Speaker: Arnak DalalyanVenue: Room 1417, Administration Building, Zijingang CampusAbstract: Generative modeling aims to produce new random examples from an unknown target distribution, given access to a finite collection of examples. Among the leading approaches, denoising diffusion probabilistic models (DDPMs) construct such examples by mapping a Brownian motion via a diffusion process driven by an estimated score function. In this work, we first provide empirical evidence that DDPMs are robust to constant-variance noise in the score evaluations. We then establish finite-sample guarantees in Wasserstein-2 distance that exhibit two key features: (i) they characterize and quantify the robustness of DDPMs to noisy score estimates, and (ii) they achieve faster convergence rates than previously known results. Furthermore, we observe that the obtained rates match those known in the Gaussian case, implying their optimality. (Joint work with V. Arsenyan and E. Vardanyan)

Speaker: Yafei LiangVenue: Room 426, School of Economics, Zijingang CampusAbstract: This study examines how large language models (LLMs) affect language barriers and disparities in publication outcomes and research impact for English-as-a-Foreign-Language (EFL) scholars. Using the release of ChatGPT-3.5 in November 2022 as a natural experiment, we analyze paper metadata from the Social Science Research Network (SSRN) across seven subject areas. We find that LLM usage increased significantly after November 2022 and that EFL-authored papers show greater improvements in language proficiency than non-EFL papers, particularly in Social Science. However, this improvement in writing quality is accompanied by only a modest increase in publication success and no detectable increase in research impact. These findings highlight both the promise and limitations of LLMs: while language gaps narrow, deeper disparities in scientific impact persist.

Speaker: Kallol RayVenue: Room 203, Haina Building 2, Zijingang CampusAbstract: This work investigates both steady and unsteady nonlinear systems that couple the generalized convective Brinkman-Forchheimer model with a system of advection-diffusion equations, commonly referred to as double-diffusion equations. The existence and uniqueness of weak solutions to the governing equations are established using Galerkin’s method. Subsequently, H(div)-conforming discontinuous Galerkin (DG) discretizations are formulated for the considered models, yielding exactly divergence-free velocity approximations. For the unsteady model, a second-order semi-implicit backward differentiation formula (BDF2) scheme is employed for temporal discretization. A rigorous analysis is then carried out to establish the well-posedness of the discrete problems. Optimal a priori error estimates are derived, ensuring that the velocity errors are pressure-robust. Furthermore, when the diffusion coefficients are constant, the velocity error estimates are Re-semi-robust at high Reynolds numbers. Numerical experiments are presented to corroborate the theoretical results and to demonstrate the performance of the proposed methods.

Speaker: Yulei LuoVenue: Room 530, Chengjun Building 7, Zijingang CampusAbstract: Using firm-level data, we document that inventory investment rises with contemporaneous sales shocks and falls with lagged sales shocks. This reversal pattern is inconsistent with standard rational-expectations stock-adjustment models, which imply weak or counterfactual inventory-sales comovement. We show that a simple extension with diagnostic expectations accounts for these dynamics. In the model, managers overweight representative demand signals, generating closed-form policy rules under which inventories initially overreact to sales shocks and subsequently reverse. Quantitatively, the model matches key moments of inventory dynamics, including the sign of inventory-sales comovement, the relative volatility of inventory investment to sales, and the persistence of inventory dynamics. These results identify belief distortions as a central force in inventory adjustment and provide a behavioral explanation for both longstanding and new puzzles in inventory dynamics.

Speaker: Jingyu ZhangVenue: Room 530, Chengjun Building 7, Zijingang CampusAbstract: Serving on the official unsecured creditors’ committee (UCC) of a bankrupt firm provides hedge funds with access to material nonpublic information (MNPI), which can facilitate their informed trading across firms and asset markets. We find that hedge funds increase equity turnover and execute more large trades in the quarters following UCC membership. In contrast, hedge funds do not exhibit such trading behavior after accessing public information about bankrupt firms or holding the bankrupt firm’s debt without committee involvement. Importantly, these large trades often target firms with close economic ties to the bankrupt entity. Returns fm these MNPI-driven trades are substantial.

