BEGIN:VCALENDAR VERSION:2.0 PRODID:-//Cambridge AI4S//Community calendar//EN CALSCALE:GREGORIAN METHOD:PUBLISH X-WR-CALNAME:Cambridge AI4S X-WR-CALDESC:AI-for-science events in Cambridge\, compiled daily by Cambr idge AI4S X-WR-TIMEZONE:Europe/London REFRESH-INTERVAL;VALUE=DURATION:PT12H X-PUBLISHED-TTL:PT12H BEGIN:VEVENT UID:83adeefcec02@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261027T130000Z DTEND:20261027T140000Z SUMMARY:Learning task structures in the hippocampus DESCRIPTION:Xinyu Zhao\, Tsinghua University\n\nAnimals can solve the sam e problem using distinct strategies\, from memorizing specific sensory d etails to making inferences based on abstract task structure. The mammal ian hippocampus has traditionally been viewed as an episodic memory syst em that encodes the rich sensory details of individual experiences.\n\nS eries: Computational Neuroscience\n\nhttps://talks.cam.ac.uk/talk/index/ 273930/ LOCATION:CBL Seminar Room\, Engineering Department\, 4th floor Baker buil ding URL:https://talks.cam.ac.uk/talk/index/273930/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:2c366b0f80c7@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261027T130000Z DTEND:20261027T140000Z SUMMARY:Towards a surface-mass balance emulator for Greenland\, the Europ ean Alps and the Himalayas: lessons learned and plans DESCRIPTION:Dr. Marco Tedesco\; Lamont Research Professor\, Marine and Po lar Geophysics\, Lamont-Doherty Earth Observatory (LDEO)\, Columbia Clim ate School Adjunct Scientist\, NASA Goddard Institute for Space Studies …\n\nRegional climate models (RCMs) such as the Modèle Atmosphérique Régional (MAR) compute surface mass balance (SMB) with physical consis tency\, but their computational cost and forcing requirements limit larg e ensembles\, long reconstructions and systematic attribution.\n\nSeries : Scott Polar Research Institute - Polar Physical Sciences Seminar\n\nht tps://talks.cam.ac.uk/talk/index/273636/ LOCATION:Scott Polar Research Institute\, main lecture theatre URL:https://talks.cam.ac.uk/talk/index/273636/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:a444d86a9d73@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261027T160000Z DTEND:20261027T170000Z SUMMARY:Beyond Academia - Financial Times\, Chief Data Officer DESCRIPTION:Kate Sargent\, Chief Data Officer\, Financial Times\n\nThe Be yond Academia: AI & Data Science Applications seminar series at the Univ ersity of Cambridge explores the transformative role of data and artific ial intelligence in the world. Designed for postgraduate researchers\, e arly-career academics\, and students\, the series bridges the gap betwee n high-level research and real-world deployment.\n\nSeries: Cambridge Ce ntre for Data-Driven Discovery (C2D3)\n\nhttps://talks.cam.ac.uk/talk/in dex/272811/ LOCATION:The Glasshouse @InnovateCambridge\, 100 Hills Road\, Cambridge\, CB2 1LQ URL:https://talks.cam.ac.uk/talk/index/272811/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:195ef5a746f3@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261028T114500Z DTEND:20261028T130000Z SUMMARY:Cambridge AI in Medicine Seminar - October 2026 DESCRIPTION:Jian Chen and Kai Yi\n\nSign up on Eventbrite: https://medai- oct2026.eventbrite.co.uk Join us for the Cambridge AI in Medicine Semina r Series\, hosted by the Cancer Research UK Cambridge Centre and the Dep artment of Radiology at Addenbrooke’s. This series brings together lea ding experts to explore cutting-edge AI applications in healthcare – f rom disease diagnosis to drug discovery.\n\nSeries: Cambridge AI in Medi cine Seminar Series\n\nhttps://talks.cam.ac.uk/talk/index/273033/ LOCATION:Jeffrey Cheah Biomedical Centre (Main Lecture Theatre)\, Univers ity of Cambridge URL:https://talks.cam.ac.uk/talk/index/273033/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:366ab87a9ad5@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261028T133000Z DTEND:20261028T143000Z SUMMARY:Bigger Picture Talk: Can we make cameras as thin as a hair? Build ing tiny imaging devices with fibres\, metasurfaces and AI DESCRIPTION:George Gordon - Department of Chemical Engineering and Biotec hnology\, University of Cambridge\n\nConventional cameras and microscope s are built by combining lenses\, detectors and illumination systems\, w ith better performance often requiring many large optical components. Bu t what happens when the imaging system needs to be thinner than a millim etre\, or even as narrow as a human hair?