Relevant Benchmark Research Articles

On the use of ensemble averaging techniques to accelerate the Uncertainty Quantification of CFD predictions in wind engineering

In this work we focus on reducing the wall clock time required to compute statistical estimators of highly chaotic incompressible flows on high performance computing systems. Our approach consists of replacing a single long-term simulation by an ensemble of multiple independent realizations, which are run in parallel. A failure probability convergence criteria must be satisfied by the statistical estimator of interest to assess convergence. The error analysis leads to the identification of two error contributions: the initialization bias and the statistical error. We propose an approach to systematically detect the burn-in time needed in order to minimize the initialization bias, as well as techniques to choose the effective time needed to keep the statistical error under control. This is accompanied by strategies to reduce simulation cost. The proposed method is assessed in application to the prediction of the drag force over high rise buildings and specifically in application to the CAARC building, a relevant benchmark for the wind engineering community.

The Regulatory Perspectives to China’s Emerging Hydrogen Economy: Characteristics, Challenges, and Solutions

Culminating two decades of industrial planning, China is now officially moving full steam ahead towards realizing a hydrogen economy under the country’s first ever Medium and Long-term Plan of the Hydrogen Industry (2021–2035). Among the relevant benchmarks established is the goal of developing a regulatory framework by 2025. This raises the question of how best to achieve a regulatory framework for China’s emerging hydrogen economy. To answer this question, the discussions of this paper are further broken-down and organized across four independent, but correlated, academic questions. One, relying on the fragmented authoritarianism model, what are the impacts of China’s current model of industrial development on an emerging regulated industry? Two, through a scientific and technological review, what are the characteristics of the hydrogen supply-chain most likely to present a regulatory challenge for China? Three, by analogy to the comparable experiences of China’s other regulated industries, what are the possible regulatory solutions? Four, and most importantly, how best to reconcile the findings to the above questions as they relate to the regulatory challenge of developing China’s emerging hydrogen economy. The results of the discussions reveal, that not all policy solutions and recommendations to the regulatory framework of the hydrogen economy should be treated equally. Rather, an integrated view of the core academic question revealed a procedural relationship among the regulatory solutions identified from the analysis above. Therefore, recognizing that each solution should synergize and correspond to different phases of regulatory development, a three-step regulatory pathway towards the hydrogen economy is proposed.

AppClassNet

The recent success of Artificial Intelligence (AI) is rooted into several concomitant factors, namely theoretical progress coupled with abundance of data and computing power. Large companies can take advantage of a deluge of data, typically withhold from the research community due to privacy or business sensitivity concerns, and this is particularly true for networking data. Therefore, the lack of high quality data is often recognized as one of the main factors currently limiting networking research from fully leveraging AI methodologies potential. Following numerous requests we received from the scientific community, we release AppClassNet, a commercial-grade dataset for benchmarking traffic classification and management methodologies. AppClassNet is significantly larger than the datasets generally available to the academic community in terms of both the number of samples and classes, and reaches scales similar to the popular ImageNet dataset commonly used in computer vision literature. To avoid leaking user- and business-sensitive information, we opportunely anonymized the dataset, while empirically showing that it still represents a relevant benchmark for algorithmic research. In this paper, we describe the public dataset and our anonymization process. We hope that AppClassNet can be instrumental for other researchers to address more complex commercial-grade problems in the broad field of traffic classification and management.

Optimal multilevel adaptive FEM for the Argyris element

The main drawback for the application of the conforming Argyris FEM is the labourious implementation on the one hand and the low convergence rates on the other. If no appropriate adaptive meshes are utilised, only the convergence rate caused by corner singularities (Blum and Rannacher, 1980), far below the approximation order for smooth functions, can be achieved. The fine approximation with the Argyris FEM produces high-dimensional linear systems and for a long time an optimal preconditioned scheme was not available for unstructured grids. This paper presents numerical benchmarks to confirm that the adaptive multilevel solver for the hierarchical Argyris FEM from Carstensen and Hu (2021) is in fact highly efficient and of linear time complexity. Moreover, the very first display of optimal convergence rates in practically relevant benchmarks with corner singularities and general boundary conditions leads to the rehabilitation of the Argyris finite element from the computational perspective.

