Multi-agent systems & convex optimization

 

Venue

The Cologne CAIML meetup no. 28 took place on January  23, hosted and sponsored by Scieneers GmbH with two presentations on large language models (LLM) and the well-established subsequent networking session.

A very special thanks for the excellent moderation goes to Dr. Fabian Hadiji, head of BI department at Lotum

Speakers and talks

More than the sum of its parts – Building Multi-Agent systems with AutoGen

“Agents that combine the intelligence of LLMs (Large Language Models) with real code execution are already very powerful. However, they are limited by context and can easily start hallucinating if the prompting is not specific enough. This talk will show how multiple agents can be combined to solve heterogeneous tasks. It will also introduce the AutoGen Framework and discuss the problems associated with using multiple agents.” ~ Nico Kreiling, Data Scientist at Scieneers GmbH

Convex optimization and word-embeddings

“After the LLM Big-Bang over one year ago, the field of (word-)embeddings has (paradoxically) gained a lot of momentum. That is remarkable, because as transformers seamlessly process Natural Language into Natural Language (or more generally, anything into whatever), what should one gain by looking into their “guts” (i.e. their sub-symbolic “thoughts”)? In this presentation, we will go beyond the King-Queen example and the nearest-neighbors search, to illustrate how, based on convex-optimization, it is possible to make high-quality inferences with sets of embeddings. An application of the technique will be illustrated with an example from the domain of knowledge graphs.” ~ Ivan Herreros, PhD Computational Neuroscience

Video impressions

www.instagram.com/reel

 

Slides

A big thank you to the speakers Nico and Ivan. You can find the corresponding slides here:

Networking

Thanks to the good organization and the strong Meetup group from CAIML, the event at Scieneers GmbH was well attended by over 60 people interested in AI and its application. It was a very productive evening thanks to the numerous personal contacts and the stimulating exchanges and technical discussions.

Many thanks also to the host and sponsor, Dr. Lars Perchalla, for providing the central location in Cologne and the delicious catering.

CAIML

Cologne AI and Machine Learning Meetup: www.meetup.com/de-DE/cologne-ai-and-machine-learning-meetup. We are looking forward to seeing you again at the next CAIML event on March 19 in Cologne.

Further information


www.scieneers.de

Retrospective

Cologne AI ML – Data mesh

Cologne AI ML #23

Applying ML to challenging clinical data

Deep Learning Cologne

CERN vs. German Mittelstand

AI for social good

Current state of AI

Automated Data Science