\n\nSeries: Chemical Engineerin g and Biotechnology\n\nhttps://talks.cam.ac.uk/talk/index/273260/ LOCATION:Lecture Theatre 2\, Department of Chemical Engineering and Biote chnology\, West Cambridge Site URL:https://talks.cam.ac.uk/talk/index/273260/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:7cf811e590e4@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261028T150500Z DTEND:20261028T155500Z SUMMARY:Ed Catmull: computers and the movies DESCRIPTION:Dr Ed Catmull - Co-founder of Pixar\n\nEd Catmull is a comput er scientist\, co-founder of Pixar Animation Studios\, and one of the fo undational figures in computer graphics. His research helped establish s ome of the core techniques of modern 3D computer graphics\, and he subse quently led Pixar and Walt Disney Animation Studios through several deca des of rapid technological and creative change.\n\nSeries: Wednesday Sem inars – Department of Computer Science and Technology\n\nhttps://talks .cam.ac.uk/talk/index/272523/ LOCATION:Lecture Theatre 1\, Computer Laboratory\, William Gates Building URL:https://talks.cam.ac.uk/talk/index/272523/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:a0b046cb1624@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261029T150000Z DTEND:20261029T160000Z SUMMARY:Curvature-Aware Optimization for High-Accuracy Physics-Informed N eural Networks DESCRIPTION:Khemraj Shukla (Brown University)\n\nEfficient and robust opt imization is essential for training neural networks\, enabling scientifi c machine learning models to converge rapidly to high accuracy while fai thfully capturing complex physical behavior governed by partial differen tial equations (PDEs).\n\nSeries: Applied and Computational Analysis\n\n https://talks.cam.ac.uk/talk/index/272768/ LOCATION:Centre for Mathematical Sciences\, MR14 URL:https://talks.cam.ac.uk/talk/index/272768/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:1894236f6609@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261029T150000Z DTEND:20261029T160000Z SUMMARY:Toward Reliable Deployment of LLM Agents: From Uncertainty Quanti fication to Progress Advantage DESCRIPTION:Sharon Li (University of Wisconsin-Madison)\n\nLLM agents are increasingly deployed on long-horizon tasks with tool use\, irreversibl e actions\, and unpredictable feedback. Yet we have few principled ways to tell\, mid-episode\, whether an agent is on track or quietly failing. Most uncertainty quantification (UQ) research still centers on single-t urn QA\, a poor match for interactive agents.\n\nSeries: Language Techno logy Lab Seminars\n\nhttps://talks.cam.ac.uk/talk/index/272902/ LOCATION:Online URL:https://talks.cam.ac.uk/talk/index/272902/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:053da5119127@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261029T150000Z DTEND:20261029T160000Z SUMMARY:Digital Twins: From Sensor Design to Real-Time Data Visualisation DESCRIPTION:Dr. Justas Brazauskas\, University of Cambridge\n\nAbstract: Digital twins combine real-world data with digital representations of ph ysical systems\, but building useful digital twins involves more than si mply collecting and displaying data.\n\nSeries: Computer Laboratory Syst ems Research Group Seminar\n\nhttps://talks.cam.ac.uk/talk/index/273183/ LOCATION:LT2 URL:https://talks.cam.ac.uk/talk/index/273183/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:ff04d91bfc7c@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261029T170000Z DTEND:20261029T180000Z SUMMARY:Panel session: a conversation on academia in the age of AI DESCRIPTION:The rapid impact and integration of AI into the research envi ronment naturally raises a number of questions and concerns\, particular ly among PhD students and early career researchers in mathematics and th eoretical physics.