•

•

•

Garden State Java User Group

May 2026 Meeting Information

We are happy to announce that Doug Drechsel will be our special guest for this month's meeting.

Traditional location tracking systems are great at answering "where," but fail to answer "what is happening?" This creates a massive problem in real-world software engineering: operators are flooded with false alerts because a safe, normal event (like driving past a school) triggers the exact same alarm as a genuine threat (like loitering near a school).

In this talk, we'll solve this context problem by building a practical, hands-on application of vector spaces. While vectors are incredibly popular right now for LLMs and text search, we will look at how to encode physical behavior - like speed, dwell time, and proximity - into a simple, 3-dimensional vector space.

You don't need any prior experience with spatial data to follow along. Using Oracle AI Database 26ai, we will walk through the database architecture and SQL required to combine basic location tracking with AI Vector Search. By the end of the session, you'll see how creating a simple vector space can reduce alert noise by up to 90% in high-stakes environments like parolee tracking and museum security.

What you will learn: Applied Vector Spaces: How to design and create a simple behavioral vector to solve a tangible logic problem The Context Gap: Why raw location data isn't enough, and how encoding behavior fixes it. Code in Action: A practical look at fusing spatial filters (SDO_GEOMETRY) and vector math (VECTOR_DISTANCE) in a single query to prioritize real threats without missing true positives.