PyData Exeter #9: April 2nd @ Innovation Hub

Enjoy three talks from our brilliant speakers after drinks and networking.
Agenda:

• 18:45 Doors open
• 19:15 Talks start
• 20:30 Talks finish - stick around for a drink and networking at the CUCKOO taproom across the road.

Talks:
Modelling human behaviour through a framework of rationality, Harry Findlay
In order to have successful human-machine teaming, sometimes referred to as collaborative AI, one assumption you can make is that the machine must internalise a model of human cognition. This is difficult as humans are incredibly adaptive, and are a unique combination of subjective preferences and physical and mental capacities. Developing a theory-driven, generative model of human cognition has provided a useful avenue to tackle this challenge. In this talk I will introduce such a theory-driven model, called computational rationality that put simply, assumes human behaviour can be predicted by the optimal solution to a constrained optimisation problem. I will then present how ideas from multimodal information fusion can be applied to computational rationality, and how a focus on how humans perceive and represent information could improve our models of human behaviour.

Maths Proofs in Lean - Introduction & Live Demo, Tariq Rashid
Open source Lean with Mathlib as an automated proof assistant has seen an unprecedented growth, and even recognition in popular media. It is being used by world-class mathematicians, and is increasingly being taught in undergraduate mathematics courses. This talk provides an overview and a short live demonstration. The talk is aimed at newcomers, not experts, and aims to remove barriers to more people trying to write their own simple proofs.

Verifiable AI, Amy Stell
A brief look into how we go about formalising and verifying neural networks in safety critical domains.

Sponsors: Butterfly Data, Exeter Innovation Hub, NumFOCUS

A little more about our speakers:
Harry Findlay is currently a PhD student in computer science at the University of Exeter, where he also completed his undergraduate degree. His research direction has evolved from formalising compiler optimisation as a reinforcement learning problem, to now ultimately developing enhanced assistive systems deployed in human-machine teaming scenarios by developing models of human cognition that the machine can use to compute the optimal, human-aligned intervention.

Tariq Rashid was originally trained as a physicist, and later gained a masters in machine learning and data mining. He’s worked in technology for 25 years, including almost a decade in central government, leading on the modernisation of technology and security. Tariq is passionate about open source, developing communities, and inspiring the next generation of scientists and engineers. He led the London Python meetup, children’s CoderDojo Cornwall, the Data Science Cornwall community, as well as the London based Algorithmic Art group which grew to over 4,500 members. He also writes books on machine learning and creative computing. He is currently developing Digital Dynamics, a business helping organisations ensure their use of data and automation through machine learning is safe, fair and ethical.

Amy Stell is a PhD student working on verifiable AI, with a focus on ensuring neural networks are safe and secure. She also works for Code First Girls in bridging the gender gap in technology.

CODE OF CONDUCT
The PyData Code of Conduct governs this meetup ([https://pydata.org/code-of-conduct/](https://pydata.org/code-of-conduct/)). To discuss any issues or concerns relating to the code of conduct or behaviour of anyone at the PyData meetup, please contact the PyData Exeter organisers, or you can submit a report of any potential Code of Conduct violation directly to NumFOCUS (https://numfocus.typeform.com/to/ynjGdT)