\n\nSeries: Isaac Newton Institute Seminar Series\n\nh ttps://talks.cam.ac.uk/talk/index/273619/ LOCATION:Seminar Room 1\, Newton Institute URL:https://talks.cam.ac.uk/talk/index/273619/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:aaa61947502b@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261030T120000Z DTEND:20261030T130000Z SUMMARY:Title to be confirmed DESCRIPTION:Rachel Lawrence (Microsoft Research Cambridge)\n\nSeries: NLI P Seminar Series\n\nhttps://talks.cam.ac.uk/talk/index/273556/ URL:https://talks.cam.ac.uk/talk/index/273556/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:8e9f1bab7c82@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261030T123000Z DTEND:20261030T133000Z SUMMARY:AI-driven materials discovery DESCRIPTION:Felix Hanke (CuspAI)\n\nSeries: AMET seminars\n\nhttps://talk s.cam.ac.uk/talk/index/273058/ LOCATION:JJ Thomson Seminar Room\, Maxwell Centre URL:https://talks.cam.ac.uk/talk/index/273058/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:5aeaad12ea28@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261030T130000Z DTEND:20261030T140000Z SUMMARY:Towards Pixel-Level Earth Observation Understanding with Multimod al Large Language Models DESCRIPTION:Yujie Zan\, University of Cambridge\n\nAbstract: Multimodal l arge language models (MLLMs) are changing how we understand Earth observ ation imagery\, making it possible to interact with satellite images thr ough natural language. However\, these models still struggle with tasks that require precise spatial understanding\, such as identifying individ ual objects\, mapping land cover\, and detecting changes over time.\n\nS eries: Energy and Environment Group\, Department of CST\n\nhttps://talks .cam.ac.uk/talk/index/272633/ LOCATION:Room GS15 at the William Gates Building and on Zoom URL:https://talks.cam.ac.uk/talk/index/272633/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:165baa1766c5@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261030T140000Z DTEND:20261030T150000Z SUMMARY:Learning Memory and Material Dependent Constitutive Laws DESCRIPTION:Dr George Stepaniants\, DAMTP\n\nThe simulation of multiscale viscoelastic materials poses a significant challenge in computational m aterials science\, requiring expensive numerical solvers that can resolv e dynamics of material deformations at the microscopic scale. The theory of homogenization offers an alternative approach to modeling\, by local ly averaging the strains and stresses of multiscale materials.\n\nSeries : Engineering - Mechanics and Materials Seminar Series\n\nhttps://talks. cam.ac.uk/talk/index/273296/ LOCATION:Oatley 1 Meeting Room\, CUED URL:https://talks.cam.ac.uk/talk/index/273296/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:311599d4d4c7@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261030T153000Z DTEND:20261030T163000Z SUMMARY:Title to be confirmed DESCRIPTION:Gašper Beguš (University of California\, Berkeley)\n\nSerie s: NLIP Seminar Series\n\nhttps://talks.cam.ac.uk/talk/index/273587/ URL:https://talks.cam.ac.uk/talk/index/273587/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:74b62f3aa5f6@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261103T093000Z DTEND:20261104T130000Z SUMMARY:Training Workshop: Uncertainty Estimation in AI DESCRIPTION:Neural networks have had success in science and engineering p redictive tasks such as computer vision\, biology\, medicine\, and weath er. To safely use (make decisions based on) the predictions of neural ne tworks\, the predictions need to have reliable uncertainty estimates. Ye t\, despite their importance\, reliable predictive uncertainty estimates are often overlooked in practice.