Optimal Trade-off Research Articles

Robust Privatization With Multiple Tasks and the Optimal Privacy-Utility Tradeoff

Robust Privatization With Multiple Tasks and the Optimal Privacy-Utility Tradeoff

Determining the Optimal Tradeoff between Compute and Accuracy for Diffusion Models in Molecular Docking

In the world of molecular biology, Molecular Docking has become a pillar for understanding complex interactions pivotal for drug design and more. The emergence of sophisticated machine learning models, such as Diffusion Generative Models, have significantly expanded our analytical capabilities. However, this evolution has introduced notable computational challenges. This study aims to examine the trade-off between computational demands and model accuracy. Our methodology, which incorporates free platforms, illuminates methods to conserve computational resources while maintaining nearoptimal accuracy. Our findings suggest that using 30 samples per complex, 15 inference steps, and 4 batch steps improves pose prediction accuracy and reduces computational resources. The proposed parameters achieve a 14% accuracy increase compared to the 40 samples per complex model and a 56.25% increase compared to the 10 samples per complex model. The optimized inference steps result in a 12.2% accuracy increase over the 20-step control using the 40 samples model and a 24.3% increase using the 10 samples model. Additionally, using 4 batch steps leads to a 40.6% increase in DiffDock Confidence for the 10-sample control and a 0.4% increase for the 40- sample control.

Exploring the Drive Motor of Electric Vehicles: Structure, Temperature Rises, and Operational Control of Permanent Magnet Motors

The drive motor is a critical component of electric vehicles. The design of its structure, along with the regulation of temperature rises and operational control, significantly impact the overall performance of electric vehicles, affecting aspects such as dynamic response, anti-interference capabilities, speed stability, and efficiency. This paper provides a review and comprehensive analysis of the Special Issue titled “Temperature Field, Electromagnetic Field, and Operation Control of Permanent Magnet Motor for Electric Vehicles”, and discusses the future development directions of permanent magnet motors in electric vehicles. The analysis reveals that the structural design is crucial for both the static and dynamic performance of permanent magnet motors. Additionally, it is imperative to identify an optimal trade-off among various performance parameters in these motors. The paper also verifies several future development directions for permanent magnet motors, including high-speed control, the use of advanced silicon steel sheets, power quality, battery charging efficiency, and energy feedback.

Pore-Scale multi-objective shape optimization of triply periodic minimal surface cellular architectures: Volumetric Nusselt number versus friction factor

Triply Periodic Minimal Surface (TPMS) cellular architectures, i.e., minimal surfaces that are periodic in three independent directions, have attracted interest in the engineering community for their potential in heat transfer applications with severe volume constraints due to their high surface-to-volume ratio, unique geometrical properties and smooth porous structure. Such performance can be conveniently controlled by tuning the TPMS unit type, relative density, and structural parameters. In this study, a multi-objective optimization process is carried out to investigate how pore-scale geometric characteristics affect the heat transfer performance of TPMS, and which combinations result in an optimal trade-off between heat dissipation capability and associated pressure drop in terms of the Nusselt number and friction factor. The analysis is performed using an in-house Genetic Algorithm routine that generates the structure, allowing for a wider and denser range of investigation, and incorporates a commercial finite element CFD software as a black box fitness function. Over 14,000 unique combinations between two heat transfer boundary conditions (constant wall heat flux, fixed wall temperature) and several flow inlet conditions are investigated. The optimization achieved an increase in performance of up to 245 % in terms of volumetric Nusselt number, which was accompanied, however, by a significant, although reasonable, increase in friction factor values. The results also show specific trends among the geometric characteristics studied, which are then contextualized through the use of CFD analysis. Finally, empirical correlations are derived to cluster the non-dominated solutions obtained, enabling the reader to choose the best cellular TPMS structure by selecting the desired level of trade-off between convection heat transfer and flow resistance.

