Counterfactual Strategies Research Articles

PSF-C-Net: A Counterfactual Deep Learning Model for Person Re-Identification Based on Random Cropping Patch and Shuffling Filling

In the task of person re-identification (re-ID), capturing the long-range dependency of instances is crucial for accurate identification. The existing methods excel at extracting local features but often overlook the global information of instance images. To address this limitation, we propose a convolution-based counterfactual learning framework, called PSF-C-Net, to focus on global information rather than local detailed features. PSF-C-Net adopts a parameter-sharing dual-path structure to perform counterfactual operations in the prediction space. It takes both the actual instance image and a counterfactual instance image that disrupts the contextual relationship as the input. The counterfactual framework enables the interpretable modeling of global features without introducing additional parameters. Additionally, we propose a novel method for generating counterfactual instance images, which effectively constructs an explicit counterfactual space, to reliably implement counterfactual strategies. We have conducted extensive experiments to evaluate the performance of PSF-C-Net on the Market-1501 and Duke-MTMC-reID datasets. The results demonstrate that PSF-C-Net achieves state-of-the-art performance.

Modelling COVID-19 Vaccination in the UK: Impact of the Autumn 2022 and Spring 2023 Booster Campaigns.

The delivery of COVID-19 vaccines was successful in reducing hospitalizations and mortality. However, emergence of the Omicron variant resulted in increased virus transmissibility. Consequently, booster vaccination programs were initiated to decrease the risk of severe disease and death among vulnerable members of the population. This study aimed to estimate the effects of the booster program and alternative vaccination strategies on morbidity and mortality due to COVID-19 in the UK. A Susceptible-Exposed-Infectious-Recovered (SEIR) model was used to assess the impact of several vaccination strategies on severe outcomes associated with COVID-19, including hospitalizations, mortality, National Health Service (NHS) capacity quantified by hospital general ward and intensive care unit (ICU) bed days, and patient productivity. The model accounted for age-, risk- and immunity-based stratification of the UK population. Outcomes were evaluated over a 48-week time horizon from September 2022 to August 2023 considering the actual UK autumn 2022/spring 2023 booster campaigns and six counterfactual strategies. The model estimated that the autumn 2022/spring 2023 booster campaign resulted in a reduction of 18,921hospitalizations and 1463 deaths, compared with a no booster scenario. Utilization of hospital bed days due to COVID-19 decreased after the autumn 2022/spring 2023 booster campaign. Expanding the booster eligibility criteria and improving uptake improved all outcomes, including averting twice as many ICU admissions, preventing more than 20% additional deaths, and a sevenfold reduction in long COVID, compared with the autumn 2022/spring 2023 booster campaign. The number of productive days lost was reduced by fivefold indicating that vaccinating a wider population has a beneficial impact on the morbidities associated with COVID-19. Our modelling demonstrates that the autumn 2022/spring 2023 booster campaign reduced COVID-19-associated morbidity and mortality. Booster campaigns with alternative eligibility criteria warrant consideration in the UK, given their potential to further reduce morbidity and mortality as future variants emerge.

Evaluating vaccine allocation strategies using simulation-assisted causal modeling

We develop a model to retrospectively evaluate age-dependent counterfactual vaccine allocation strategies against the coronavirus disease 2019 (COVID-19) pandemic. To estimate the effect of allocation on the expected severe-case incidence, we employ a simulation-assisted causal modeling approach that combines a compartmental infection-dynamics simulation, a coarse-grained causal model, and literature estimates for immunity waning. We compare Israel's strategy, implemented in 2021, with counterfactual strategies such as no prioritization, prioritization of younger age groups, or a strict risk-ranked approach; we find that Israel's implemented strategy was indeed highly effective. We also study the impact of increasing vaccine uptake for given age groups. Because of its modular structure, our model can easily be adapted to study future pandemics. We demonstrate this by simulating a pandemic with characteristics of the Spanish flu. Our approach helps evaluate vaccination strategies under the complex interplay of core epidemic factors, including age-dependent risk profiles, immunity waning, vaccine availability, and spreading rates.

Using a counterfactual-based intervention to change eating intentions: Results from Hispanic and non-Hispanic undergraduate and community samples

Finding ways to improve eating behavior has become a major focus of interventions designed to improve health outcomes. Counterfactual thinking (i.e., mental simulations of how a past outcome could have been different) is a cognitive strategy that has been shown to improve behavior regulation and could be a promising intervention to improve eating behavior. The purpose of the current research was to examine the effect of a counterfactual-based intervention to shift contemplation to change eating behavior, motivation towards eating healthier, and intentions to engage in healthier eating behaviors. Participants in both studies were randomly assigned to either a counterfactual-based intervention or a control condition. In Study 1, those in the counterfactual-based intervention condition reported an increase in readiness to change their eating behavior, higher extrinsic motivation towards eating healthy, and higher intentions to eat healthy compared to those in the control condition. Within the counterfactual condition, individuals who self-identified as Hispanic had higher intentions to use their counterfactual strategies than those who identified as non-Hispanic. Study 2 found similar results using a more diverse community sample. Those in the counterfactual-based intervention condition reported an increase in readiness to change, intentions to eat healthy, and intentions to use their counterfactual strategies, with this effect being stronger for self-identified Hispanic participants. These studies provide initial evidence for the use of a counterfactual-based intervention to improve eating behavior in diverse populations.