\n\nhttps://science.ai.cam.ac.uk/event s/training-workshop-uncertainty-estimation-in-ai URL:https://science.ai.cam.ac.uk/events/training-workshop-uncertainty-est imation-in-ai CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:9e3fa674ccb9@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261103T140000Z DTEND:20261103T150000Z SUMMARY:The Invisible Hand of Stability DESCRIPTION:Geelon So\, ETH Zürich\n\nThe division of labor enables us t o build complex human and machine systems. Today\, decentralized multi-a gent systems with learned behavior are becoming more prevalent. At the s ame time\, decentralization can come with a costly price of anarchy—ag ents learn to behave in ways that leave everyone worse off. Can we under stand and guard against collectively irrational behavior?\n\nSeries: Mac hine learning theory\n\nhttps://talks.cam.ac.uk/talk/index/273900/ LOCATION:MR19 (Potter Room\, Pavilion B)\, CMS URL:https://talks.cam.ac.uk/talk/index/273900/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:6502816f8990@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261103T140000Z DTEND:20261103T150000Z SUMMARY:"Operando X-ray correlative imaging and processing of 3D structur es inside lithium batteries" DESCRIPTION:Professor Chun Ann Huang - RSC Marlow Prizewinner 2025 - Impe rial College London\n\nIt is conventionally difficult to measure lithium ion concentrations inside commercially standard battery cell configurat ions using X-rays because the weak signals of lithium cannot be distingu ished from the strong signals of other heavy transition metal elements t hat are present in batteries.\n\nSeries: Materials Chemistry Research In terest Group\n\nhttps://talks.cam.ac.uk/talk/index/273598/ LOCATION:Dept of Chemistry\, Wolfson Lecture Theatre URL:https://talks.cam.ac.uk/talk/index/273598/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:40eb9062af9c@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261103T160000Z DTEND:20261103T170000Z SUMMARY:Out-of-Distribution Generalisation\, Predictive Uncertainty\, and Continual Learning DESCRIPTION:Mark van der Wilk\, University of Oxford\n\nHow can we create a learning algorithm that generalises “out of distribution”\, i.e. beyond the training data? What can we expect about errors when a model g eneralises OoD? And what should happen when an error is made? In this ta lk\, we will discuss how OccamLab’s approach to these questions.\n\nSe ries: Computational Neuroscience\n\nhttps://talks.cam.ac.uk/talk/index/2 73882/ LOCATION:CBL Seminar Room\, Engineering Department\, 4th floor Baker buil ding URL:https://talks.cam.ac.uk/talk/index/273882/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:1158a05bc846@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261104T143000Z DTEND:20261104T153000Z SUMMARY:Data-driven acceleration of materials discovery and optimization for energy applications DESCRIPTION:Professor Silvana Botti\, Rohr University Bochum\n\nThe searc h for better materials for energy conversion and storage is limited less by our understanding of the underlying physics than by the sheer size o f the chemical and structural space to be explored.\n\nSeries: Theory - Chemistry Research Interest Group\n\nhttps://talks.cam.ac.uk/talk/index/ 246972/ LOCATION:Unilever Lecture Theatre\, Yusuf Hamied Department of Chemistry URL:https://talks.cam.ac.uk/talk/index/246972/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:b1a05f648f82@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261105T160000Z DTEND:20261105T170000Z SUMMARY:Perspectives on Cross-Lingual Consistency in LLMs for Medical Que stions DESCRIPTION:Minh Duc Bui (Johannes Gutenberg University Mainz)\n\nShould multilingual LLMs answer medical questions the same way in every languag e\, or adapt to cultural cues? In this talk\, we discuss these two persp ectives. Multilingual medical benchmarks typically treat cross-lingual v ariation as error\, while cultural adaptation research argues that appro priate answers can legitimately differ.