DOCKSTRING: Easy Molecular Docking Yields Better Benchmarks for Ligand Design.

The field of machine learning for drug discovery is witnessing an explosion of novel methods. These methods are often benchmarked on simple physicochemical properties such as solubility or general druglikeness, which can be readily computed. However, these properties are poor representatives of objective functions in drug design, mainly because they do not depend on the candidate compound’s interaction with the target. By contrast, molecular docking is a widely applied method in drug discovery to estimate binding affinities. However, docking studies require a significant amount of domain knowledge to set up correctly, which hampers adoption. Here, we present dockstring, a bundle for meaningful and robust comparison of ML models using docking scores. dockstring consists of three components: (1) an open-source Python package for straightforward computation of docking scores, (2) an extensive dataset of docking scores and poses of more than 260,000 molecules for 58 medically relevant targets, and (3) a set of pharmaceutically relevant benchmark tasks such as virtual screening or de novo design of selective kinase inhibitors. The Python package implements a robust ligand and target preparation protocol that allows nonexperts to obtain meaningful docking scores. Our dataset is the first to include docking poses, as well as the first of its size that is a full matrix, thus facilitating experiments in multiobjective optimization and transfer learning. Overall, our results indicate that docking scores are a more realistic evaluation objective than simple physicochemical properties, yielding benchmark tasks that are more challenging and more closely related to real problems in drug discovery.

Strengthening systems of care for people with or at risk for HIV, HCV and opioid use disorder: a call for enhanced data collection

Background The syndemic between opioid use disorder (OUD), hepatitis C virus (HCV), and human immunodeficiency virus (HIV) results in excessive burdens on the healthcare system. Integrating these siloed systems of care is critical to address all three conditions adequately. In this implementation project, we assessed the data capacity of the health system to measure a cascade of care (COC) across HIV, HCV and OUD services in five states to help guide public health planning. Materials and methods Data for this study were gathered from publicly available datasets and reports from government (SAMSHA, CMS, HRSA and CDC) sites. We created, where possible, COCs for HIV, HCV, and OUD spanning population estimate, diagnosis, treatment initiation, treatment retention, and patient outcomes for each of five states in the study. Results The process of data collection showed that baseline COCs examining the intersections of OUD, HIV, and HCV cannot be produced and that there are missing data in all states examined. Collection of specific data points is not consistent across all states. States are better at reporting HIV cascades due to federal requirements. Only gross estimates could be made for OUD cascades in all states because data are separated by payer source, leaving no central point of data collection from all sources. Data for HCV were not publicly available. Conclusion It is difficult to assess the strategies needed or the progress made towards increasing treatment access and decreasing the burden of disease without the ability to construct an accurate baseline. Using integrated COCs with relevant benchmarks can not only guide public health planning, but also provide meaningful targets for intervention. KEY MESSAGES While HIV COCs are available for most states at least annually, they are not disaggregated for populations with co-occurring OUD or HCV. Data to calculate HCV COC are not available and data to calculate OUD COC are partially available, but only for specific payers. States do not have systems in place to measure the scope of the syndemic or to identify targets for quality improvement activities.

A Note on the Optimality of Domain-specific Liability

Abstract This note analyzes the socially optimal allocation of liability when both consumers and the environment incur harm from the activity of a monopolistic firm. We show that the marginal welfare effect from a greater extent of loss shifting depends on the domain of harm (consumer vs. environment) and the relationship between the harm level and the level of output (proportional vs. non-proportional). Starting from the relevant benchmark of full compensation in both domains, reducing the firm’s liability for environmental harm is welfare-improving whereas reducing the firm’s liability for consumer harm is welfare-decreasing when harm increases more than proportionally with the quantity produced.

