## Colloquium - Faculty Threads

- Calendar
- Mathematics & Statistics

- Date
- 09.17.2021 3:30 pm - 4:30 pm

### Description

Title: Faculty Threads

Speakers: Ben Bolker (McMaster), Ian Hambleton (McMaster) and Tom Hurd (McMaster)

Abstract:

Statisticians have spent large amounts of effort on methods for choosing among discrete hypotheses; e.g. for the purpose of choosing among models where different effects are present or absent (null). In multifactorial systems from the biological and social sciences (ecology, epidemiology, economics, psychology ...), effects are never really absent. Approaches to model simplification and inference that work by choosing among discrete null models are computationally wasteful and usually give overconfident predictions. I will discuss some alternatives, including nonparametric and regularized models.

The topology and total curvature of a Riemann surface is determined by a single integer, the Euler characteristic (Leonhard Euler, 1707-1783). In dimension four, the Euler characteristic gives an interesting invariant for finitely presented groups. The talk will survey some recent joint work with Alejandro Adem on this theme.

Like several department members, I’ve recently become involved with NSERC’s EIDM (Emerging Infectious Diseases Modelling) Initiative. This thread talk will introduce stochastic agent based models for air-borne diseases, providing a glimpse of how they are structured and how they relate to the widely used deterministic SIR compartment models.

Date/Time: Friday September 17, 3:30 - 4:30

Location: Virtual

Join Zoom Meeting

Meeting details to follow

Speakers: Ben Bolker (McMaster), Ian Hambleton (McMaster) and Tom Hurd (McMaster)

Abstract:

**Nothing is zero in a multifactorial world (and why it matters) - Ben Bolker**Statisticians have spent large amounts of effort on methods for choosing among discrete hypotheses; e.g. for the purpose of choosing among models where different effects are present or absent (null). In multifactorial systems from the biological and social sciences (ecology, epidemiology, economics, psychology ...), effects are never really absent. Approaches to model simplification and inference that work by choosing among discrete null models are computationally wasteful and usually give overconfident predictions. I will discuss some alternatives, including nonparametric and regularized models.

**Euler Characteristics and 4-manifolds - Ian Hambleton**The topology and total curvature of a Riemann surface is determined by a single integer, the Euler characteristic (Leonhard Euler, 1707-1783). In dimension four, the Euler characteristic gives an interesting invariant for finitely presented groups. The talk will survey some recent joint work with Alejandro Adem on this theme.

**Micro and Macro Models of Infectious Disease Transmission - Tom Hurd**Like several department members, I’ve recently become involved with NSERC’s EIDM (Emerging Infectious Diseases Modelling) Initiative. This thread talk will introduce stochastic agent based models for air-borne diseases, providing a glimpse of how they are structured and how they relate to the widely used deterministic SIR compartment models.

Date/Time: Friday September 17, 3:30 - 4:30

Location: Virtual

Join Zoom Meeting

https://mcmaster.zoom.us/j/92441720085?pwd=VDAwRHQvV0tBK2xZYlloTE45bkRndz09

Meeting details to follow

Go Back