\n\nSeries: Language Technology L ab Seminars\n\nhttps://talks.cam.ac.uk/talk/index/273238/ LOCATION:Online URL:https://talks.cam.ac.uk/talk/index/273238/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:8293d74747a0@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261106T120000Z DTEND:20261106T130000Z SUMMARY:Title to be confirmed DESCRIPTION:Lisa Alazraki (Imperial College)\n\nSeries: NLIP Seminar Seri es\n\nhttps://talks.cam.ac.uk/talk/index/273264/ URL:https://talks.cam.ac.uk/talk/index/273264/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:b2ebcb4a596e@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261109T130500Z DTEND:20261109T135500Z SUMMARY:Microsoft: Building Future AI Infrastructure: Inside Microsoft Sp ecialized Clouds R&D DESCRIPTION:Ryan Bailey and Rich Hewitt\n\nJoin Ryan Bailey and Rich Hewi tt\, Principal Group Engineering Manager at Microsoft\, for a look insid e Microsoft Specialised Clouds R&D and how the team is tackling the engi neering challenges behind AI agents and infrastructure.\n\nSeries: Techn ical Talks - Department of Computer Science and Technology\n\nhttps://ta lks.cam.ac.uk/talk/index/272548/ LOCATION:FW26\, William Gates Building URL:https://talks.cam.ac.uk/talk/index/272548/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:6865883df58f@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261110T160000Z DTEND:20261110T170000Z SUMMARY:Beyond Academia - BBC\, Head of Applied AI DESCRIPTION:Nikolay Burlutskiy\, Head of Applied AI\, BBC\n\nThe Beyond A cademia: AI & Data Science Applications seminar series at the University of Cambridge explores the transformative role of data and artificial in telligence in the world. Designed for postgraduate researchers\, early-c areer academics\, and students\, the series bridges the gap between high -level research and real-world deployment.\n\nSeries: Cambridge Centre f or Data-Driven Discovery (C2D3)\n\nhttps://talks.cam.ac.uk/talk/index/27 2812/ LOCATION:Room E\, West Hub\, JJ Thomson Ave\, Cambridge CB3 0US URL:https://talks.cam.ac.uk/talk/index/272812/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:3ed95e8a238e@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261111T093000Z DTEND:20261111T133000Z SUMMARY:Training Workshop: Hands on Agentic AI for Science DESCRIPTION:This half-day workshop introduces agentic AI through a mix of short presentations\, live demonstrations\, and hands-on notebook exerc ises. Participants will explore how LLM-powered agents can be designed t o reason over tasks\, use tools\, call external APIs\, work with structu red prompts\, and carry out multi-step workflows.\n\nhttps://science.ai. cam.ac.uk/events/training-workshop-hands-on-agentic-ai-for-science URL:https://science.ai.cam.ac.uk/events/training-workshop-hands-on-agenti c-ai-for-science CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:4260bb83456d@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261112T100000Z DTEND:20261126T120000Z SUMMARY:AI Clinic slots - online DESCRIPTION:Have you thought about using AI in your research but aren’t sure how to get started? Or are you already using AI and have run into challenges with implementation? Meet the Accelerate Programme’s team o f AI experts to find the support you need.\n\nhttps://science.ai.cam.ac. uk/events/ai-clinic-slots-online-35 URL:https://science.ai.cam.ac.uk/events/ai-clinic-slots-online-35 CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:005e4070be2e@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261112T140000Z DTEND:20261112T150000Z SUMMARY:Title to be confirmed DESCRIPTION:Jindřich Libovický (Charles University)\n\nSeries: Language Technology Lab Seminars\n\nhttps://talks.cam.ac.uk/talk/index/273239/ LOCATION:Online URL:https://talks.cam.ac.uk/talk/index/273239/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:f4746aee4d32@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261112T150000Z DTEND:20261112T160000Z SUMMARY:Expander Datacenters: From First Principles to the Cloud DESCRIPTION:Prof. Michael Schapira\, Hebrew University of Jerusalem\n\nAb stract: AWS recently adopted an expander-based network topology as the d efault design for its new datacenters\, bringing an idea rooted in graph theory into operation at unprecedented scale. In this talk\, we trace t he research journey that led to this development.