Efficiency Tradeoff Research Articles

Efficient and lightweight layer-wise in-situ defect detection in laser powder bed fusion via knowledge distillation and structural re-parameterization

Powder bed quality is critical in Laser Powder Bed Fusion (LPBF) as defects in the powder bed will affect the quality of the forming parts. Recently, numerous powder bed defect detection methods based on semantic segmentation algorithms have been developed. However, due to the computational resource constraints and real-time requirements in industry scenarios, these heavy models face challenges in deploying into the LPBF process monitoring system. Meanwhile, existing lightweight models struggle to achieve sufficient performance on industrial data due to their limited feature extraction capabilities. To address the challenges in industrial deployment, this paper proposes a lightweight and efficient LPBF layer-wise defect detection paradigm to achieve a balance between performance and efficiency. Aiming at improving the performance of the lightweight model, this paper proposes a structured Knowledge Distillation framework, in which a pre-trained high-performance segmentation model (UNet++) is used to guide the training process of a lightweight model. To address the impact of model discrepancies on the effectiveness of Knowledge Distillation and further improve the inference speed, this paper constructs a specific lightweight model and applies Structural Re-parameterization to deeply compress the model. Moreover, we created a dataset comprising 406 images of powder-bed defects, with each image annotated at the pixel level. Massive experiments based on the dataset demonstrate the effectiveness of the proposed paradigm, with an excellent trade-off in performance and efficiency. The speed of the proposed model is approximately 5 times faster than that of high-performance models while maintaining comparable performance.

No immediate attentional bias towards or choice bias for male secondary sexual characteristics in Bornean orang-utans (Pongo pygmaeus).

Primate faces provide information about a range of variant and invariant traits, including some that are relevant for mate choice. For example, faces of males may convey information about their health or genetic quality through symmetry or facial masculinity. Because perceiving and processing such information may have bearing on the reproductive success of an individual, cognitive systems are expected to be sensitive to facial cues of mate quality. However, few studies have investigated this topic in non-human primate species. Orang-utans are an interesting species to test mate-relevant cognitive biases, because they are characterised by male bimaturism: some adult males are fully developed and bear conspicuous flanges on the side of their face, while other males look relatively similar to females. Here, we describe two non-invasive computerised experiments with Bornean orang-utans (Pongo pygmaeus), testing (i) immediate attention towards large flanges and symmetrical faces using a dot-probe task (N = 3 individuals; 2F) and (ii) choice bias for pictures of flanged males over unflanged males using a preference test (N = 6 individuals; 4F). In contrast with our expectations, we found no immediate attentional bias towards either large flanges or symmetrical faces. In addition, individuals did not show a choice bias for stimuli of flanged males. We did find exploratory evidence for a colour bias and energy efficiency trade-offs in the preference task. We discuss our null results and exploratory results in the context of the evolutionary history of Bornean orang-utans, and provide suggestions for a more biocentric approach to the study oforang-utan cognition.

Effects of higher education subsidies on equity and efficiency across developmental stages

This study investigates the impact of taxes and subsidies for higher education on equity–efficiency trade-offs under various phases of higher education development. A scholarship program emerges as the most effective subsidy for higher education across all developmental stages. During both early and later phases, a universal grant proves more effective than a scholarship grant in reducing inequality. At the later stages of higher education development, the enrollment rate increases for universal grants but decreases for other policies, implying that the recent shift away from universal grant schemes in the UK could lead the enrollment rate to decline.

Testing the LSST Difference Image Analysis Pipeline Using Synthetic Source Injection Analysis

We evaluate the performance of the Legacy Survey of Space and Time Science Pipelines Difference Image Analysis (DIA) on simulated images. By adding synthetic sources to galaxies on images, we trace the recovery of injected synthetic sources to evaluate the pipeline on images from the Dark Energy Science Collaboration Data Challenge 2. The pipeline performs well, with efficiency and flux accuracy consistent with the signal-to-noise ratio of the input images. We explore different spatial degrees of freedom for the Alard–Lupton polynomial-Gaussian image subtraction kernel and analyze for trade-offs in efficiency versus artifact rate. Increasing the kernel spatial degrees of freedom reduces the artifact rate without loss of efficiency. The flux measurements with different kernel spatial degrees of freedom are consistent. We also here provide a set of DIA flags that substantially filter out artifacts from the DIA source table. We explore the morphology and possible origins of the observed remaining subtraction artifacts and suggest that given the complexity of these artifact origins, a convolution kernel with a set of flexible bases with spatial variation may be needed to yield further improvements.

