blog - Generative Feedback Loops: https://weaviate.io/blog/generative-feedback-loops-with-llms
    Jerry Liu in Llama Index Blog - A New Document Summary Index for LLM-powered QA Systems: https://medium.com/llamaindex-blog/a-new-document-summary-index-for-llm-powered-qa-systems-9a32ece2f9ec
    Learning to Retrieve Passages without Supervision (Spider): https://arxiv.org/pdf/2112.07708.pdf
    Weaviate Blog - Analysis of Spider - https://weaviate.io/blog/research-insights-spider
    Chapters
    0:00 Introduction
    0:13 Quick Overview
    7:30 How to use in Weaviate!
    7:50 Background
    12:08 Motivation
    14:20 Prompts Used
    18:14 More Details of training
    21:12 Human Evaluation Study
    22:40 My Takeaways from the Paper

    29 min
  • Kapa AI with Emil Sorensen and Finn Bauer - Weaviate Podcast #50!

    Hey everyone, thank you so much for watching the 50th (!!!) Weaviate Podcast with Emil Sorensen and Finn Bauer from Kapa AI! Are you curious about taking either your, or your company's, specific information and putting into a Vector DB + LLM system? Emil and Finn are doing this at the highest level, taking the documentation of software companies like Weaviate and building these LLM-augmetnted assistant systems for them. This podcast takes a complete tour from Data Ingestion to Cleaning, Chunking, LLM latency, and emerging trends in LLMs such as cheap fine-tuning with LoRA or Long Context Windows such as GPT-4 32K, MPT-7B 65K, or Anthropic Claude's 100k. I learned so much from speaking with Emil and Finn! Please let us know any questions you have or ideas you would like to discuss!

    Check out Kapa here! https://www.kapa.ai/
    Chapters
    0:00 Welcome Emil and Finn!
    0:42 Origin Story of Kapa
    2:08 Data Ingestion
    5:10 Data Cleaning
    6:20 Slack / Discord / Forum Ingestion
    9:05 Testing Models on Support QA
    11:14 Selling Kapa to Weaviate and friends
    12:37 Hallucinations in LLMs
    14:06 Trends in Open-Source LLMs
    15:20 Long Input LLMs (32K, 65K, 100K, …)
    16:54 Retrieval-Augmentation for Long Input LLMs
    18:08 Fine-Tuning LLMs
    23:00 As much or as refined content as possible?
    24:40 Adding Docs from Integrations
    26:15 Generative Feedback Loops
    29:00 What in AI excites you the most?

    36 min
  • Neurosymbolic AI in Search with Professor Laura Dietz - Weaviate Podcast #49!

    Hey everyone, thank you so much for watching the 49th episode of the Weaviate Podcast!! This podcast features Professor Laura Dietz from the University of New Hampshire! I came across Dr. Dietz's tutorial at ECIR on Neuro-Symbolic Approaches for Information Retrieval and am so grateful that she was interested in joining the Weaviate Podcast! I learned so much about Neurosymbolic Search, especially around the role of Entity Linking and Entity Re-Ranking -- as well as the topic of Knowledge Graphs and Vector Search. We also discussed Prof. Dietz and collaborators latest perspectives paper on Large Language Models for Relevance Judgment. TLDR this describes the idea of using LLMs to either generate synthetic queries for documents or say annotate the relevance for query, document pairs. We discussed this kind of idea with Leo Boytsov and his work on InPars, and have presented Promptagator on past episodes of the Weaviate Air show. Although this idea comes with a lot of potential, Dr. Dietz explains the potentials for bias and poor judgements, as well as generally diving more into the details of this kind of idea! I really hope you enjoy the podcast, we are more than happy to answer any questions you might have about these ideas, or discuss any of your ideas! Thanks so much for watching!

    Check out Laura Dietz's Publications here: https://scholar.google.com/citations?user=IIXpJ8oAAAAJ&hl=en&oi=ao
    ECIR 23 Tutorial: Neuro-Symbolic Approaches
    for Information Retrieval: https://www.cs.unh.edu/~dietz/appendix/dietz2023neurosymbolic.pdf
    Please check this paper out below, I think this is a severely underrated work in the Search / Information Retrieval community:
    Perspectives on Large Language Models for Relevance Judgment: https://arxiv.org/pdf/2304.09161.pdf
    Chapters
    0:00 Introduction
    0:15 Neurosymbolic Search
    4:50 Entity Parsing and Vector Semantics
    10:56 Query Intent Understanding
    15:35 Knowledge Graphs and Vector Search
    17:37 Symbolic Re-Ranking
    22:10 ColBERT and Entity Ranking
    26:25 Example - South America and Zika Virus IR
    29:15 Knowledge Graph Query Languages with LLMs
    35:10 We need more Knowledge Graphs!!
    37:30 PrimeKG from Harvard BMI
    39:40 Filtered Vector Search
    42:20 LLM Entity Linking - “The” example
    47:30 Cross Encoder Entity Focus?
    48:25 Perspectives on LLMs for Relevance Judgments
    55:28 Spectrum of Human-Machine Collaboration for Labeling
    57:30 Use LLM to Create Relevance Labeling Interfaces
    1:02:30 Importance for Weaviate
    1:03:45 12 Authors’ 3 Conclusions
    1:04:40 IR Research Community Challenge
    1:06:55 Query Generation for Weaviate Users
    1:13:05 Clustering Queries
    1:17:30 Final Thoughts

    1 hr 31 min
  • Unstructured with Brian Raymond - Weaviate Podcast #48!

    Hey everyone, thank you so much for watching the 48th episode of the Weaviate Podcast!! This is a SUPER exciting one, welcoming Brian Raymond the CEO / Founder of Unstructured! Unstructured is a perfect complimenting technology for Weaviate, helping people get their Unstructured data into Weaviate! The podcast dives into the nuances of this task, but it generally revolves around Unstructured's abstraction of Partitioning, Cleaning, and Staging! Unstructured is making groundbreaking innovations on using Visual Document Layout models for Partitioning, for example saying that this part of the PDF is the header, body, image caption, and so on. Cleaning then describes removing pesky details like whitespaces or odd characters. Staging then describes the transformations of say formatting a text chunk with it's metadata into the JSON for a Weaviate object upload! I really hope you find this podcast interesting! We are publishing a blog post as well showing an example of how to use Unstructured to get PDF data into Weaviate, please please check that out and let us know if it works for your data and how we can improve it! This blog post can be found on weaviate.io and we will be managing discussions around it both in the Weaviate slack, as well as Unstructured! Thank you so much for listening!

    Check out Unstructured here! https://www.unstructured.io/
    Chapters
    0:00 Welcome Brian!!
    0:27 What is Unstructured?
    5:42 Why now? New Advancements in Unstructured
    8:02 Thoughts on Data Connectors Hub
    10:55 PDFs to Weaviate with Unstructured
    13:53 State-of-the-Art in OCR and Document Parsing
    16:10 How to get the data from Weaviate.io?
    18:06 Foundation Models from Unstructured
    20:45 Evaporate-Code+
    23:15 CSV, Parquet, JSON transformations in Staging
    25:08 Cleaning Bricks
    28:02 Visual Document Examples
    30:45 Text Chunking with Metadata
    33:25 Knowledge Graphs with Goldman Sachs example
    39:10 LLM Hallucinations
    42:10 Announcements from Brian!

    44 min

About Weaviate Podcast

From the publisher's feed

Join Connor Shorten as he interviews machine learning experts and explores Weaviate use cases from users and customers.