\n\nSeries: Computer La boratory Systems Research Group Seminar\n\nhttps://talks.cam.ac.uk/talk/ index/273185/ LOCATION:FN05 URL:https://talks.cam.ac.uk/talk/index/273185/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:afd15877b75d@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261113T120000Z DTEND:20261113T130000Z SUMMARY:Title to be confirmed DESCRIPTION:Anne-Marie Lutgen (University of Luxembourg)\n\nSeries: NLIP Seminar Series\n\nhttps://talks.cam.ac.uk/talk/index/273205/ URL:https://talks.cam.ac.uk/talk/index/273205/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:6cae2627c226@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261113T160000Z DTEND:20261113T170000Z SUMMARY:Examining fast hydrological responses to anthropogenic forcing ag ents using machine-learning climate emulators DESCRIPTION:Bill Collins\, University of East Anglia\n\nThe response of t he climate system to increased greenhouse gases is governed by a combina tion of fast responses and slow feedbacks. Slow feedbacks are typically activated in response to changes in ocean temperatures on decadal timesc ales and manifest as changes in climatic state with no recent historical analogues.\n\nSeries: Fluid Mechanics (DAMTP)\n\nhttps://talks.cam.ac.u k/talk/index/273504/ LOCATION:MR2 URL:https://talks.cam.ac.uk/talk/index/273504/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:177a4795b309@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261117T140000Z DTEND:20261117T150000Z SUMMARY:MRC Biostatistics Unit seminar DESCRIPTION:Pawel Morzywolek\, University of Copenhagen\n\nInference on V ariable Importance for Treatment Effect Heterogeneity: Shapley Values an d Beyond We provide an inferential framework to assess variable importan ce for heterogeneous treatment effects. This assessment is especially us eful in high-risk domains such as medicine\, where decision makers hesit ate to rely on black-box treatment recommendation algorithms.\n\nSeries: MRC Biostatistics Unit Seminars\n\nhttps://talks.cam.ac.uk/talk/index/2 73523/ LOCATION:on-line seminar URL:https://talks.cam.ac.uk/talk/index/273523/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:b506aefc5693@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261118T143000Z DTEND:20261118T153000Z SUMMARY:Machine Learning for Materials Discovery: From Screening to Advan ced Simulations DESCRIPTION:Professor Janine George\, BAM Berlin\n\nThe development of ad vanced materials relies on a continuous feedback loop between computatio nal prediction\, experimental synthesis\, and characterization. Computat ional methods support this process in many ways.\n\nSeries: Theory - Che mistry Research Interest Group\n\nhttps://talks.cam.ac.uk/talk/index/272 153/ LOCATION:Unilever Lecture Theatre\, Yusuf Hamied Department of Chemistry URL:https://talks.cam.ac.uk/talk/index/272153/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:981084f8a0fe@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261118T150500Z DTEND:20261118T155500Z SUMMARY:Synthesizing Correct Hardware Without Hallucination DESCRIPTION:Dr Satnam Singh - Harmonic. Visiting academic at the Departme nt of Computer Science and Technology\, University of Cambridge\n\nHardw are can be formally verified by writing theorems about its desired corre ct behaviour and then proving these theorems correct against a formal ma thematical representation of the circuit. Until now this has been a chal lenging task requiring rare deep expertise both in using interactive the orem provers and in hardware design and verification.\n\nSeries: Wednesd ay Seminars – Department of Computer Science and Technology\n\nhttps:/ /talks.cam.ac.uk/talk/index/272522/ LOCATION:Lecture Theatre 1\, Computer Laboratory\, William Gates Building URL:https://talks.cam.ac.uk/talk/index/272522/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:6047ae39a467@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261119T110000Z DTEND:20261119T120000Z SUMMARY:How Do Language Models Reason? Foundations\, Limitations and Neur o-Symbolic Approaches DESCRIPTION:André Freitas (Idiap Research Institute & University of Manc hester)\n\nHow do language models reason\, and what would it take to bui ld AI systems capable of sustained\, complex reasoning? This lecture dev elops the foundations for answering these questions\, connecting what we know about the internal workings of language models with methods for ev aluating and extending their capabilities.