'Stories about Hollywood's hippest mathematics: celebrating women in STEM.' Join us on 12 February 2025 for this public event showcasing female contributions to STEM! Tickets here:

Compass student @babichmorrowc's blog on regional sensitivity analysis: a method for understanding how #uncertainty in model input affects uncertainty in output, particularly in settings with continuous inputs and discrete outputs: #datascience #statistics

Our Compass Conference took place in #Bristol recently, enabling our students, alumni, #industry partners and #research colleagues to explore 'The Future of Data Science'. Read a student perspective of the event: #AI #machinelearning #datascience #maths

Compass CDT student Xinrui Shi's blog: How to optimally allocate a fixed participant number across trials in a pairwise meta-analysis, to achieve the most precise estimates, considering a trade-off between the precision/quantity of individual observations:

New Student perspectives blog post by Compass CDT Student Sam Perren. This blog highlights Sam's research on network meta-analysis, addressing challenges in disconnected networks and single-arm studies, and extending the multinma package.

NEW Student perspectives blog post by Compass CDT student Sam Bowyer: How can we efficiently finetune a pretrained AI model (which is too large for us to train from scratch) on a downstream task whilst accurately estimating the model's uncertainty?

New Student Perspectives blog post by Compass CDT Student Dylan Dijk! The vector autoregressive model is popular for fitting multiple time series. This post highlights the need for a factor adjustment step when dealing with a large number of series.

New blog post by Compass CDT Student Rachel Wood on how can we spot anomalies in networks. In this blog post, Rachel explains how we can do this to identify changes between networks observing the same subjects (e.g. snapshots of the same graph over time).

APERIODIC, a festival of art, science, music and performance, taking place in Bristol from 1- 14 July, is featured in @bristol247 Read more:

New Student Perspectives blog post by Compass CDT Student Qi Chen on Variational inference, which is a popular method for finding approximate posterior distributions in Bayesian settings. Read the full article here:

RT @GreshamCollege: Today at 6pm: Logarithms: Mobile Phones, Modelling & Statistics? Watch: Prof @BristOliver of…

APERIODIC is a two-week festival exploring the realm of aperiodic patterns, from 1st-14th of July at Kit Form Gallery, Stokes Croft. With public talks, music, dance, performances, and an open gallery, the festival is intended for everybody. Find out more:

New Student Perspectives blog post by #CompassCDT Student @henrybourne_ . How do you learn a probability model when you can't normalise it? How can you learn general informative representations from data? How about just avoiding your problems!

RT @BristolUni: Huge congratulations to Dr Laura Monk from @BristolUniMaths who mingled with Hollywood A-listers and fellow top researchers…

New blog post by Compass CDT #PhDstudent Emma Tarmey which introduces the problem of confounder handling within the wider covariate selection problem and model selection for causal inference.

Registration is open for the Workshop on Functional Inference and Machine Intelligence! Held in the School of Math's Fry Building, from 25- 27 March, you can find out more and sign up for free here: #AI #MachineLearning #AIWorkshop #MLWorkshop #DataScience

Registration is open for this year's CMI-HIMR Summer School on Symmetry and Randomness - where probability theory comes together with analysis, geometry and group theory. For more information and to see this year's speakers please visit the website

New blog post by Compass CDT student Emma Ceccherini: over-parametrised models, like neural networks, exhibit a double descent, a second decrease of the test error after the interpolation threshold, contrasting with classic knowledge on overfitting.