Wed 18 Feb 15:05: Geometric Gaussian Processes
Gaussian processes (GPs) are often considered to be the gold standard in settings where well-calibrated predictive uncertainty is of key importance, such as decision making.
It is important for applications to have a class of “general purpose” GPs. Traditionally, these are the stationary processes, e.g. RBF or Matérn GPs, at least for the usual vectorial inputs. For non-vectorial inputs, however, there is often no such class. This state of affairs hinders the use of GPs in a number of application areas ranging from robotics to drug design.
In this talk, I will consider GPs taking inputs on a manifold, on a node set of a graph, or in a discrete “space” of graphs. I will discuss a framework for defining the appropriate general purpose GPs, as well as the analytic and numerical techniques that make them tractable.
Link to join virtually: https://cam-ac-uk.zoom.us/j/89473073451
This talk is being recorded. If you do not wish to be seen in the recording, please avoid sitting in the front three rows of seats in the lecture theatre. Any questions asked will also be included in the recording. The recording will be made available on the Department’s webpage
- Speaker: Dr Viacheslav Borovitskiy - School of Informatics Institute for Adaptive and Neural Computation, University of Edinburgh
- Wednesday 18 February 2026, 15:05-15:55
- Venue: Lecture Theatre 1, Computer Laboratory, William Gates Building.
- Series: Wednesday Seminars - Department of Computer Science and Technology ; organiser: Ben Karniely.
Wed 28 Jan 15:05: Title to be confirmed
Abstract to be confirmed
Link to join virtually: https://cam-ac-uk.zoom.us/j/89473073451
This talk is being recorded. If you do not wish to be seen in the recording, please avoid sitting in the front three rows of seats in the lecture theatre. Any questions asked will also be included in the recording. The recording will be made available on the Department’s webpage
- Speaker: Professor Steve Awodey - Department of Philosophy, Carnegie Mellon University
- Wednesday 28 January 2026, 15:05-15:55
- Venue: Lecture Theatre 1, Computer Laboratory, William Gates Building.
- Series: Wednesday Seminars - Department of Computer Science and Technology ; organiser: Ben Karniely.
Wed 03 Dec 15:05: Title to be confirmed
Abstract to be confirmed
Link to join virtually: https://cam-ac-uk.zoom.us/j/89473073451
This talk is being recorded. If you do not wish to be seen in the recording, please avoid sitting in the front three rows of seats in the lecture theatre. Any questions asked will also be included in the recording. The recording will be made available on the Department’s webpage
- Speaker: Dr Prakash Murali - Department of Computer Science and Technology, University of Cambridge
- Wednesday 03 December 2025, 15:05-15:55
- Venue: Lecture Theatre 1, Computer Laboratory, William Gates Building.
- Series: Wednesday Seminars - Department of Computer Science and Technology ; organiser: Ben Karniely.
Wed 26 Nov 15:05: Title to be confirmed
Abstract to be confirmed
Link to join virtually: https://cam-ac-uk.zoom.us/j/89473073451
This talk is being recorded. If you do not wish to be seen in the recording, please avoid sitting in the front three rows of seats in the lecture theatre. Any questions asked will also be included in the recording. The recording will be made available on the Department’s webpage
- Speaker: Dr Ayush Terwari - Department of Engineering, University of Cambridge
- Wednesday 26 November 2025, 15:05-15:55
- Venue: Lecture Theatre 1, Computer Laboratory, William Gates Building.
- Series: Wednesday Seminars - Department of Computer Science and Technology ; organiser: Ben Karniely.
Wed 19 Nov 15:05: Title to be confirmed
Abstract to be confirmed
Link to join virtually: https://cam-ac-uk.zoom.us/j/89473073451
This talk is being recorded. If you do not wish to be seen in the recording, please avoid sitting in the front three rows of seats in the lecture theatre. Any questions asked will also be included in the recording. The recording will be made available on the Department’s webpage
- Speaker: Professor Gabriele Gradoni - Institute for Communication Systems, University of Surrey and visiting academic at the Department of Computer Science and Technology, University of Cambridge
- Wednesday 19 November 2025, 15:05-15:55
- Venue: Lecture Theatre 1, Computer Laboratory, William Gates Building.
- Series: Wednesday Seminars - Department of Computer Science and Technology ; organiser: Ben Karniely.
