We use cookies to ensure that we give you the best experience on our website. You can change your cookie settings at any time. Otherwise, we'll assume you're OK to continue.

# Research lectures, seminars and events

The events listed in this area are research seminars, workshops and lectures hosted by Durham University departments and research institutes. If you are not a member of the University, but  wish to enquire about attending one of the events please contact the organiser or host department.

November 2020
SunMonTueWedThuFriSat
October 2020 December 2020
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30

## Events for 13 November 2020

### Marcin Lis: On delocalization in the six-vertex model.

1:00pm, Zoom

In this talk I will show that the six-vertex model with parameter c in [\sqrt 3, 2] on a square lattice torus has an ergodic infinite-volume limit as the size of the torus grows to infinity.
Moreover I will prove that for c in [\sqrt{2+\sqrt 2}, 2], the associated height function on the infinite square lattice has unbounded variance.

The proof relies on an extension of the Baxter--Kelland--Wu representation of the six-vertex model
to multi-point correlation functions of the associated spin model.
Other crucial ingredients are the uniqueness and percolation properties of the critical random cluster measure for $q\in[1,4]$, and recent results relating the decay of correlations
in the spin model with the delocalization of the height function.

Best,
Marcin

### Neelanjan Bhowmik and Jack W Barker: The Good, the Bad and the Ugly: Evaluating Convolutional Neural Networks for Prohibited Item Detection Using Real and Synthetically Composited X-ray Imagery

1:00pm, Online

Detecting prohibited items in X-ray security imagery is pivotal in maintaining border and transport security against a wide range of threat profiles. Convolutional Neural Networks (CNN) with the support of a significant volume of data have brought advancement in such automated prohibited object detection and classification. However, collating such large volumes of X-ray security imagery remains a significant challenge. This work opens up the possibility of using synthetically composed imagery, avoiding the need to collate such large volumes of hand-annotated real-world imagery. Here we investigate the difference in detection performance achieved using real and synthetic X-ray training imagery for CNN architecture detecting three exemplar prohibited items, {Firearm, Firearm Parts, Knives}, within cluttered and complex X-ray security baggage imagery. Our extended evaluation demonstrates both challenge and promise of using synthetically composed images to diversify the X-ray security training imagery for automated detection algorithm training. Additionally, what's the future on synthetic data generation techniques: we introduce briefly our initial works on synthetically composited X-ray imagery using Generative Adversarial Networks (GAN).