How Many Are Too Many? Analyzing Dockless Bikesharing Systems with a Parsimonious Model

Using a parsimonious model, this paper analyzes a dockless bikesharing (DLB) system in a city, where the DLB service competes with walking and a generic motorized mode. The DLB system operator chooses a fleet size and a fare schedule that dictate the utility of its service. The access time to the service---a crucial component in the utility---is modeled as a function of the number of unique bike locations, which is estimated from the number of idle bikes through an empirical relationship. The market equilibrium is formulated as a solution to a nonlinear equation system, over which three design problems are formulated to maximize (i) profit; (ii) ridership; and (iii) social welfare. We calibrate the model against empirical data collected in Chengdu, China and test three counterfactual strategies, each for one of the three design problems. We find: (i) the current fleet limit set by Chengdu should be cut by roughly two thirds in order to avoid severe oversupply and waste; (ii) maximizing ridership keeps the fare at an affordable level to maintain revenue neutrality, while raising ridership and social welfare close to system optimum; (iii) for a regulator seeking to influence the DLB operator for social good, the choice of policy instruments depends on the operator's objective. When it focuses on profit, limiting price is a much more effective policy than limiting fleet size. If, instead, the operator aims to grow its market share, then setting a limit on fleet size becomes a dominant strategy.

In Defence of Lost Causes: Counter(F)Actual Recasting of Biographies in Polish Contemporary Theatre

Summary This article considers the impact of counterfactual strategies on the most recent Polish theatrical practices dealing with biographies of “historical” figures. The re-occurrence of these past agents on the stage will be viewed in light of the biographical turn in the humanities as well as from the perspective of Jacques Derrida’s concept of hauntology. Seemingly, both trends share a need to create an alternative space for the expression of a contemporary self which is marked by disunity and disintegration. Subjects of current semi-biographical projects are those whose voices have once been neglected, marginalised, or oppressed because of their gender, social background, or political views. This account examines the ways in which counterfactual strategies enable us to grasp the polyphonic condition of a modern subject and to see, in traces left by different Other(s), touchstones for social and political change. By taking the play Tu Wersalu nie będzie! (No Versailles over here!) by Rabih Mroué as the core case study of the analysis, I aim to demonstrate how counterfactual strategies animate emancipatory potential ascribed to the arrival of the phantom of controversial Polish politician Andrzej Lepper. His death in unknown circumstances becomes a point of divergence in which Lepper’s existence layers into counterfactual scenarios. Counterfactual strategies enable many approaches to view Lepper’s figure without the ethically dubious act of speaking in his name. By unsettling claims of truth, counterfactual strategies unravel how “facts” about Lepper resurfaced in mass media, thereby constructing his stereotyped and over-generalised image. The play has a form of investigation which, by employment of counterfactualism, reenacts the oppression of a mainstream media discourse against the disturbing Other epitomised by Lepper.

Messianism and UFO Sightings as Counterfactual Strategies of Enchantment of the World

Summary In this text, I argue that there are numerous affinities between 19th century messianism and testimonies of UFO sightings, both of which I regarded as forms of secular millennialism. The common denominator for the comparison was Max Weber’s concept of “disenchantment of the world” in the wake of the Industrial Revolution which initiated the era of the dominance of rational thinking and technological progress. However, the period’s counterfactual narratives of enchantment did not repudiate technology as the source of all social and political evil—on the contrary, they variously redefined its function, imagining a possibility of a new world order. In this context, I analysed the social projects put forward by Polish Romantics in the first half of the 19th century, with emphasis on the role of technology as an agent of social change. Similarly, the imaginary technology described by UFO contactees often has a redemptive function and is supposed to bring solution to humanity’s most dangerous problems.

Implementation intentions forge a strong cue–response link and boost prospective memory performance

Prospective memory (PM) performance (i.e., remembering to perform intentions) is often in the service of higher goals and thus could be improved through better self-regulation. We examined the effectiveness of three self-regulatory strategies in a laboratory event-based PM task: implementation intention (i.e., planning in an if/when–then format), rehearsal (i.e., planning not in an if/when–then format) and upward counterfactual (“if only…” thoughts about a past performance). The cue and response mentioned in all three strategies were held constant, such that only format and timing (past vs. future orientation) of the strategy varied. Results indicated that the implementation intention strategy strengthened the cue–response link and improved PM performance over baseline, relative to a control condition. Neither rehearsal nor counterfactual strategies had significant benefits, relative to a control condition. These findings suggest that the if/when–then plan format of implementation intentions may be a critical feature of their effectiveness in retrieving and executing intentions.

Counterfactual geographies: worlds that might have been

This paper argues for a renewed consideration of counterfactuals within geography. Drawing upon Doreen Massey's emphasis on notions of ‘possibility’, ‘chance’, ‘undecidability’ and ‘happenstance’, we argue for an engagement with approaches in the humanities that have addressed such issues directly. We review previous uses of counterfactual method in historical geography, particularly as related to cliometrics and the ‘new economic history’ of the 1960s, but argue that a recent upsurge of interest in other disciplines indicates alternative ways that ‘what-if’ experiments might work in the sub-discipline. Recent counterfactual work outside of geography has had a notably spatial cast, often thinking through the nature of alternative worlds, or using counterfactual strategies that are explicitly concerned with space as well as temporal causality. We set out possible agendas for counterfactual work in historical geography. These include: consideration of the historical geographies within existing counterfactual writings and analyses; suggestions for distinctive ways that historical geographers might think and write counterfactually, including experiments in geographies of happenstance, and the exploration of more-than-human possibilities; analyses of the geography of and in counterfactual writing; and study of the political, ethical and emotional demands that counterfactuals make. This discussion and framework provides an extended introduction to this special feature on counterfactual geographies.