Multi-objective optimization design of automobile seat backrest considering coupling effect

To investigate the impact of the coupling effects of carbon fiber reinforced polymer in the seat back layer on the performance of car seats, this paper presents a comprehensive optimization design method for composite materials. In detail, the finite element models firstly established and validated through five typical working conditions of automotive seats based on experimental data. Then, the optimized variables are divided and determined through backrest stress nephograms for each working conditions of the automotive seats, in which the various perspective are taken into account, such as the total mass of seat backrest, safety performance, and comfort index. Subsequently, an optimization strategy for unequal thickness layers lay-up design is constructed, which combines strength factors, optimal Latin hypercube sampling, best-worst method, gray relational analysis, and Visekriterijumsko KOmpromisno Rangiranje method for the optimal design of the automotive CFRP seat backrest. Additionally, the impact of layer coupling effects on different performance indices of the seat is examined through simulating and analyzing the seat backrest with various layer coupling types, while incorporating the classical laminate theory. The study reveals that by minimizing the laminate coupling effect, the comfort of the seat can be enhanced. Finally, a comprehensive comparative analysis of the optimal trade-off solution is carried out in terms of optimization strategies. The results show that ensuring the safety performance, the total mass of the seat backrest decreased by 21.3%, as a result of the optimization strategy proposed in this paper, and the comfort performance is also improved to some extent. Therefore, the multi-objective optimization strategy proposed in this paper performs well in terms of effectiveness and provides a reliable reference for related composite material multi-objective optimization.

Comparative and optimal study of energy, exergy, and exergoeconomic performance in supercritical CO2 recompression combined cycles with organic rankine, trans-critical CO2, and kalina cycle

The bottom cycle in supercritical carbon dioxide (sCO2) recompression Brayton combined cycle (RCBC) effectively recovers substantial waste heat for electricity generation, enhancing energy utilization. The recovery efficiency is significantly influenced by the bottom cycle configurations and specific working conditions. Besides, implementing these bottom cycles introduces challenges related to increased system dimensions and design complexity. The present study aims to understand and evaluate the characteristics and optimal trade-offs of various bottom cycles, including trans-critical, subcritical, and non-azeotropic mixed working fluids, by establishing 3E (energy, exergy and Exergoeconomic) models for sCO2/tCO2, sCO2/ORC, and sCO2/KC combined cycle systems. To analysis and enhance interpretability, statistical Global Sensitivity Analysis (GSA) alongside unsupervised learning-based Self-Organizing Map (SOM) techniques and parametric analysis are employed. These methods identify key parameters and discern system patterns. Subsequently, the Non-Dominated Sorting Whale Optimization Algorithm (NSWOA) and entropy weight-based Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) are applied to establish a Pareto-optimal decision space and identify compromised ideal design points for each combined cycle. The results quantitatively rank sensitivity for objective variables within the design space in different combined cycles and visually represent system nonlinearity and coupling relationships in 2D weight planes using SOM. The TOPSIS trade-off points based on the Pareto frontier for the three combined cycles are ηth = 45.08 %, cp,tot = 8.5199 $/GJ, ηth = 45.13 %, cp,tot = 8.3739 $/GJ, and ηth = 48.97 %, cp,tot = 8.4783 $/GJ, respectively. Integrating machine learning techniques provides a comprehensive understanding of system patterns, offering valuable insights for decision-makers in the design and optimization of combined cogeneration cycles.

Achieving the Exactly Optimal Privacy-Utility Trade-Off With Low Communication Cost via Shared Randomness

Achieving the Exactly Optimal Privacy-Utility Trade-Off With Low Communication Cost via Shared Randomness

Neighbor patches merging reduces spatial redundancy to accelerate vision transformer

Vision Transformers (ViTs) deliver outstanding performance but often require substantial computational resources. Various token pruning methods have been developed to enhance throughput by removing redundant tokens; however, these methods do not address the peak memory consumption, which remains equivalent to that of the unpruned networks. In this study, we introduce Neighbor Patches Merging (NEPAM), a method that significantly reduces the maximum memory footprint of ViTs while pruning tokens. NEPAM targets spatial redundancy within images and prunes redundant patches at the onset of the model, thereby achieving the optimal throughput-accuracy trade-off without fine-tuning. Experimental results demonstrate that NEPAM can accelerate the inference speed of the Vit-Base-Patch16-384 model by 25% with a negligible accuracy loss of 0.07% and a notable 18% reduction in memory usage. When applied to VideoMAE, NEPAM doubles the throughput with a 0.29% accuracy loss and a 48% reduction in memory usage. These findings underscore the efficacy of NEPAM in mitigating computational requirements while maintaining model performance.