\n\nSeries: Language Technolog y Lab Seminars\n\nhttps://talks.cam.ac.uk/talk/index/273552/ LOCATION:GR04 (English Faculty Building\, 9 West Road\, Sidgwick Site) an d online ( URL:https://talks.cam.ac.uk/talk/index/273552/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:be933dc5caee@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261119T140000Z DTEND:20261119T150000Z SUMMARY:Lewis Lecture 1: Application of NMR Crystallography to Study Diso rder in Mixed-Metal Materials DESCRIPTION:Professor Sharon Ashbrook\, University of St Andrews\n\nInorg anic materials play vital roles in our everyday lives\; understanding th e atomic-scale structure of these solids\, and how that affects their ph ysical and chemical properties\, is key to the future design of function al materials. Mixed-metal materials offer the opportunity to vary or con trol properties through changes in the composition and the distribution of atoms.\n\nSeries: Materials Chemistry Research Interest Group\n\nhttp s://talks.cam.ac.uk/talk/index/273740/ LOCATION:Dept of Chemistry\, Wolfson Lecture Theatre URL:https://talks.cam.ac.uk/talk/index/273740/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:88845fd6b541@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261120T120000Z DTEND:20261120T130000Z SUMMARY:Aligning LLMs to the Human Brain DESCRIPTION:Mariya Toneva (Max Planck Institute for Software Systems)\n\n Series: NLIP Seminar Series\n\nhttps://talks.cam.ac.uk/talk/index/273206 / URL:https://talks.cam.ac.uk/talk/index/273206/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:179d56d01a4d@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261123T173000Z DTEND:20261123T183000Z SUMMARY:Do Children and Other Animals Think Like Scientists? DESCRIPTION:Professor Marta Halina is a Senior Research Fellow at the Lev erhulme Centre for the Future of Intelligence\n\nFor decades\, psycholog ists have characterised young children as little scientists\, a comparis on made famous by the 1999 book The Scientist in the Crib. A core claim of this perspective is that children (and other animals such as chimpanz ees\, dogs\, and birds) use abstract theories to navigate the world\, mu ch like researchers discovering and applying general laws.\n\nSeries: Ca mbridge Philosophical Society\n\nhttps://talks.cam.ac.uk/talk/index/2737 94/ LOCATION:Bristol-Myers Squibb Lecture Theatre\, Department of Chemistry URL:https://talks.cam.ac.uk/talk/index/273794/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:7c566f972cd8@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261124T100000Z DTEND:20261124T180000Z SUMMARY:AI for Science Summit 2026 DESCRIPTION:AI for Science Summit 2026\n\nhttps://science.ai.cam.ac.uk/ev ents/ai-for-science-summit-2026 URL:https://science.ai.cam.ac.uk/events/ai-for-science-summit-2026 CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:1ab6dbd24c04@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261124T140000Z DTEND:20261124T150000Z SUMMARY:Understanding and Protecting Interdependent Privacy in LLMs DESCRIPTION:Xiao Zhan\, University of Cambridge\n\nInterdependent privacy arises when one person's use of technology affects the privacy of other s. In this talk\, I will present an empirical study exploring how and wh y people disclose others' information in everyday interactions with LLMs \, as well as how they perceive and navigate the associated privacy conc erns.