Wed 12 Nov 15:05: Title to be confirmed
Abstract to be confirmed
Link to join virtually: https://cam-ac-uk.zoom.us/j/89473073451
This talk is being recorded. If you do not wish to be seen in the recording, please avoid sitting in the front three rows of seats in the lecture theatre. Any questions asked will also be included in the recording. The recording will be made available on the Department’s webpage
- Speaker: Dr Daniel Hugenroth - Department of Computer Science and Technology, University of Cambridge
- Wednesday 12 November 2025, 15:05-15:55
- Venue: Lecture Theatre 1, Computer Laboratory, William Gates Building.
- Series: Wednesday Seminars - Department of Computer Science and Technology ; organiser: Ben Karniely.
Wed 29 Oct 15:05: Title to be confirmed
Abstract to be confirmed
Link to join virtually: https://cam-ac-uk.zoom.us/j/89473073451
This talk is being recorded. If you do not wish to be seen in the recording, please avoid sitting in the front three rows of seats in the lecture theatre. Any questions asked will also be included in the recording. The recording will be made available on the Department’s webpage
- Speaker: Dr Greg Lavender - Former Chief Technology Officer of Intel
- Wednesday 29 October 2025, 15:05-15:55
- Venue: Lecture Theatre 1, Computer Laboratory, William Gates Building.
- Series: Wednesday Seminars - Department of Computer Science and Technology ; organiser: Ben Karniely.
Wed 22 Oct 15:00: Wheeler Lecture 2025: Efficiency, Resilience, and Artificial Intelligence
We are delighted to announce that Professor Moshe Vardi will be delivering the 2025 Wheeler Lecture on Wednesday 22nd October hosted by Professor Anuj Dawar.
Guest Speaker: Professor Moshe Vardi https://profiles.rice.edu/faculty/moshe-y-vardi
Abstract: ‘In both computer science and economics, efficiency is a cherished property. The field of algorithms is almost solely focused on their efficiency. The goal of AI research is to increase efficiency by reducing human labor. In economics, the main advantage of the free market is that it promises “economic efficiency”. A major lesson from many recent disasters is that both fields have over-emphasized efficiency and under-emphasized resilience. I argue that resilience is a more important property than efficiency and discuss how the two fields can broaden their focus to make resilience a primary consideration. I will conclude by raising serious questions on the goal of the AI research program.’
Time: 15:00-16:00 The Wheeler Lecture 16:00-17:00 Afternoon Tea
If you plan attend, please sign up on Eventbrite registration by 13/10/25:
- Speaker: Speaker to be confirmed
- Wednesday 22 October 2025, 15:00-17:00
- Venue: Lecture Theatre 1, Computer Laboratory.
- Series: The Wheeler Lectures in Computer Science; organiser: Ben Karniely.
Mon 01 Dec 13:05: Dojo: Title to be confirmed
Abstract to be confirmed.
Some catering will be provided.
- Speaker: Speaker to be confirmed
- Monday 01 December 2025, 13:05-13:55
- Venue: FW26, William Gates Building.
- Series: Technical Talks - Department of Computer Science and Technology ; organiser: Ben Karniely.
Mon 17 Nov 13:05: Gearset: Title to be confirmed
Abstract to be confirmed.
Some catering will be provided.
- Speaker: Speaker to be confirmed
- Monday 17 November 2025, 13:05-13:55
- Venue: FW26, William Gates Building.
- Series: Technical Talks - Department of Computer Science and Technology ; organiser: Ben Karniely.
Mon 03 Nov 13:05: Jane Street: Title to be confirmed
Abstract to be confirmed
Some catering will be provided
- Speaker: Speaker to be confirmed
- Monday 03 November 2025, 13:05-13:55
- Venue: FW26, William Gates Building.
- Series: Technical Talks - Department of Computer Science and Technology ; organiser: Ben Karniely.
Mon 27 Oct 13:05: GSA Capital: Technology at GSA
Overview: In this presentation, we will provide an overview of GSA and what a systematic quantitative fund is. We will outline the role of our Technology teams, including the divisions, technology stacks and how it supports researchers in the development and deployment of trading strategies. Finally, one of our software developers will share a recent project that involved compressing exchange trade and order data into a binary format, reducing the replay time by approximately ten-fold, and enabling rapid generation of intraday equity trading signals.