Abstract LB220: The anticalin-antibody bispecific PRS-352/S095025 strongly stimulates human CD4+ T cells in a PD-L1-dependent manner

Abstract Background: OX40, a member of the tumor necrosis factor receptor superfamily, is a costimulatory molecule promoting survival and cytokine release by effector T cells and inhibiting human regulatory T-cell activity. OX40 activation has been explored as a potential treatment for solid malignancies, using mainly agonistic anti-OX40 antibodies. These antibodies rely on Fc gamma receptor (FcγR)-mediated crosslinking to activate the OX40 pathway and have shown so far only limited antitumoral efficacy in clinical settings. One underlying hypothesis is that FcγR crosslinking of these antibodies may not allow optimal activation of the OX40 pathway to fully exploit the potential of this target for immunotherapy of cancer. To overcome this limitation, we designed PRS-352/S095025, a strong PD-L1-dependent OX40 agonistic bispecific molecule, that allows combination of OX40 costimulation and PD-L1 blockade. Methods: We generated an Anticalin protein that binds with high affinity to human OX40. Anticalin® proteins are approximately 18-20 kDa protein therapeutics derived from human lipocalins that can be engineered to bind with high affinity and specificity to different targets. The PRS-352/S095025 bispecific fusion protein was obtained by genetic fusion of OX40-targeting Anticalin protein to a PD-L1-targeting monoclonal antibody with a modified IgG4 backbone, allowing an optimal activation of OX40 in the presence of PD-L1, but not FcγRs. Results: We showed that PRS-352/S095025 retains binding to PD-L1 like the anti PD-L1 antibody backbone alone and binds with high affinity to both human and cynomolgus OX40. We showed that PRS-352/S095025 inhibited the PD-1/PD-L1 pathway, as the parental PD-L1 targeting antibody and an approved anti PD-L1 antibody. In addition to its PD-L1 blocking properties, we showed that PRS-352/S095025 activates OX40, and that this activity is dependent on the expression of PD-L1 and not on FcγR, consistent with the desired mechanism of action. Furthermore, we demonstrated that PRS-352/S095025 activity is superior to both PD-L1 benchmark antibodies and to the combination of clinically relevant OX40 and PD-L1 benchmark antibodies. We could also show that PRS352/S095025 strongly stimulated human CD4 T cells. In vivo, we showed that PRS-352/S095025 has a PK profile similar to that of the PD-L1 antibody alone. Conclusions: We provide here the preclinical characterization of the fusion protein PRS-352/S095025, a bispecific molecule composed of a PD-L1 blocking moiety and an Anticalin protein agonizing OX40. In vitro, this molecule showed the desired MoA, PD-L1 blocking and potent OX40 agonism driven by binding to PD-L1, with superior activity to a clinical stage OX40 agonist. Citation Format: Lucia Pattarini, Marina Pavlidou, Aizea Kastresana Morales, Janet Peper-Gabriel, Matthieu Riviere, Didier Demarles, Alix Scholer-Dahirel, Veronique Blanc, Shane Olwill. The anticalin-antibody bispecific PRS-352/S095025 strongly stimulates human CD4+ T cells in a PD-L1-dependent manner [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr LB220.

Multiple Treatment Modeling for Target Marketing Campaigns: A Large-Scale Benchmark Study

Machine learning and artificial intelligence (ML/AI) promise higher degrees of personalization and enhanced efficiency in marketing communication. The paper focuses on causal ML/AI models for campaign targeting. Such models estimate the change in customer behavior due to a marketing action known as the individual treatment effect (ITE) or uplift. ITE estimates capture the value of a marketing action when applied to a specific customer and facilitate effective and efficient targeting. We consolidate uplift models for multiple treatments and continuous outcomes and perform a benchmarking study to demonstrate their potential to target promotional monetary campaigns. In this use case, the new models facilitate selecting the optimal discount amount to offer to a customer. Large-scale analysis based on eight marketing data sets from leading B2C retailers confirms the significant gains in the campaign return on marketing when using the new models compared to relevant model benchmarks and conventional marketing practices.

