324139 research outputs found
Sort by
Data-driven performance guarantees for parametric optimization problems
Get PDFWe propose a data-driven method to establish probabilistic performance guarantees for parametric optimization problems solved via iterative algorithms. Our approach addresses two key challenges: providing convergence guarantees to characterize the worst-case number of iterations required to achieve a predefined tolerance, and upper bounding a performance metric after a fixed number of iterations. These guarantees are particularly useful for online optimization problems with limited computational time, where existing performance guarantees are often unavailable or unduly conservative. We formulate the convergence analysis problem as a scenario optimization program based on a finite set of sampled parameter instances. Leveraging tools from scenario optimization theory enables us to derive probabilistic guarantees on the number of iterations needed to meet a given tolerance level. Using recent advancements in scenario optimization, we further introduce a relaxation approach to trade the number of iterations against the risk of violating convergence criteria thresholds. Additionally, we analyze the trade-off between solution accuracy and time efficiency for fixed-iteration optimization problems by casting them into scenario optimization programs. Numerical simulations demonstrate the efficacy of our approach in providing reliable probabilistic convergence guarantees and evaluating the trade-off between solution accuracy and computational cost
Supervised rehabilitation versus self-managed rehabilitation for people with an acute first-time or recurrent patellar dislocation: the PRePPeD (Physiotherapy Rehabilitation Post Patellar Dislocation) external pilot randomised controlled trial and embedded qualitative study
No full textAims: To determine the feasibility of a full-scale randomised controlled trial (RCT) comparing two exercise-based rehabilitation interventions for people with an acute patellar dislocation.
Methods: A two-group external pilot RCT and embedded qualitative study conducted in five English National Health Service hospitals. Participants were aged ≥14 years with an acute (recruited ≤21 days of injury) first-time or recurrent patellar dislocation. Randomisation was 1:1 to supervised rehabilitation (4-6 physiotherapy sessions of tailored advice and prescribed home exercise) or self-managed rehabilitation (one physiotherapy session of advice, exercise instruction, and provision of materials to guide self-management). Quantitative feasibility objectives were 1.) patients’ willingness to be randomised, 2.) participant recruitment, 3.) intervention adherence (overall proportion of supervised rehabilitation participants that attended at least four physiotherapy sessions and self-managed rehabilitation participants that attended at least one session), and 4.) retention. Follow-up was three, six, and nine months after randomisation. There was no blinding. Semi-structured interviews aimed to understand participants’ experience of recovery, and the acceptability to them of the RCT interventions and follow-up methods.
Results: 50/88 (57%, 95% CI 46% to 67%) eligible patients were willing to be randomised. Sites recruited a mean of 1.4 (95% CI 0.6 to 1.8) participants per month, intervention adherence was 72% (95% CI 58% to 83%), and nine-month participant retention was 62% (95% CI 48% to 74%). During follow-up, three participants redislocated their patella and another underwent patellar stabilisation surgery. Interviews with nine pilot RCT participants found that the experience of recovery was conveyed through the themes ‘coming to terms with the initial injury’ and ‘regaining my former self’. Interviews also indicated that the RCT interventions and follow-up methods were generally acceptable to participants.
Conclusions: A full-scale RCT comparing two exercise-based rehabilitation interventions is feasible with minor modifications. Modifications should prioritise improving retention and participant attendance at physiotherapy sessions.