Speaker: Junhui Yang joined GSA in 2023 as a software developer supporting the firm’s founder. He works on building and maintaining data pipelines, back testing frameworks, monitoring systems and other critical tools essential to the success of a quantitative fund. Prior to GSA , he read Mathematics and Computer Science at Merton College, University of Oxford and spent a brief period at Meta. Outside of work, he enjoys playing the violin and piano, and is a tea enthusiast.
Audience: We welcome students from all academic levels in Computer Science, Engineering, Mathematics, Statistics and other closely related disciplines to attend. The event will start with a short presentation followed by informal chats with representatives from GSA over food and drinks. Please sign up via our event listings at https://www.gsacapital.com/join-us so we can gauge numbers for catering purposes.
About GSA : GSA Capital is a multi-award winning investment manager that has been delivering exceptional results for over 15 years. With offices in London and New York, GSA Capital combines a world-class proprietary platform with innovative thinking to develop and deploy systematic and process-driven investment strategies across all asset classes, geographies and timescales. For more information, visit our website at www.gsacapital.com
Some catering will be provided
- Speaker: Junhui Yang
- Monday 27 October 2025, 13:05-13:55
- Venue: FW26, William Gates Building.
- Series: Technical Talks - Department of Computer Science and Technology ; organiser: Ben Karniely.
Mon 20 Oct 13:05: Bloomberg: Observability in Action: Designing Effective Dashboards
Dashboards are a critical part of your observability stack. Looking at your dashboards—are they actually helping you understand the system? When well-designed, they surface intuitive information and help quickly diagnose outages. But poorly designed dashboards can do more harm than good: they generate noise instead of insights, contribute to alert fatigue, and bury real issues under a flood of useless data. In this tech talk, we’ll explore common design pitfalls through a fun fast food shop scenario that mimics real-world systems. We’ll look at examples of dashboards that mislead and ones that empower, and discuss how to create dashboards that deliver actionable insights, reduce noise, and support fast, confident decision-making. You’ll learn how to identify key signals, tailor visualizations to specific roles and needs, and align your observability strategy with real-world goals. Whether you’re a student, early-career engineer, or just curious about system reliability, you’ll walk away with practical tips for building dashboards that drive clarity, confidence, and impact.
Registration Link: https://bloomberg.avature.net/su/2ebe52ef137ab02f
Some catering will be provided
- Speaker: Speaker to be confirmed
- Monday 20 October 2025, 13:05-13:55
- Venue: FW26, William Gates Building.
- Series: Technical Talks - Department of Computer Science and Technology ; organiser: Ben Karniely.
Fri 28 Nov 16:30: What does the high heritability of psychological traits mean for psychologists and educators?
Behavioural genetics has a hundred-year history of pointing to the high heritability of psychological traits relevant to people’s everyday lives, and outcomes such as educational attainment linked to wellbeing and professional success. Recently, the availability of polygenic indices, derived from large-scale genomic studies, has brought behavioural genetics to the individual level, generating individualised predictions of genetic potential. This advance brings into focus the relevance of genomic information for both personal and policy decisions. The drawbacks of behavioural genetics are twofold: it tells us only about current population outcomes, not how they could be different under different circumstances; and it is relatively silent on the developmental mechanisms that deliver heritable outcomes. In this talk, I will use a computational framework to focus on developmental mechanisms. I will show how simulations of interventions to alter developmental outcomes for whole populations, which have been designed to show differing levels of trait heritability, can produce surprising results. These results point to the policy relevance of the notion of heritability itself.
Host: Prof Dénes Szücs (ds377@cam.ac.uk)
- Speaker: Prof Michael Thomas, Birkbeck, University of London
- Friday 28 November 2025, 16:30-18:00
- Venue: Ground Floor Lecture Theatre, Department of Psychology.
- Series: Zangwill Club; organiser: Psychology Reception.
Tue 28 Apr 16:00: Title to be confirmed
Abstract not available
- Speaker: Professor Thomas Nikolaus (Universität Münster)
- Tuesday 28 April 2026, 16:00-17:00
- Venue: MR2, CMS.
- Series: Mordell Lectures; organiser: HoD Secretary, DPMMS.
Tue 28 Apr 16:00: Title to be confirmed
Abstract not available
- Speaker: Professor Thomas Nikolaus (Universität Münster)
- Tuesday 28 April 2026, 16:00-17:00
- Venue: MR2, CMS.
- Series: Mordell Lectures; organiser: HoD Secretary, DPMMS.
