Unstructured Data Meetup South Bay Edition
This is an in-person event! Registration is required in order to get in.
Topic: Connecting your unstructured data with Generative LLMs
What we’ll do:
Have some food and refreshments. Hear three exciting talks about unstructured data and generative AI.
5:30 - 6:00 - Welcome/Networking/Registration
6:05 - 6:30 - Dinesh Chandrasekhar, Challenges in Structured Document Data Extraction at Scale with LLMs
6:35 - 7:00 - James Luan, Dense Embeddings != Complete Search - a sneak peak of Milvus 2.5
7:05 - 7:30 - Rob Quiros, Beyond RAG Partitions: Per-User, Per-Chunk Access Policy
7:30 - 8:00 - Networking
Tech Talk 1: Challenges in Structured Document Data Extraction at Scale with LLMs
Speaker: Dinesh Chandrasekhar, Unstract
Abstract: All businesses have to deal with unstructured documents at some level. Some have to deal with them at scale. While an LLM-powered approach to this problem is most certainly head and shoulders above traditional machine learning-based approaches, it is not without its challenges. Top concerns being accuracy and cost, which can really begin to hurt at scale.
In this talk, we will look at how Unstract, an open source platform purpose-built for structured document data extraction, solves these challenges. Dealing with 5M+ pages of structured content extraction per month, Unstract uses various techniques to attain accuracy and cost efficiency.
Topics Covered
- Introduction to Unstructured Data Processing
- Processing Document Data
- Extraction Difficulties
- Unstract to the rescue
- Demo
Tech Talk 2: Dense Embeddings != Complete Search - a sneak peak of Milvus 2.5
Speaker: James Luan, VP of Engineering, Zilliz
Abstract:
Dense embeddings miss exact matches. Keyword search misses semantic meaning. Running two separate systems is a maintenance nightmare. We'll show how Milvus 2.5's hybrid search tackles this with a unified solution, preview its sparse-based BM25 implementation, and share performance numbers against current Elasticsearch-based architectures.
Key Points:
Where dense embeddings fall short and how a unified system architectures address the search needs
Sneak Peak of Milvus 2.5 - Quick look at our BM25 implementation and sparse vector optimizations
Benchmark results comparing hybrid search latency and throughput vs ElasticSsearch
What's Next - Brief overview of upcoming features in our technical roadmap
Tech Talk 3: Beyond RAG Partitions: Per-User, Per-Chunk Access Policy
Speaker: Rob Quiros, CEO & Co-Founder, Caber Systems, Inc.
Abstract: Partitioning vector databases has proven to be a useful tool for privacy and per-tenant isolation. Recent releases of vector db software, including Milvus, have continued to improve partitioning capabilities such as pushing the number of partitions into the millions and providing improved selection of partitions per tenant.
Despite these advances, management overhead increases with the number of partitions. Relative to the capabilities enterprises require and have come to expect from their existing storage systems and databases, there is still a shortfall. New capabilities specific to how vector databases store data and how they are used in RAG applications are needed.
Topics Covered:
Origins of enterprise requirements for granular access control and policy in storage systems.
Sensitive data identification: data classification versus access control.
The problem data-duplication in enterprise datasets presents when copying permissions from documents to chunks.
How enterprise access requirements can be met with per-user, per-chunk access control
Case study and example implementation.
When:
Nov 13, 2024
5:30PM
Where:
This is an in-person event. Registration is required to get into the event. Registration in advance will close 2 days before the event.