GenAI Network inaugural meetup!

We're excited to announce the inaugural meetup of the GenAI Network Melbourne meetup!

We're keen to bring people together to discuss generative AI, its potential application in industry and impact on society. Our goal is to ensure that Australia adapts to this new technology and paves a new way forward for the benefit of all.

Join us at 511 Church St, East Richmond as we explore various ideas around GenAI and connect with like minded people!

We'll have two speakers for this session (and pizza!):

Semantic hide and seek: a gentle introduction to embeddings with a vector search game - Dave Colls

LLMs have thrust embeddings and vector search to prominence, with many new technologies in this space. Embeddings capture some useful semantic - or meaningful - relationships between items. But how do you get your head around meaning in 500 dimensions?

We’ll explore embeddings with a game of semantic hide and seek. Fans of the Wordle gamification universe may be familiar with Semantle. In this game, players seek a hidden word. A little like the game of Marco Polo, players “shout” out guesses, and the game “responds” with a semantic similarity score, from (roughly) 0 to 100. Behind the scenes, embeddings are doing all the work.

We’ll see when Semantle does and doesn’t match our intuition for similar meaning - as not all embeddings are equal. We’ll also explore how biases can exist in embeddings. We’ll compare human and multiple machine-driven game strategies and their robustness to divergent semantics. We’ll discuss what all this means for building solutions with the current generation of LLMs.

Maybe we’ll even have some live demos and see if ChatGPT can play this game of semantic hide and seek!

Engineering Generative AI: A reflection on Intelligence, Agency, Learning, Representation, Language & Emergence - Rajesh Vasa

Infusing non-deterministic Generative AI components into existing real-world systems creates unique engineering challenges. For instance, testing of linguistic interfaces, change management due to components that are adaptive beyond the speed of humans, and project management & integration issues driven by use of co-pilots. These challenges can be (partially) addressed when we see the core concepts from a philosophical perspective.

In this short talk, I want to help present ideas to add nuance & depth to the solution concepts in generative AI.

Speaker bios:

Dave Colls: Dave is at home on the leading edge of technology, with 25 years’ experience delivering R&D, world first software solutions, change programs and strategy in AI, data and digital. Dave has worked extensively with academia, industry, enterprise and government. Over the last 6 years, Dave has led the development of Thoughtworks Data and AI capabilities and service offerings.

Rajesh Vasa: Professor Vasa specialises in building robust AI infused technology. He has a strong record as a technology leader with work spanning defence technologies for contextual adaptation, smart manufacturing processes, building intelligent homes, data driven traffic congestion models, large scale complex simulations to support decision making, and improving patient health outcomes in clinical settings.

His core research area is in improving robustness in complex systems with an active focus on “AI infused systems”; he actively publishes in top ranked journals and conferences in software engineering/complex system development.

Currently, Rajesh leads a multi-disciplinary team focused on solving high- impact societal problems. Prof Vasa brings over a decade of experience from technology industry across the world having held roles spanning engineering and executive leadership. His career spans roles in technology engineering, operations, and executive leadership with an international work profile. The concepts he developed resulted in spinouts of 4 companies; one of his smart-home spinouts listed on ASX in Jan 2022.

He is an advisor to various organisations globally in the space of complex systems, technology development and Artificial Intelligence.