AERO: Design Space Exploration Framework for Resource-Constrained CNN Mapping on Tile-Based Accelerators

Analog In-Memory Compute (AIMC) arrays can store weights and perform matrix-vector multiplication operations for Deep Convolutional Neural Networks (CNNs). A number of recent efforts have integrated AIMC arrays into hybrid digital-analog accelerators in a multi-layer parallel manner to achieve energy efficiency and high throughput. Multi-layer parallelism on large-scale tile-based architectures need efficient mapping support at the processing element (PE)-level ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e.g.</i> , digital or analog processing elements) and tile-level. To find the most efficient architectures, fast and accurate design space exploration (DSE) support is required. In this paper, a novel DSE framework, AERO, is presented to characterize a CNN inference workload executing on hybrid tile-based architectures that supports multi-layer parallelism. Three characteristics can be seen in our DSE framework: (1) It presents a hierarchical Tile/PE-level mapping exploration strategy including inter-layer interaction, and allowing layer fusion/splitting configurations for PE-level mapping optimization. (2) It unlocks different Performance, Power and Area (PPA) exploration points under both sufficient and limited resource constraints, while limited resource case is not considered in prior works of multi-layer parallel architectures. The impact of weight loading and weight stationary mapping are analyzed for better insights into hybrid tile-based architectures. (3) It incorporates a detailed PPA model that supports a broad range of hybrid digital and analog units in a tile. Experimental case-studies are performed for realistic and relevant benchmarks such as MLP, CNNs (Lenet-5, Resnet-18,-34,-50 and −101).

Study of the Skills of Balanced Fund Managers in Poland

The market timing is one of the active methods used by portfolio managers to do their investments more effective. It allows to separate management skills on a micro and macro scale. The market timing applies to the appropriate selection of assets for the portfolio and the right moment to change its structure. The aim of the presented research was to check whether the managers of balanced open-end mutual funds operating on the Polish market apply market timing skills. Nine balanced funds have been accepted for the research, which have existed since at least 2003 year. The research period covered the years 2003-2019. WIBOR 1M was used as the risk-free rate, and the market factors were the main WSE indexes. The research was first conducted based on basic market timing research models, i.e. the Henriksson-Merton and Treynor-Mazuy. Then, these models were expanded to include factors related to the bond market. The Henriksson-Merton and Treynor-Mazuy models and their extensions with additional factors were compared among themselves. Studies show that models with additional factors have proven to be more appropriate for balanced open-end mutual funds. It has occurred that regardless of the model used, market timing skills were similar. In most cases the fund managers did not achieve higher results than the results of the relevant benchmark. Managers tried to follow the trend rather than anticipate it. In most cases, there was also no ability to select assets or market-timing. Most of the parameters standing by these variables were not statistically significant.

Novel benchmark values for redo liver transplantation – does the outcome justify the effort?

Abstract Objective In the era of organ shortage, redo liver transplantation (reLT) is frequently discussed in terms of expected poor outcome, high cost and therefore wasteful resources. However, there is a lack of benchmark data to reliably assess outcomes after reLT. The aim of this study was to define the ideal reLT case, and to establish clinically relevant benchmark values for best achievable outcome in reLT. Methods We collected data on reLT between January 2010 and December 2018 from 22 high volume transplant centers on three continents. Benchmark cases were defined as recipients with model of end-stage liver disease score &amp;lt;=25, absence of portal vein thrombosis, no mechanical ventilation before surgery, receiving a graft from a donor after brain death. In addition, early reLT including those for primary non-function (PNF) were excluded. Clinically relevant endpoints covering intra- and postoperative course were selected and complications were graded by severity using the Clavien-Dindo classification and the comprehensive complication index (CCI). The benchmark cutoff for each outcome was derived from the 75th percentile of the median values of all benchmark centers, indicating the “best achievable” result. To assess the utility of the newly established benchmark values, we analyzed patients who received reLT for PNF (non-benchmark patients). Results Out of 1110 reLT 413 (37.2%) qualified as benchmark cases. Benchmark values included: Length of intensive care unit and hospital stay: &amp;lt;=6 and &amp;lt;=24 days, respectively; Clavien-Dindo grade &amp;gt;=3a complications and the CCI at 1 year: &amp;lt;=76% and &amp;lt;=72.2, respectively; in-hospital and 1-year mortality rates: &amp;lt;=14.0% and &amp;lt;=14.3%, respectively. The cutoffs for transplant-specific complications such as biliary complications at 1 year, outflow problems at 1 year and hepatic artery thrombosis at discharge were &amp;lt;=27.3%, &amp;lt;=2.5% and &amp;lt;=4.8%, respectively. Patients receiving a reLT for PNF showed mean outcome values all outside the reLT benchmark values. In-hospital mortality rate was 34.4% and the mean CCI at discharge 68.8. Conclusion ReLT remains associated with high morbidity and mortality. The availability of benchmark values for outcome parameters of reLT may serve for comparison in any future analyses of individuals, patient groups, or centers, but also in the evaluation of new therapeutic strategies and principles.