Registration: ISRCTN1423523
Collective asymmetric synthesis of the Strychnos alkaloids via thiophene S,S-dioxide cycloadditions
No full textThe Strychnos alkaloids have long been regarded as landmark targets for chemical synthesis due to their captivating architectures and notorious biological properties. However, the design of approaches that access multiple family members in an asymmetric, concise, and atom economical fashion remains a significant challenge. Here we show that thiophene S,S-dioxides (TDOs) offer a modular, rapid entry to the Strychnos natural products via inverse electron demand Diels-Alder cascades. Exceptional levels of stereocontrol are demonstrated in asymmetric TDO cycloadditions, which afforded tricyclic indolines of utility in medicinal chemistry research, and enabled stereoselective syntheses of eight Strychnos alkaloids by the shortest routes described to date, including the first synthesis of the iconic family member brucine. Using a machine-learning approach, computational studies provide insight into the source of stereoinduction, and reveal an intriguing and unexpected spontaneous cheletropic extrusion of SO2
Citizen deliberation and a communalist ethics of freedom
No full textUnlike many mainstream accounts of deliberation, Kwasi Wiredu offers a distinctly ethical conceptualization that foregrounds a robust accounting of pluralism in understanding the distinctive, general good of deliberation. Wiredu’s account partially emerges from his considerations of an African political geography where the pluralism of high ethnic stratification is structured into all political life. His broader arguments underscore the benefits to Africa and everywhere else of a robust consideration of pluralism and its ethical significance to deliberative understanding. Indeed, developing Wiredu’s arguments corrects a longstanding oversight in much deliberative theory which fails to treat pluralism robustly. Consequently, much mainstream theory misses a crucial, ethical, understanding of deliberation and its distinctive good. Developing Wiredu’s political and moral arguments, the article argues that deliberation fulfils a distinctly ethical function in enabling citizens’ moral consideration of each other and of those interests that are most meaningful to them in ways that are fundamental to the communal fabric of free society. This challenges the dominant justificatory view that sees deliberation predominantly as a means of justifying law and policy. By drawing out the moral and philosophical implications of Wiredu’s arguments, the article argues that the justificatory view fails to account for a normatively coherent understanding of deliberation because it fails, also, at an ethically robust understanding of the kind of difference—and agreement—with which deliberation is concerned. Further, with a field experiment, the article demonstrates that the empirical claim embedded in the justificatory view about the adherence of participants to deliberative decisions lacks support
ELIP: enhanced visual-language foundation models for image retrieval
Get PDFThe objective in this paper is to improve the performance of text-to-image retrieval. To this end, we introduce a new framework that can boost the performance of large-scale pre-trained vision-language models, so that they can be used for text-to-image re-ranking. The approach, Enhanced LanguageImage Pre-training (ELIP), uses the text query, via a simple MLP mapping network, to predict a set of visual prompts to condition the ViT image encoding. ELIP can easily be applied to the commonly used CLIP, SigLIP and BLIP-2 networks. On the evaluation side, we set up two new out-of-distribution (OOD) benchmarks, Occluded COCO and ImageNet-R, to assess the zeroshot generalisation of the models to different domains. The results demonstrate that ELIP significantly boosts CLIP/SigLIP/SigLIP2 text-to-image retrieval performance and outperforms BLIP-2 on several benchmarks, as well as providing an easy means to adapt to OOD datasets
On the origins of theories of compounding and the question of Indian influence on modern linguistics
No full textThe ancient Indian linguistic tradition has been implicated in various ways in the development of modern linguistics. This paper starts from one superficially obvious point of influence: the use of Sanskrit terms for compound classes in modern approaches to compound classification. I show that the use of these terms does not reflect direct influence based on a proper understanding of the Indian tradition, and that in fact many interesting and important aspects of compound theorization in ancient India have been overlooked in modern linguistics
Paired parton trial states for the superfluid-fractional Chern insulator transition
Get PDFWe consider a model of hard-core bosons on a lattice, half-filling a Chern band such that the system has a continuous transition between a fractional Chern insulator (FCI) and a superfluid state (SF) depending on the bandwidth to bandspacing ratio. We construct a parton-inspired trial wavefunction ansatz for the ground states that has remarkably high overlap with exact diagonalization in both phases and throughout the phase transition. Our ansatz is stable to adding some bosonic interactions beyond the on-site hard core constraint. We confirm that the transition is well described by a projective translation symmetry-protected multiple parton band gap closure, as has been previously predicted. However, unlike prior work, we find that our wavefunctions require anomalous (BCS-like) parton correlations to describe the phase transition and SF phase accurately
The effect of Indigenous American genomic ancestry on type 2 diabetes in Mexico: an analysis of 134 548 individuals from the Mexico City Prospective Study
No full textBackground: The prevalence of type 2 diabetes in Mexico is among the highest in the world and a major public health problem. The aim of this study was to examine the association between the percentage of Indigenous American genomic ancestry (AMR) and the prevalence of both prediabetes and type 2 diabetes in a large study of Mexican adults.
Methods: In this cross-sectional study, we analysed data from 134 548 individuals from the Mexico City Prospective Study (MCPS), including sociodemographic, clinical, and genetic data. Type 2 diabetes was defined as a self-reported previous diagnosis, glucose-lowering medication, or glycated hemoglobin (HbA1c) of at least 6·5%. Prediabetes was defined as HbA1c between 5·7% and less than 6·5%. Individuals with probable type 1 diabetes were excluded. Logistic regression models were used to estimate associations between higher AMR ancestry percentage and prediabetes and type 2 diabetes, after adjustment for age, sex, and other diabetes risk factors.