Accurate preference-based method to obtain the deterministically optimal and satisfactory fairness-efficiency trade-off

The resource allocation problem is a classic multi-objective challenge, particularly when balancing the fairness-efficiency trade-off. To achieve a deterministically optimal and satisfactory solution, researchers frequently employ preference-based methods, including selecting among Pareto solutions based on the decision-maker's a posteriori preference and using deterministic models incorporating a priori preferences. In this study, we address two main challenges—specifically, (1) the limitations in measuring the abstract concepts of fairness and efficiency and (2) finding a deterministically optimal and satisfactory balance between fairness and efficiency. We apply a Gini impurity index derived from the classification and regression tree to calculate fairness, ensuring the Gini index function's differentiability. Additionally, we unify the scales of fairness and efficiency to facilitate calculation. Using accurate preference information, we employ the extended interval goal programming method to solve the model and achieve a deterministically optimal and satisfactory solution. The comparative analysis results demonstrate that our model (1) efficiently addresses the real-world water resource allocation problem concerning the fairness-efficiency trade-off; and (2) generates fewer penalties, with an average improvement ratio of 8% in the case study, using more refined penalty functions that align closer to the decision-maker's real and nonlinear preferences.

Can input reconstruction be used to directly estimate uncertainty of a dose prediction U-Net model?

The reliable and efficient estimation of uncertainty in artificial intelligence (AI) models poses an ongoing challenge in many fields such as radiation therapy. AI models are intended to automate manual steps involved in the treatment planning workflow. We focus in this study on dose prediction models that predict an optimal dose trade-off for each new patient for a specific treatment modality. They can guide physicians in the optimization, be part of automatic treatment plan generation or support decision in treatmentindication. Most common uncertainty estimation methods are based on Bayesian approximations, like Monte Carlo dropout (MCDO) or Deep ensembling (DE). These two techniques, however, have a high inference time (i.e., require multiple inference passes) and might not work for detecting out-of-distribution (OOD) data (i.e., overlapping uncertainty estimate for in-distribution (ID) and OOD). In this study, we present a direct uncertainty estimation method and apply it for a dose prediction U-Net architecture. It can be used to flag OOD data and give information on the quality of the doseprediction. Our method consists in the addition of a branch decoding from the bottleneck which reconstructs the CT scan given as input. The input reconstruction error can be used as a surrogate of the model uncertainty. For the proof-of-concept, our method is applied to proton therapy dose prediction in head and neck cancer patients. A dataset of 60 oropharyngeal patients was used to train the network using a nested cross-validation approach with 11 folds (training: 50 patients, validation: 5 patients, test: 5 patients). For the OOD experiment, we used 10 extra patients with a different head and neck sub-location. Accuracy, time-gain, and OOD detection are analyzed for our method in this particular application and compared with the popular MCDO andDE. The additional branch did not reduce the accuracy of the dose prediction model. The median absolute error is close to zero for the target volumes and less than 1% of the dose prescription for organs at risk. Our input reconstruction method showed a higher Pearson correlation coefficient with the prediction error (0.620) than DE (0.447) and MCDO (between 0.599 and 0.612). Moreover, our method allows an easier identification of OOD (no overlap for ID and OOD data and a Z-score of 34.05). The uncertainty is estimated simultaneously to the regression task, therefore requires less time and computationalresources. This study shows that the error in the CT scan reconstruction can be used as a surrogate of the uncertainty of the model. The Pearson correlation coefficient with the dose prediction error is slightly higher than state-of-the-art techniques. OOD data can be more easily detected and the uncertainty metric is computed simultaneously to the regression task, therefore faster than MCDO or DE. The code and pretrained model are available on the gitlab repository: https://gitlab.com/ai4miro/ct-reconstruction-for-uncertainty-quatification-of-hdunet.

Slow but flexible or fast but rigid? Discrete and continuous processes compared

A tradeoff exists when dealing with complex tasks composed of multiple steps. High-level cognitive processes can find the best sequence of actions to achieve a goal in uncertain environments, but they are slow and require significant computational demand. In contrast, lower-level processing allows reacting to environmental stimuli rapidly, but with limited capacity to determine optimal actions or to replan when expectations are not met. Through reiteration of the same task, biological organisms find the optimal tradeoff: from action primitives, composite trajectories gradually emerge by creating task-specific neural structures. The two frameworks of active inference – a recent brain paradigm that views action and perception as subject to the same free energy minimization imperative – well capture high-level and low-level processes of human behavior, but how task specialization occurs in these terms is still unclear. In this study, we compare two strategies on a dynamic pick-and-place task: a hybrid (discrete-continuous) model with planning capabilities and a continuous-only model with fixed transitions. Both models rely on a hierarchical (intrinsic and extrinsic) structure, well suited for defining reaching and grasping movements, respectively. Our results show that continuous-only models perform better and with minimal resource expenditure but at the cost of less flexibility. Finally, we propose how discrete actions might lead to continuous attractors and compare the two frameworks with different motor learning phases, laying the foundations for further studies on bio-inspired task adaptation.