Connecting the Dots: Numerical Randomized Hamiltonian Monte Carlo with State-Dependent Event Rates

Numerical generalized randomized Hamiltonian Monte Carlo is introduced, as a robust, easy to use and computationally fast alternative to conventional Markov chain Monte Carlo methods for continuous target distributions. A wide class of piecewise deterministic Markov processes generalizing Randomized HMC (Bou-Rabee and Sanz-Serna) by allowing for state-dependent event rates is defined. Under very mild restrictions, such processes will have the desired target distribution as an invariant distribution. Second, the numerical implementation of such processes, based on adaptive numerical integration of second order ordinary differential equations (ODEs) is considered. The numerical implementation yields an approximate, yet highly robust algorithm that, unlike conventional Hamiltonian Monte Carlo, enables the exploitation of the complete Hamiltonian trajectories (hence, the title). The proposed algorithm may yield large speedups and improvements in stability relative to relevant benchmarks, while incurring numerical biases that are negligible relative to the overall Monte Carlo errors. Granted access to a high-quality ODE code, the proposed methodology is both easy to implement and use, even for highly challenging and high-dimensional target distributions. Supplementary materials for this article are available online.

A multi-quantile regression time series model with interquantile lipschitz regularization for wind power probabilistic forecasting

Modern decision-making processes require uncertainty-aware models, especially those relying on non-symmetric costs and risk-averse profiles. The objective of this work is to propose a dynamic model for the conditional non-parametric distribution function (CDF) to generate probabilistic forecasts for a renewable generation time series. To do that, we propose an adaptive non-parametric time-series model driven by a regularized multiple-quantile-regression (MQR) framework. In our approach, all regression models are jointly estimated through a single linear optimization problem that finds the global-optimal parameters in polynomial time. An innovative feature of our work is the consideration of a Lipschitz regularization of the first derivative of coefficients in the quantile space, which imposes coefficient smoothness. The proposed regularization induces a coupling effect among quantiles creating a single non-parametric CDF model with improved out-of-sample performance. A case study with realistic wind-power generation data from the Brazilian system shows: 1) the regularization model is capable to improve the performance of MQR probabilistic forecasts, and 2) our MQR model outperforms five relevant benchmarks: two based on the MQR framework, and three based on parametric models, namely, SARIMA, and GAS with Beta and Weibull CDF.

Supervised promoter recognition: a benchmark framework

MotivationDeep learning has become a prevalent method in identifying genomic regulatory sequences such as promoters. In a number of recent papers, the performance of deep learning models has continually been reported as an improvement over alternatives for sequence-based promoter recognition. However, the performance improvements in these models do not account for the different datasets that models are evaluated on. The lack of a consensus dataset and procedure for benchmarking purposes has made the comparison of each model’s true performance difficult to assess.ResultsWe present a framework called Supervised Promoter Recognition Framework (‘SUPR REF’) capable of streamlining the complete process of training, validating, testing, and comparing promoter recognition models in a systematic manner. SUPR REF includes the creation of biologically relevant benchmark datasets to be used in the evaluation process of deep learning promoter recognition models. We showcase this framework by comparing the models’ performances on alternative datasets, and properly evaluate previously published models on new benchmark datasets. Our results show that the reliability of deep learning ab initio promoter recognition models on eukaryotic genomic sequences is still not at a sufficient level, as overall performance is still low. These results originate from a subset of promoters, the well-known RNA Polymerase II core promoters. Furthermore, given the observational nature of these data, cross-validation results from small promoter datasets need to be interpreted with caution.

