Mini-Projects

Each year, AIMS holds an Annual Meeting where our students and industrial partners come together.  This year, the first (in person) event for over 18 months, took place at Lady Margaret Hall.

As part of the event, our students get to showcase their mini projects with a poster session, and present their research.

Below are the posters which were part of the Research Showcase.

Cohort 2020

• Ondrej Bajgar - Epistemic Max-Min with Bayesian Optimisation
• Freddie Bickford-Smith - Continual Learning via Function-Space Variational Inference
• Jonathan Carter - Remote Heart Rate Monitoring on the Ward
• Benjamin Ellis - Investigating Ration Clipping in Multi-agent Reinforcement Learning
• Francisco Girbal Eiras - ANCER: Anisotropic Certification via Sample-wise Volume Maximization
• Laurynas Karazija - CLEVRTEX: A Texture-Rich Benchmark for Unsupervised Multi-Object Segmentation
• Dominik Kloepher - Learning How to Learn Where You Are: Meta-Learning for Few-Shot Camera Localization
• Pierre Osselin - Certifiable Robustness in Graph Classification via Community-Aware Randomized Smoothing
• Benjamin Ramtoula - Better Segmentation by Adversarial Attacks on Style
• Lisa Schut - Can we use Ensemble Uncertainty in the Infinite Width Limit?
• Aleks Shtedritski - Text-Conditional Image Generation from Discrete Representations
• Alex Stephens - Dissimilar Path Generation for Multi-Agent Pathfinding via Combinatorial Auction
• Baskaran Sripathmanathan - Paying Attention to Curvature on Surfaces with Graph Convolutions
• Charig Yang - It's About Time: Analog Clock Reading in the Wild
• Zheng Xiong - On the Practical Consistency of Meta-Reinforcement Learning

Cohort 2019

• Jan Brauner & Mrinank Sharma - How Robust are the Estimated Effects of Nonpharmaceutical Interventions against COVID-19?
• Alec Edwards - Formal Synthesis of Lyapunov Functions and Barrier Certificates
• Jonas Beuchert & Amanda Matthes - SnapperGPS A Small, Low-Cost, Low-Power Wildlife Tracking System
• James Fox - Multi-agent Influence Diagrams Offer a Compact and Complete Alternative Graphical Representation of Games
• Shu Ishia - Towards Real-world Navigation with Deep Differentiable Planners
• Cong Lu - Augmented World Models Facilitate Zero-Shot Dynamics Generalization From a Single Offline Environment
• Tim Reichelt - Adapting Probabilistic Programming Systems to Estimate Expectations Efficiently