Findings: Between April 14, 1998, and Sept 28, 2004, 159 755 participants were recruited into MCPS. Among these, 134 548 participants were selected for subsequent analyses (mean age 52 years [SD 12·6]; 90 688 [67·4%] women and 43 860 [32·6%] men); mean AMR ancestry percentage was 66·2% (SD 17·9). Across AMR tenths (mean AMR ranging from 34·8% [SD 7·9] to 94·7% [SD 2·7]), the prevalence of prediabetes increased from 19·2% (2573 of 13 376) to 26·3% (3529 of 13 421) and of type 2 diabetes from 13·5% (1805 of 13 376) to 23·4% (3142 of 13 421). After adjustment for age and sex, each 20% absolute increase in AMR was associated with substantially increased odds of type 2 diabetes (odds ratio [OR] 1·45 [95% CI 1·43–1·48]) and of prediabetes (OR 1·28 [1·26–1·30]). Further adjustment for socioeconomic status, lifestyle factors, and adiposity reduced these ORs to 1·33 (1·31–1·36) for type 2 diabetes and 1·18 (1·16–1·20) for prediabetes. Consequently, at the population mean AMR ancestry percentage, these fully adjusted ORs were 2·59 (2·44–2·75) for type 2 diabetes and 1·75 (1·65–1·85) for prediabetes, whereas among Indigenous populations with 100% AMR ancestry, ORs were 4·23 (3·85–4·64) for type 2 diabetes and 2·33 (2·14–2·54) for prediabetes. ORs for type 2 diabetes and prediabetes associated with higher AMR were higher for women than for men, for younger than for older participants, and were further reduced in magnitude following additional adjustment for a type 2 diabetes polygenic risk score.
Interpretation: The percentage of inherited AMR ancestry is strongly associated with prediabetes and type 2 diabetes in this admixed Mexican population. These findings suggest that most of the Mexican population has a higher genetic susceptibility to type 2 diabetes than European-ancestry populations, underscoring the need for health systems to implement earlier, more intensive, and population-targeted preventive strategies
Exponential distillation of dominant eigenproperties
Get PDFEstimating observable expectation values in eigenstates of quantum systems has a broad range of applications and is an area where early fault-tolerant quantum computers may provide practical quantum advantage. We develop a hybrid quantum-classical algorithm that enables the estimation of an arbitrary observable expectation value in an eigenstate, given an initial state is supplied that has dominant overlap with the targeted eigenstate – but may overlap with any other eigenstates. Our approach builds on, and is conceptually similar to purification-based error mitigation techniques; however, it achieves exponential suppression of algorithmic errors using only a single copy of the quantum state. The key innovation is that random time evolution is applied in the quantum computer to create an average mixed quantum state, which is then virtually purified with exponential efficacy. We prove rigorous performance guarantees and conclude that the complexity of our approach depends directly on the energy gap in the problem Hamiltonian and remarkably, can be compared to phase estimation combined with amplitude estimation in terms of its scaling with respect to a target precision. We demonstrate in a broad range of numerical simulations the applicability of our framework in near-term and early fault-tolerant settings. Furthermore, we demonstrate in a 100-qubit example that direct classical simulation of our approach enables the prediction of ground and excited state properties of quantum systems using tensor network techniques, which we recognize as a quantum-inspired classical approach
Analysing student engagement with large language models in higher education: prompts as channels of communication with AI
Get PDFBackground: The aim of this study is to establish concrete insights into the ways in which students in higher education are engaging with large language models (LLM) through an analysis of the prompts they have written. Integrating frameworks from Human-Machine Communication (HMC) studies and student engagement literature, this research intends on elucidating prompts as a valuable source of analysis for the effect that LLMs have on students’ approaches to their academic work. These findings will inform scholarly perspectives on the precise strategies, participations and mannerisms students employ within their communications with LLMs.Method: This research utilises a questionnaire to collect samples of conversations and individual prompts from students' interaction with LLMs in academic contexts. It also asks for their academic background, such as what programme they enrolled in and previous areas of academic expertise. It analyses over 800 prompts from 23 students in higher education within the United Kingdom.Analysis of this data employed an integrative approach, relating Guzman and Lewis’s (2022) framework for human-AI communications with dimensions of student engagement. This approach enables the analysis of prompt data not only within its specific context as a unique mode of communication with a technological entity but also sheds light on the pedagogical implications of these interactions.Results: The results from this study presents four main types of engagement with large language models from students in higher education: linguistic manipulation, engagement with corpus data, controlling the LLM and ideation. These engagements reveal how students are mostly benefitting from the LLM’s ability to reword large bodies of text into different styles of language, as well as the ability to navigate the corpus data as a resource for information. Students also utilised LLMs to generate responses that would assist with the exploration of new ideas, as well as the evaluation of ideas. A large portion of prompts were also employed to control the LLM, directing the LLM to respond in certain ways or attempting to fix the LLM’s interpretation of previous prompts