121P Treatment of unresectable stage III NSCLC: Real-world study and literature review

The standard treatment of unresectable stage III non-small cell lung cancer (NSCLC) is combined chemoradiotherapy (CRT), delivered either concurrently (cCRT) or sequentially (sCRT). There is, however, limited data on the outcomes and safety of CRT in real-world setting. We performed an analysis of the Leuven Lung Cancer Group (LLCG) experience with CRT for unresectable stage III NSCLC prior to the era of consolidation treatment with immunotherapy. We conducted a retrospective, monocentric cohort study at the university hospital UZ Leuven, Belgium, and retrieved 163 consecutive patients who were treated with CRT for unresectable stage III NSCLC between January 1st 2011 and December 31st 2018. Patient and tumor characteristics, treatment patterns, toxicity, and primary outcome parameters were analyzed. Chemotherapy was given concurrently in 108 patients, sequentially in 55. Overall tolerability was good with two thirds of patients having no severe adverse events: severe febrile neutropenia (9,2%), ³ grade 2 pneumonitis (16%), or ³ grade 3 esophagitis (19%). Adverse events were more frequent in the cCRT than sCRT group. Median PFS was 13.2 months (95% CI 10.3-16.2). In the patients with progressive disease, the site of first progression was locoregional in 36% and distant in 64%, with no significant difference between cCRT and sCRT. Median OS was 23.3 months (95% CI 18.3-28.0), with a 47.5% survival rate at 2 years and 29.4% at 5 years. In multivariate Cox regression analysis, Karnofsky performance status, stage (UICC TNM 8th edition) and COPD were significant factors for OS, while sequence of therapy (cCRT versus sCRT) lost significance in multivariate testing. Our real-world PFS, OS and toxicity results are comparable to those reported in randomized controlled trials (RTOG 0617, PROCLAIM). Sequence of therapy did not independently associate with OS, likely since the choice for a sequential versus concurrent approach is strongly influenced by other main drivers of survival such as performance status. This study provides a clinically relevant real-world benchmark on the outcomes of cCRT and sCRT before the start of the immunotherapy era in this setting.

Real-world burden of adverse events for apalutamide- or enzalutamide-treated non-metastatic castration-resistant prostate cancer patients in the United States

BackgroundSecond-generation androgen receptor inhibitors (ARIs) have been associated with adverse events (AEs) such as fatigue, falls, fractures, and rash in non-metastatic castration-resistant prostate cancer (nmCRPC) patients as identified in clinical trials. The objectives of this study were to describe the incidence and management of AEs in patients receiving apalutamide and enzalutamide.MethodsThis retrospective chart review study was conducted in nmCRPC-treating sites in the United States. Patients starting apalutamide or enzalutamide between February 1, 2018 and December 31, 2018 were included and any AEs they experienced were recorded. AEs, including those considered to be of special interest as defined in the pivotal clinical trials of the second-generation ARIs, were analyzed and grouped retrospectively in this study. Detailed chart data (patient demographics, clinical characteristics, treatment history, type of AE, outcomes, and resource utilization) were then collected for a randomly selected subset among patients with ≥1 AE to characterize AEs and their management. Descriptive results were summarized.ResultsForty-three sites participated in the study. A total of 699 patients were included, of whom 525 (75.1%) experienced ≥1 AE. The most common AEs were fatigue/asthenia (34.3%), hot flush (13.9%), and arthralgia (13.6%). In the subset of 250 patients randomly selected from those who experienced ≥1 AE, patients were primarily White (72.0%), the mean age was 71 years, 86.0% had an Eastern Cooperative Oncology Group score of 0–1 at nmCRPC diagnosis, and the average prostate specific antigen (PSA) value at diagnosis was 23.2 ng/mL. PSA-doubling time < 10 months was chosen as reason to initiate treatment in 40% of patients. The median duration of follow-up was 1.1 years, with 14.4% of patients progressing to metastasis by end of study period. Grade 3–4 and Grade 5 AEs occurred in 14.4 and 0.4% of patients, respectively. Actions taken to manage AEs included AE-directed treatment (38.0%), ARI discontinuation (10.4%), dose reduction (7.6%), and AE-related hospitalization (4.8%).ConclusionsThis study highlights the burden of AEs among nmCRPC patients treated with apalutamide or enzalutamide, providing a relevant real-world benchmark as clinical trial evidence and the treatment landcape for nmCRPC continues to evolve.