Mon 13 Oct 13:05: Perplexity AI: Inference at Perplexity AI
Abstract: Perplexity is a search and answer engine which leverages LLMs to provide high-quality citation-backed answers. The AI Inference team within the company is responsible for serving the models behind the product, ranging from single-GPU embedding models to multi-node sparse Mixture-of-Experts language models. This talk provides more insight into the in-house runtime behind inference at Perplexity, with a particular focus on efficiently serving some of the largest available open-source models.
Biography:Nandor Licker is an AI Inference Engineer at Perplexity, focusing on LLM runtime implementation and GPU performance optimization.
Some catering will be provided
- Speaker: Speaker to be confirmed
- Monday 13 October 2025, 13:05-13:55
- Venue: FW26, William Gates Building.
- Series: Technical Talks - Department of Computer Science and Technology ; organiser: Ben Karniely.
Mon 13 Oct 13:00: Predicting wind extremes in a warming climate: from general circulation to storm-resolving models via improved turbulence representation
A wave of unprecedented extreme weather events, breaking records worldwide, has raised urgent questions about the ability of current weather and climate models to anticipate the emerging impacts of climate change on human life and infrastructure. Among these, extreme wind speeds and gusts, often associated with midlatitude cyclones and low-level jets, pose a growing threat to critical sectors of society. In this talk, I will first present projections of near-surface extreme winds over the midlatitudes of both hemispheres under an idealized warming scenario, based on CMIP -class models. I will then illustrate how global kilometer-scale simulations may provide new insight into how the structure and intensity of North Atlantic midlatitude cyclones respond to climate warming. Finally, I will discuss results from a set of experiments with the GFDL -AM4 model that incorporate improved turbulence representation via the CLUBB scheme. These highlight the role of prognosed momentum fluxes in better capturing low-level jet dynamics and improving the simulation of the diurnal precipitation cycle. Together, these studies demonstrate the importance of refined physics and high-resolution modeling for advancing our understanding and prediction of wind extremes in a warming climate.
- Speaker: Emanuele Silvio Gentile, University of Reading
- Monday 13 October 2025, 13:00-14:00
- Venue: MR3, CMS.
- Series: Quantitative Climate and Environmental Science Seminars; organiser: Bethan Wynne-Cattanach.
Fri 21 Nov 12:00: Bayesian Brains Without Probabilities
Over the past few decades, waves of complex probabilistic explanations have swept through cognitive science, explaining behaviour as tuned to environmental statistics in domains from intuitive physics and causal learning, to perception, motor control and language. Yet people produce stunningly incorrect answers in response to even the simplest questions about probabilities. How can a supposedly rational brain paradoxically reason so poorly with probabilities? Perhaps our minds do not represent or calculate probabilities at all and are, indeed, poorly adapted to do so. Instead, the brain could be approximating Bayesian inference through sampling: drawing samples from its distribution of likely hypotheses over time. Only with infinite samples does a Bayesian sampler conform to the laws of probability, and in this talk I show how using a finite number of samples systematically generates classic probabilistic reasoning errors in individuals, and how an extended model explains estimates, choices, response times, and confidence judgments in a variety of tasks.
Host: Dr Deborah Talmi (dt492@cam.ac.uk)
- Speaker: Prof Adam Sanborn, University of Warwick
- Friday 21 November 2025, 12:00-13:30
- Venue: Ground Floor Lecture Theatre, Department of Psychology.
- Series: Zangwill Club; organiser: Psychology Reception.
Fri 14 Nov 16:30: Naive Wisdom: Behavioral Evidence from Newborn Chicks
For many years, the scientific community neglected or even denied the existence of anything resembling a mind in newborn animals, whether human or non-human. However, since the latter half of the twentieth century, a series of seminal studies has revealed a dramatically different scenario. Today it is well established that newborn animals enter the world equipped with a rich repertoire of innate predispositions and skills that facilitate learning and enable them to successfully navigate their social and physical environments. In this talk, I will present an overview of research highlighting key aspects of the newborn mind, with particular focus on behavioural methodologies and findings in which I have been directly involved—chiefly investigating the newborn domestic chick, as well as extending some findings to human infants.
Host: Prof Nicky Clayton (nsc22@cam.ac.uk)
- Speaker: Prof Lucia Regolin, University of Padua
- Friday 14 November 2025, 16:30-18:00
- Venue: Ground Floor Lecture Theatre, Department of Psychology.
- Series: Zangwill Club; organiser: Psychology Reception.