Trade-Off in Energy Efficiency and Response Performance in Green CRNs: Perspective of Two Sleep Timers and Bursty Traffic

Trade-Off in Energy Efficiency and Response Performance in Green CRNs: Perspective of Two Sleep Timers and Bursty Traffic

UAV-Enabled Federated Learning in Dynamic Environments: Efficiency and Security Trade-Off

UAV-Enabled Federated Learning in Dynamic Environments: Efficiency and Security Trade-Off

Intercomparisons of Tracker v1.1 and four other ocean particle-tracking software packages in the Regional Ocean Modeling System

Abstract. Particle tracking is widely utilized to study transport features in a range of physical, chemical, and biological processes in oceanography. In this study, a new offline particle-tracking package, Tracker v1.1, is introduced, and its performance is evaluated in comparison to an online Eulerian dye, one online particle-tracking software package, and three offline particle-tracking software packages in a small, high-resolution model domain and a large coarser model domain. It was found that both particle and dye approaches give similar results across different model resolutions and domains when they were tracking the same water mass, as indicated by similar mean advection pathways and spatial distributions of dye and particles. The flexibility of offline particle tracking and its similarity against online dye and online particle tracking make it a useful tool to complement existing ocean circulation models. The new Tracker was shown to be a reliable particle-tracking package to complement the Regional Ocean Modeling System (ROMS) with the advantages of platform independence and speed improvements, especially in large model domains achieved by the nearest-neighbor search algorithm. Lastly, trade-offs of computational efficiency, modifiability, and ease of use that can influence the choice of which package to use are explored. The main value of the present study is that the different particle and dye tracking codes were all run on the same model output or within the model that generated the output. This allows some measure of intercomparison between the different tracking schemes, and we conclude that all choices that make each tracking package unique do not necessarily lead to very different results.

Possible Challenges and Recommendations for Risk Managers in Financial Institutions

The article analyses the importance and potential challenges of risk management in financial institutions, highlights the difficulties financial institutions face in identifying, assessing and mitigating the various risks within their organisations, and makes a number of recommendations to The article analyses the importance and potential challenges of risk management in financial institutions, highlights the difficulties financial institutions face in identifying, assessing and mitigating the various risks within their organisations, and makes a number of recommendations to address these issues.This paper explores these challenges, which include spillover effects, cybersecurity issues, efficiency trade-offs between costs and processes, regulatory compliance, and the impact of globalisation. Recommendations are made to address these challenges, such as adopting standardised procedures, embracing technological innovation, identifying emerging risks, and increasing automation of risk management processes. Effective risk management practices are essential to address the complexity of the financial industry and ensure organisational resilience.

Continuous Organizational Agility in a Highly Regulated Industry

Unlike those in other industries, highly regulated entities, such as banks, are required to be agile to business uncertainties and unforeseen risks while complying with the growing regulations. Moreover, banks pose constant challenges in the form of digitization, globalization, automation, analytics, and other forces of change that will continue to accelerate; thus, implementing a short-term agile transformation strategy will burden the company with inefficiencies. Expanding organizational and strategic management theory for the conceptualization, continuous organizational agility will follow as the manifestation of the intertwined routine dynamics-dynamic capabilities-organizational agility relationship. This study reviewed, synthesized, extended the literature, and proposed the research idea based on the relationship between routine dynamics, dynamic capabilities, and their impact on organizational agility in an integrated model. Generally, the proposed model aims to clarify: (1) the ability of an agile company to maintain stability while being dynamic and (2) the development of a more efficient trade-off between agility and efficiency during the agile transformation process. This study may provide fruitful avenues for enriching literature and future study since a dearth of empirical evidence addresses the relationship.

Medium spatiotemporal characteristics based global optimization method for energy efficiency trade-off issue in variable flow rate HVAC system

During the optimization and control processes of variable flow rate heating, ventilation, and air conditioning (HVAC) systems, the medium temporal and spatial characteristics often significantly influence the control effectiveness and energy efficiency of the overall system. Unfortunately, much of the existing research predominantly concentrates on enhancing the performance of specific individual components, neglecting the impact of medium temporal and spatial characteristics on the overall HVAC system performance. Consequently, issues such as energy efficiency trade-offs in practical systems remain inadequately addressed. In this paper, the medium spatiotemporal characteristics of HVAC systems are systematically illustrated and modeled. A novel optimization method based on these characteristics is proposed to improve the overall performance of the HVAC system. This method adopts an event-driven optimization strategy to determine the optimal timing with consideration of temporal characteristics, and an adaptive hybrid algorithm is proposed to determine the optimal setpoints with consideration of spatial characteristics. By comparing to conventional methods with constant setpoints, genetic algorithm (GA) based setpoints, and differential evolution (DE) based setpoints in both simulation test and experiment, the proposed method demonstrates the ability to reduce electricity consumption by 21.2%, 13.1%, and 10.0%, respectively. The methodology in this research provides significant guidance for engineers to promote the decoupling control and overall optimization of HVAC systems. Additionally, the results have referable values for implementing the most appropriate algorithm in practice.

