Location: William Gates Building, Department of Computer Science and Technology, 15 JJ Thomson Avenue, Cambridge CB3 0FD
Programme Overview
NeurIPS@Cam 2025 is part of the AI for Science Summit (9 to 10 December 2025) hosted by the Accelerate Programme for Scientific Discovery, a two day event celebrating advances in machine learning and their impact across scientific domains. On the second day of the Summit, NeurIPS@Cam brings together students, researchers, and engineers across the Cambridge community for a full day dedicated to discussing work presented at NeurIPS and related advances in AI research. The day is organised in partnership with the Cambridge Centre for Data Driven Discovery (C2D3), and the first half of the programme is delivered jointly with C2D3.
The joint C2D3/NeurIPS@Cam morning session will include talks on new ways to solve problems along with a keynote. The afternoon programme (see C2D3 events page for parallel session run by C2D3) consists of two presentation sessions followed by a panel discussion. All speakers present their ongoing research, and the event provides a space to share ongoing research, exchange ideas, and strengthen connections across the local AI community.
Joint C2D3/NeurIPS@Cam - Session 1
09:40 – 11:15
| Time | Speaker | Talk |
|---|---|---|
| 09:45 – 10:00 | Tom Cowperthwaite | GPGreen: Learning Linear Operators using Gaussian Processes |
| 10:00 – 10:15 | Yihang She | PILA: Physics-Informed Low Rank Adaptation for Interpretable Earth Observation |
| 10:15 – 10:30 | Anna Pazola | Predicting and Interpolating Spatiotemporal Environmental Data: A Case Study of Groundwater Storage in Bangladesh |
| 10:30 – 10:45 | Xiangjian Jiang | TabStruct: Measuring Structural Fidelity of Tabular Data |
Break (45 minutes)
Joint C2D3/NeurIPS@Cam - Keynote
11:30 – 12:30
Title: Unveiling Social Tipping Mechanisms in Climate Discourse Using Computational Social Sciences
Speaker: Dr Ramit Debnath
Lunch (1 hour)
NeurIPS@Cam - Session 2A
13:30 – 14:55
| Time | Speaker | Talk |
|---|---|---|
| 13:35 – 13:45 | Zakhar Shumaylov | When is a System Discoverable from Data? |
| 13:45 – 13:55 | Tiansi Dong | Neural Reasoning for Sure Through Constructing Explainable Models |
| 13:55 – 14:05 | John Bronskill | Stroke Patches: Customizable Artistic Styling Using Regression |
| 14:05 – 14:15 | Oscar Hill | Hierarchical Concept-based Interpretable Models |
| 14:15 – 14:25 | Cristiana Diaconu | Estimating Interventional Distributions with Uncertain Causal Graphs through Meta-Learning |
| 14:25 – 14:35 | Shangqi Gao | Towards Generalizable Retinal Vessel Segmentation with Deformable Graph Priors |
| 14:35 – 14:45 | Mohammad Zaid | AI-driven MEMS Acoustic Sensors |
| 14:45 – 14:55 | Amy Yijie Zheng | Towards Equitable Photoacoustic Imaging: Digital-Twin Inspired Skin Pigmentation Bias Correction |
Break (20 minutes)
NeurIPS@Cam - Session 2B
15:15 – 15:45
| Time | Speaker | Talk |
|---|---|---|
| 15:15 – 15:25 | Rares Grozovescu | Koopman Autoencoder for Ocean State Forecasting |
| 15:25 – 15:35 | Wagar Ashraf | Neural Network-enabled Domain-consistent Robust Optimisation for Global CO₂ Reduction Potential of Gas Power Plants |
| 15:35 – 15:45 | Sinan Deger | AI-enabled Insights Into Galaxy Evolution |
NeurIPS@Cam - Panel Discussion
15:45 – 16:30
Topic:
Beyond Tokens and GPUs: What Survives When the AI Bubble Pops?
Moderator: Emilio Luz-Ricca (AI4ER)
Panellists:
Simon Driscoll (AI4Climate, ICCS), Kristina Khutsishvili (AI Ethics, ai@cam), Carl Henrik Ek (Core AI/ML, Cambridge), Adrian Alexa (Daset Labs, AI4HealthTech)
Drinks Reception (30 minutes)
Contact
For questions regarding NeurIPS@Cam 2025, please contact: Pritthijit Nath (pn341@cam.ac.uk)
