107159 research outputs found
Sort by
Testing stationarity and change point detection in reinforcement learning
No full textWe consider reinforcement learning (RL) in possibly nonstationary environments. Many existing RL algorithms in the literature rely on the stationarity assumption that requires the state transition and reward functions to be constant over time. However, this assumption is restrictive in practice and is likely to be violated in a number of applications, including traffic signal control, robotics and mobile health. In this paper, we develop a model-free test to assess the stationarity of the optimal Q-function based on pre-collected historical data, without additional online data collection. Based on the proposed test, we further develop a change point detection method that can be naturally coupled with existing state-of-the-art RL methods designed in stationary environments for online policy optimization in nonstationary environments. The usefulness of our method is illustrated by theoretical results, simulation studies, and a real data example from the 2018 Intern Health Study. A Python implementation of the proposed procedure is publicly available at https://github.com/limengbinggz/CUSUM-RL
Defining a framework for sustainable global biosimilars markets using findings from a targeted literature review
No full textA biosimilar is a biologic medication that is highly similar to and has no clinically meaningful differences from an existing approved biologic referred to as "reference product." From the introduction of the first biosimilar in 2006 to today, a variety of challenges to biosimilar development and uptake have arisen across global markets, threatening sustainability. Consequences of an unsustainable market can include drug shortages, limited competition, and less innovation. However, there are few frameworks to facilitate systematic evaluation and action to address these threats. This study used a contemporary, targeted review of the global biosimilars literature to establish the key dimensions of biosimilar market sustainability. The most commonly referenced stakeholder groups were healthcare payers, government/legal/regulatory authorities, healthcare providers, biologic manufacturers, patients, and biologic purchasers. The most prevalent sustainability dimensions discussed were pricing and cost-savings, legal and regulatory barriers to market entry and access, manufacturer processes, provider choice in selecting biologic therapy, knowledge and preferences, and procurement processes. We incorporated these findings into a framework of biosimilar market sustainability dimensions that should be considered by stakeholders looking to ensure the long-term viability of the market
Captivity’s collections: science, natural history, and the British transatlantic slave trade, by Kathleen S. Murphy
No full textKathleen S. Murphy, Captivity’s Collections: Science, Natural History, and the British Transatlantic Slave Trade. Chapel Hill: University of North Carolina Press, 2023. xiv + 239 pp. (Paper US$ 29.95
How data governance happens (and why it matters)
No full textWhat does data governance mean? The vague meaning has consequences for ethics and AI. Chris Wiggins and Alison Powell argue it’s time to articulate a vision of governance that centres social participation, shared decision-making and equitable distribution of benefits and risks
Xi, Putin and the struggle for "history"
No full textHistory and memory as much as interests play a critical role in sustaining the Russia-China relationship today, writes Michael Cox
Bridging research and practice: funders as drivers for social equity
No full textFunders of projects to research and reduce global inequalities increasingly look for academics and practitioners to work together, writes Aygen Kurt-Dickson. But how do these “AcPrac” collaborations come about? Who is funding such schemes, and what determines their success
Why is food insecurity worsening in Africa?
No full textHow Africa Eats: Trade, Food Security and Climate Risks examines why food insecurity is so prevalent in Africa and how a confluence of factors (including trade and agriculture policies and climate change) shape the issue. In this extract from the introduction to the book, editor David Luke traces how and why food deprivation has deepened over the past decade, including inflation, the COVID-19 pandemic and Russia’s war in Ukraine. How Africa Eats: Trade, Food Security and Climate Risks. David Luke (ed.). LSE Press. 2025
Can investor coalitions drive corporate climate action?
No full textThis paper investigates the effectiveness of collective investor engagement in driving corporate climate action. Empirically, I focus on Climate Action 100+ (CA100+), the world’s largest investor coalition on climate change. To address common measurement issues in previous research, I conduct a multidimensional assessment of companies’ climate action. In particular, I collect new primary data on the ambition of carbon emission reduction targets and use the ClimateBERT model to analyse climate-related disclosure. To isolate the causal impact of CA100+, I examine the selection of the coalition’s focus companies and employ a Difference-in-Differences analysis. While the findings suggest that CA100+ has had no effect on companies’ disclosures or reductions in carbon emissions, I observe a significant impact on targets. However, this effect holds only for medium- and long-term targets, not in the short-term, and is exclusively driven by companies potentially selected based on prior investor knowledge. Overall, this study finds limited effectiveness of collective engagement through CA100+. It raises questions about the importance of investor selectivity for engagement success and highlights the risk of companies backloading their decarbonisation efforts
Insights into weighted sum sampling approaches for multi-criteria decision making problems
No full textIn this paper we explore several approaches for sampling weight vectors in the context of weighted sum scalarisation approaches for solving multi-criteria decision making (MCDM) problems. This established method converts a multi-objective problem into a (single) scalar optimisation problem. It does so by assigning weights to each objective. We outline various methods to select these weights, with a focus on ensuring computational efficiency and avoiding redundancy. The challenges and computational complexity of these approaches are explored and numerical examples are provided. The theoretical results demonstrate the trade-offs between systematic and randomised weight generation techniques, highlighting their performance for different problem settings. These sampling approaches will be tested and compared computationally in an upcoming paper