Speaker: Sven HirschVenue: Room 204, Haina Building 2, Zijingang CampusAbstract: We analyze the causal type of both Killing vector fields and Killing spinors. As an application we give a proof of Bartnik's stationary vacuum conjecture from 1989 and geometrically characterize Siklos wave spacetimes. This talk is based on joint work with Lan-Hsuan Huang and Yiyue Zhang.

Speaker: Geoffrey SmithVenue: Lecture Hall, Floor 1, School of Medicine, Zijingang CampusAbstract: Professor Geoffrey Smith earned his PhD in 1981 at the laboratory of Alan Hay at the National Institute for Medical Research, London, UK (the predecessor of the Francis Crick Institute), focusing on the replication mechanism of influenza virus. Between 1981 and 1984, he conducted postdoctoral research at the laboratory of Bernard Moss at the National Institutes of Health (NIH) in the United States. During this period, he pioneered the development of recombinant vaccine technology using vaccinia virus (the smallpox vaccine) as a vector, proposing the original theory of using genetic engineering to modify viruses as recombinant vaccines. This principle has since been widely applied to the development of vaccines against various viruses and microorganisms. Professor Smith was elected a Fellow of the Royal Society in 2003 and a Member of the German National Academy of Sciences Leopoldina in 2011.Professor Smith has served as Professor of Pathogen Microbiology at the University of Oxford, Head of the Department of Virology at Imperial College London, and Head of the Division of Pathology at the University of Cambridge. In 2023, he returned to the Centre for Molecular Pathology at the University of Oxford to continue his research. In 2025, he joined the Shanghai Institute of Immunity and Infection, Chinese Academy of Sciences. He currently serves as Chief Scientist of the Vaccine Research Center at the Shanghai Institute of Materia Medica, Chinese Academy of Sciences, and is Emeritus Professor at the Centre for Molecular Pathology, University of Oxford. His team’s main research directions include the interaction between poxviruses and host cells, the development of novel vaccines, and oncolytic viruses.

Speaker: Michael BarnettVenue: Room 901, Department of Public Affairs, Zijingang CampusAbstract: Global governance was first coined in the early and optimistic days following the end of the Cold War. Capitalizing on an ending to the Cold War no one expected, the international community experienced thoughts of new beginnings and possibilities. Confrontation and fear would now yield to cooperation and optimism. The world began building and expanding existing multilateral organizations to produce not only mutual gains but also a sense of global community. Around the turn of the century, though world order began to darken and notions of progress began to replace precarity. The drivers of the new worldview has led to a new round of global governance defined by precarity, uncertainty, and risk. What kinds of global institutions can and will be built in this new world order?

Speaker: Dr. Jeremy ArolesVenue: Room A723, School of Management Building, Zijingang CampusAbstract: Organizations in nascent categories face a distinctive legitimation challenge as they must establish temporal credibility – a persuasive account of where they came from and where they are going – when both the organisational past and the sector’s direction are unsettled. Existing research on uses of the past has primarily examined organisations with rich temporal resources; we know less about how temporal legitimation works when the past itself must be constructed rather than curated. We ask: How do actors in nascent organisational categories construct temporal legitimacy when the temporal resources available to them are scarce? We address this question through a qualitative study of coworking spaces across four global cities, through which we identify five temporal legitimation strategies operating under two interdependent mechanisms: constructing temporal ground (manufacturing a credible past) and claiming temporal trajectory (naturalising a credible future). We theorise these as forming an iterative cycle in which constructed depth makes trajectory claims credible, while projected futures redefine which pasts are usable. We contribute to research on legitimation in nascent categories by theorising temporal legitimation as a distinct mechanism of category construction, and to the uses-of-the-past literature by showing how temporal work operates under conditions of historical scarcity.