\n\nSeries: Computer Laboratory Security Seminar\n\nhttps://talks.c am.ac.uk/talk/index/273731/ LOCATION:Webinar & Lecture Theatre 2\, Computer Laboratory\, William Gate s Building URL:https://talks.cam.ac.uk/talk/index/273731/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:9e433bc0a251@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261126T110000Z DTEND:20261126T120000Z SUMMARY:Title to be confirmed DESCRIPTION:Bálint Gyevnár (CMU)\n\nSeries: Language Technology Lab Sem inars\n\nhttps://talks.cam.ac.uk/talk/index/272978/ LOCATION:GR04 (English Faculty Building\, 9 West Road\, Sidgwick Site) an d online ( URL:https://talks.cam.ac.uk/talk/index/272978/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:da76aaf4d49b@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261130T123000Z DTEND:20261130T133000Z SUMMARY:The threat of AI-agent responding across online survey platforms DESCRIPTION:Tanay Katiyar\, MRC CBU\n\nTBC\n\nSeries: CBU Monday Methods Meeting\n\nhttps://talks.cam.ac.uk/talk/index/273257/ LOCATION:MRC-CBU\, 15 Chaucer Road\, Cambridge URL:https://talks.cam.ac.uk/talk/index/273257/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:a223f7cb12aa@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261203T140000Z DTEND:20261203T150000Z SUMMARY:Title to be confirmed DESCRIPTION:Simon Ostermann (German Research Centre for Artificial Intell igence)\n\nSeries: Language Technology Lab Seminars\n\nhttps://talks.cam .ac.uk/talk/index/272979/ LOCATION:Online URL:https://talks.cam.ac.uk/talk/index/272979/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:f0ec514c878f@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261204T163000Z DTEND:20261204T180000Z SUMMARY:The Psychology of AI: How Artificial Intelligence is Changing Ind ividual\, Group\, and Organisational Behaviour DESCRIPTION:Jessica Charter\, EY\n\nArtificial Intelligence is transformi ng more than technology. It is reshaping how people think\, collaborate\ , and lead. As AI becomes embedded in everyday work\, its influence exte nds beyond productivity to fundamentally changing human behaviour across individuals\, teams\, and organisations. This session explores three ke y questions.\n\nSeries: Zangwill Club\n\nhttps://talks.cam.ac.uk/talk/in dex/272971/ LOCATION:Ground Floor Lecture Theatre\, Department of Psychology URL:https://talks.cam.ac.uk/talk/index/272971/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:deadline-2026-12-08-1@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART;VALUE=DATE:20261208 DTEND;VALUE=DATE:20261209 SUMMARY:Deadline: University Postgraduate Funding Competition\, first rou nd DESCRIPTION:Most PhDs\, the MPhil MLMI\, the BBSRC DTP and NanoDTC's firs t review. This round reaches the Cambridge Trust\, Harding and departmen tal studentships. CATEGORIES:Deadline END:VEVENT BEGIN:VEVENT UID:64b61e0c7f85@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART:20261210T114500Z DTEND:20261210T130000Z SUMMARY:Cambridge AI in Medicine Seminar - December 2026 DESCRIPTION:Shivani Sinha\n\nSign up on Eventbrite: details coming soon J oin us for the Cambridge AI in Medicine Seminar Series\, hosted by the C ancer Research UK Cambridge Centre and the Department of Radiology at Ad denbrooke’s. This series brings together leading experts to explore cu tting-edge AI applications in healthcare – from disease diagnosis to d rug discovery.\n\nSeries: Cambridge AI in Medicine Seminar Series\n\nhtt ps://talks.cam.ac.uk/talk/index/273034/ LOCATION:Jeffrey Cheah Biomedical Centre (Main Lecture Theatre)\, Univers ity of Cambridge URL:https://talks.cam.ac.uk/talk/index/273034/ CATEGORIES:Cambridge END:VEVENT BEGIN:VEVENT UID:deadline-2027-01-06-2@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART;VALUE=DATE:20270106 DTEND;VALUE=DATE:20270107 SUMMARY:Deadline: University Postgraduate Funding Competition\, second ro und DESCRIPTION:MASt Mathematics\, MPhil Scientific Computing\, MPhil Computa tional Biology and the Sensor CDT. CATEGORIES:Deadline END:VEVENT BEGIN:VEVENT UID:deadline-2027-05-13-3@cambridgeai4s.org DTSTAMP:20261025T143531Z DTSTART;VALUE=DATE:20270513 DTEND;VALUE=DATE:20270514 SUMMARY:Deadline: Most course deadlines DESCRIPTION:Departmental PhDs and several MPhils. Applying this late usua lly means applying without funding. CATEGORIES:Deadline END:VEVENT END:VCALENDAR