A Novel Control Strategy for Optimal Tradeoff between Overshoot and Switching Loss Based on Double Closed-Loop Self-Regulating Active Gate Driver

A Novel Control Strategy for Optimal Tradeoff between Overshoot and Switching Loss Based on Double Closed-Loop Self-Regulating Active Gate Driver

Multiobjective optimal TCSC placement using multiobjective grey wolf optimizer for power losses reduction

This study investigates the application of the multiobjective grey wolf optimizer (MOGWO) for optimal placement of thyristor-controlled series compensator (TCSC) to minimize power loss in power systems. Two conflicting objectives are considered: (1) minimizing real and reactive power loss, and (2) minimizing real power loss and TCSC capital cost. The Pareto-optimal method is employed to generate the Pareto front for these objectives. The fuzzy set technique is used to identify the optimal trade-off solution, while the technique for order preference by similarity to the ideal solution suggests multiple optimal solutions catering to diverse utility preferences. Simulations on an IEEE 30 bus test system demonstrate the effectiveness of TCSC placement for power loss minimization using MOGWO. The superiority of MOGWO is confirmed by comparing its results with those obtained from a multiobjective particle swarm optimization algorithm. These findings can assist power system utilities in identifying optimal TCSC locations to maximize their performance.

Quantitative Portfolio Management: Review and Outlook

This survey aims to provide insightful and objective perspectives on the research history of quantitative portfolio management strategies with suggestions for the future of research. The relevant literature can be clustered into four broad themes: portfolio optimization, risk-parity, style integration, and machine learning. Portfolio optimization attempts to find the optimal trade-off of future returns per unit of risk. Risk-parity attempts to match the exposure of various asset classes such that no single asset class dominates portfolio risk. Style integration combines risk factors on a security level such that rebalancing differences cancel out. Finally, machine learning utilizes large arrays of tunable parameters to predict future asset behavior and solve non-convex optimization problems. We conclude that machine learning will likely be the focus of future research.

Optimal performance objectives in the highly conserved bone morphogenetic protein signaling pathway

Throughout development, complex networks of cell signaling pathways drive cellular decision-making across different tissues and contexts. The transforming growth factor β (TGF-β) pathways, including the BMP/Smad pathway, play crucial roles in determining cellular responses. However, as the Smad pathway is used reiteratively throughout the life cycle of all animals, its systems-level behavior varies from one context to another, despite the pathway connectivity remaining nearly constant. For instance, some cellular systems require a rapid response, while others require high noise filtering. In this paper, we examine how the BMP-Smad pathway balances trade-offs among three such systems-level behaviors, or “Performance Objectives (POs)”: response speed, noise amplification, and the sensitivity of pathway output to receptor input. Using a Smad pathway model fit to human cell data, we show that varying non-conserved parameters (NCPs) such as protein concentrations, the Smad pathway can be tuned to emphasize any of the three POs and that the concentration of nuclear phosphatase has the greatest effect on tuning the POs. However, due to competition among the POs, the pathway cannot simultaneously optimize all three, but at best must balance trade-offs among the POs. We applied the multi-objective optimization concept of the Pareto Front, a widely used concept in economics to identify optimal trade-offs among various requirements. We show that the BMP pathway efficiently balances competing POs across species and is largely Pareto optimal. Our findings reveal that varying the concentration of NCPs allows the Smad signaling pathway to generate a diverse range of POs. This insight identifies how signaling pathways can be optimally tuned for each context.

Novel Power-Efficient Fast-Locking Phase-Locked Loop Based on Adaptive Time-to-Digital Converter-Aided Acceleration Compensation Technology

This paper proposes an adaptive acceleration lock compensation technology for phase-locked loops (PLLs) based on a novel dual-mode programmable ring voltage-controlled oscillator (ring-VCO). In addition, a time-to-digital converter (TDC) is designed to accurately quantify the phase difference from the phase frequency detector (PFD) in order to optimize the dead-zone effect while dynamically switching an auxiliary charge pump (CP) module to realize fast phase locking. Furthermore, a TDC-controlled three/five-stage dual-mode adaptively continuously switched VCO is proposed to optimize the phase noise (PN) and power efficiency, leading to an optimal performance tradeoff of the PLL. Based on the 180 nm/1.8 V standard CMOS technology, the complete PLL design and a corresponding simulation analysis are implemented. The results show that, with a 1 GHz reference signal as the input, the output frequency is 50–324 MHz, with a wide tuning range of 260 MHz and a low phase noise of −98.07 dBc/Hz@1 MHz. The key phase-locking time is reduced to 1.11 μs, and the power dissipation is lowered to 1.86 mW with a layout area of 66 μm × 128 μm. A significantly remarkable multiobjective performance tradeoff with topology optimization is realized, which is in contrast to several similar design cases of PLLs.