ABE Squared: Accurately Benchmarking Efficiency of Attribute-Based Encryption

Measuring efficiency is difficult. In the last decades, several works have contributed in the quest to successfully determine and compare the efficiency of pairing-based attribute-based encryption (ABE) schemes. However, many of these works are limited: they use little to no optimizations, or use underlying pairingfriendly elliptic curves that do not provide sufficient security anymore. Hence, using these works to benchmark ABE schemes does not yield accurate results. Furthermore, most ABE design papers focus on the efficiency of one important aspect. For instance, a new scheme may aim to have a fast decryption algorithm. Upon realizing this goal, the designer compares the new scheme with existing ones, demonstrating its dominance in this particular aspect. Although this approach is intuitive and might seem fair, the way in which this comparison is done might be biased. For instance, the schemes that are compared with the new scheme may be optimized with respect to another aspect, and appear in the comparison consequently inferior.In this work, we present a framework for accurately benchmarking efficiency of ABE: ABE Squared. In particular, we focus on uncovering the multiple layers of optimization that are relevant to the implementation of ABE schemes. Moreover, we focus on making any comparison fairer by considering the influence of the potential design goals on any optimizations. On the lowest layer, we consider the available optimized arithmetic provided by state-of-the-art cryptographic libraries. On the higher layers, we consider the choice of elliptic curve, the order of the computations, and importantly, the instantiation of the scheme on the chosen curves. Additionally, we show that especially the higher-level optimizations are dependent on the goal of the designer, e.g. optimization of the decryption algorithm. To compare schemes more transparently, we develop this framework, in which ABE schemes can be justifiably optimized and compared by taking into account the possible goals of a designer. To meet these goals, we also introduce manual, heuristic type-conversion techniques where existing techniques fall short. Finally, to illustrate the effectiveness of ABE Squared, we implement several schemes and provide all relevant benchmarks. These show that the design goal influences the optimization approaches, which in turn influence the overall efficiency of the implementations. Importantly, these demonstrate that the schemes also compare differently than existing works previously suggested.

Empirically-Derived Effect Size Distributions of Interventions for Young Children on the Autism Spectrum

ABSTRACT Objective The purpose of this study is to present a set of empirically derived effect size distributions to provide field-based benchmarks for interpreting the observed effects of interventions for young children on the autism spectrum, and for planning future studies. Method We generated effect size distributions and reported quartile values for each by outcome domain, and by boundedness, proximity, and assessment approach using 1552 effect sizes from 144 early childhood autism intervention studies gathered for a previously published meta-analysis. Results Quartile values represent considerable heterogeneity in effect size distributions across outcome domains, as well as variability as a function of outcome boundedness, proximity, and assessment approach. Conclusions Our results serve as field- and outcome-specific benchmarks (e.g., contextual guides for small, medium, and large effects) that will help autism intervention researchers easily incorporate information from relevant prior empirical literature when conducting power analyses to plan for future studies. Benchmarks will also assist researchers seeking to interpret the magnitude of observed effects in clinical trials relative to the broader distribution of intervention effects on similar outcomes. Nuanced discussions that contextualize study findings in light of relevant empirical benchmarks will better assist practitioners in understanding the magnitude and scope of demonstrated change relative to studies with similar outcomes and selecting interventions for clinical practice. We discuss the limitations of these data, our analyses, as well as directions for future work.

Xor-And-Inverter Graphs for Quantum Compilation

Quantum compilation is the task of translating a high-level description of a quantum algorithm into a sequence of low-level quantum operations. We propose and motivate the use of Xor-And-Inverter Graphs (XAG) to specify Boolean functions for quantum compilation. We present three different XAG-based compilation algorithms to synthesize quantum circuits in the Clifford + T library, hence targeting fault-tolerant quantum computing. The algorithms are designed to minimize relevant cost functions, such as the number of qubits, the T-count, and the T-depth, while allowing the flexibility of exploring different solutions. We present novel resource estimation results for relevant cryptographic and arithmetic benchmarks. The achieved results show a significant reduction in both T-count and T-depth when compared with the state-of-the-art.