Unidimensional community detection: A monte carlo simulation, grid search, and comparison.

Unidimensionality is fundamental to psychometrics. Despite the recent focus on dimensionality assessment in network psychometrics, unidimensionality assessment remains a challenge. Community detection algorithms are the most common approach to estimate dimensionality in networks. Many community detection algorithms maximize an objective criterion called modularity. A limitation of modularity is that it penalizes unidimensional structures in networks, favoring two or more communities (dimensions). In this study, this penalization is discussed and a solution is offered. Then, a Monte Carlo simulation using one- and two-factor models is performed. Key to the simulation was the condition of model error or the misfit of the population factor model to the generated data. Based on previous simulation studies, several community detection algorithms that have performed well with unidimensional structures (Leading Eigenvalue, Leiden, Louvain, and Walktrap) were compared. A grid search was performed on the tunable parameters of these algorithms to determine the optimal trade-off between unidimensional and bidimensional recovery. The best-performing parameters for each algorithm were then compared against each other as well as maximum likelihood factor analysis and parallel analysis (PA) with mean and 95th percentile eigenvalues. Overall, the Leiden and Louvain algorithms and PA methods were the most accurate methods to recover unidimensional and bidimensional structures and were the most robust to model error. More nuanced method recommendations for specific unidimensional and bidimensional conditions are provided. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Potential implications of using locally validated risk factors for drug-resistant pathogens in patients with community-acquired pneumonia in US hospitals: A cross-sectional study.

The 2019 ATS/IDSA community-acquired pneumonia (CAP) guidelines recommend that clinicians prescribe empiric antibiotics for MRSA or P. aeruginosa only if locally validated risk factors (or 2 generic risk factors if local validation is not feasible) are present. It remains unknown how implementation of this recommendation would influence care. This cross-sectional study included adults hospitalized for CAP across 50 hospitals in the Premier Healthcare Database from 2010-2015 and sought to describe how the use of extended-spectrum antibiotics (ESA) and the coverage for patients with CAP due to restraint organisms would change under the two approaches described in 2019 ATS/IDSA guidelines. To do this, the proportion of ESA use in patients with CAP and the proportion of ESA coverage among patients with infections resistant to recommended CAP therapy were measured. In the 50 hospitals, 19%-75% of patients received ESA, and 42%-100% of patients with resistant organisms received ESA. The median number of risk factors identified per hospital was 9 (interquartile range [IQR], 6-12). Overall, treatment according to local risk factors reduced the number of patients receiving ESA by 38.8 percentage points and using generic risk factors by 47.5 percentage points. However, the effect varied by hospital. The use of generic risk factors always resulted in less ESA use and less coverage for resistant organisms. Using locally validated risk factors resulted in a similar outcome in all but one hospital. Future guidelines should explicitly define the optimal trade-off between adequate coverage for resistant organisms and ESA use.

An optimization framework for achieving optimal hydrocyclone's performance aligning with decision-makers' preferences

Previous hydrocyclone optimization often neglected interactions among key performance objectives, which limits hydrocyclone wide applications to meet increasingly diverse industry demands. This study proposes an optimization framework to identify the most suitable hydrocyclone design and operating conditions with conflicting key performance objectives. An advanced multi-objective evolutionary algorithm (PICEA-g) is employed to generate Pareto-optimal solutions that capture trade-offs among multiple conflicting objectives. A novel data-driven predictive algorithm, INFO-ELM, is introduced to establish nonlinear relationships between key variables and performance objectives, thereby accelerating the search for Pareto-optimal solutions by PICEA-g. Furthermore, a multi-criteria decision-making method (TOPSIS) is utilized to determine the optimal solution based on decision-makers' preferences, ensuring consistency between preference information and decision outcomes. The framework's effectiveness is validated across various separation scenarios using two decision-making strategies. This study offers a comprehensive approach to address trade-offs in hydrocyclone optimization, widening its application in diverse separation scenarios.

Optimal cost-emission trade-off for plug-in hybrid electric vehicles around zero emission zones using a supervisory energy and emissions management strategy

Optimal cost-emission trade-off for plug-in hybrid electric vehicles around zero emission zones using a supervisory energy and emissions management strategy