AC-FPSK modulation for low-energy and low data rate optical wireless communications

This article proposes an amalgam of phase-shift keying and asymmetrically clipped frequency-shift keying (i.e., AC-FPSK) for low-energy/low-data rate optical wireless communications. We design a near-optimal, low-complexity harmonic receiver by analyzing the AC-FPSK time-domain (TD) and frequency-domain (FD) waveforms. The Euclidean distance analysis between AC-FPSK waveform pairs yields lower and upper bounds for theoretical bit error probability. Using numerical simulations, we present an exhaustive evaluation of the optimal cardinality of PSK, which illustrates the superiority of AC-FPSK. We show that AC-FPSK provides better energy efficiency versus spectral efficiency trade-off than state-of-the-art AC-FSK, direct current-FPSK, pulse amplitude modulation, and asymmetrically clipped optical orthogonal frequency division multiplexing. We also demonstrate that by making a suitable choice for the cardinality of PSK, the proposed 2-tap harmonic receiver for AC-FPSK can reach a similar complexity as the AC-FSK harmonic receiver.

Boosting Residual Networks with Group Knowledge

Recent research understands the residual networks from a new perspective of the implicit ensemble model. From this view, previous methods such as stochastic depth and stimulative training have further improved the performance of the residual network by sampling and training of its subnets. However, they both use the same supervision for all subnets of different capacities and neglect the valuable knowledge generated by subnets during training. In this manuscript, we mitigate the significant knowledge distillation gap caused by using the same kind of supervision and advocate leveraging the subnets to provide diverse knowledge. Based on this motivation, we propose a group knowledge based training framework for boosting the performance of residual networks. Specifically, we implicitly divide all subnets into hierarchical groups by subnet-in-subnet sampling, aggregate the knowledge of different subnets in each group during training, and exploit upper-level group knowledge to supervise lower-level subnet group. Meanwhile, we also develop a subnet sampling strategy that naturally samples larger subnets, which are found to be more helpful than smaller subnets in boosting performance for hierarchical groups. Compared with typical subnet training and other methods, our method achieves the best efficiency and performance trade-offs on multiple datasets and network structures. The code is at https://github.com/tsj-001/AAAI24-GKT.

Building back better: Modeling decentralized recovery in sociotechnical systems using strategic network dynamics

The ability of sociotechnical systems to recover from disruptions through decentralized, bottom-up processes is a crucial element of their overall resilience, but it is also among the most overlooked in resilience research. This paper proposes a computational framework leveraging dynamic multi-agent networks to conceptualize, model, and quantify multi-agent, decentralized recovery process building upon strategic network formation methods. This innovative approach addresses limitations in conventional strategic network formation models by accounting for the fluidity of structural shifts induced by the strategic behaviors of individual agents, and by acknowledging the inherent heterogeneity within agent populations. The central objective of this study is to highlight the critical role of bottom-up recovery and present a systematic, adaptable framework for modeling these processes in diverse real-world scenarios. We achieve this through four key elements: modeling post-disruption responses, exploring their interdependencies, accommodating multiple equilibria, and generating metrics for scenario comparison and forecasting. Our dynamic strategic network formation model offers valuable insights into crucial trade-offs in decentralized recovery processes and facilitates the integration of both bottom-up and top-down resilience strategies. Furthermore, it could serve as a predictive tool informing initial network architectural decisions by estimating likely post-disruption equilibrium states based on the initial network structure. Ultimately, this study provides a robust platform for analyzing network efficiency and resilience trade-offs, setting the stage for more informed decision-making in complex sociotechnical systems.

Histogram shifting based reversible data hiding with multiple expansion bin pairs

As the most popular reversible data hiding (RDH) technique, histogram shifting (HS) is widely investigated due to its efficient capacity-distortion trade-off. With a given histogram, the performance of HS-based RDH mainly depends on the selection of expansion bins. However, the determination of expansion bins, especially for the case of multiple pairs of expansion bins, usually involves massive solution space and is computationally expensive. To efficiently optimize the embedding performance, in this paper, the HS-based RDH with multiple expansion bin pairs (EBPs) is theoretically investigated, and some general results about the optimal EBPs are firstly derived. Based on the theoretical results, we conclude that a part of optimal EBPs can be directly determined without any computation burden for adaptive embedding. Then, inspired by the derived results, a simple yet efficient method to find the nearly optimal EBPs is proposed with a fairly low computational cost. Finally, the proposed method is further extended to multiple histograms based embedding for better capacity-distortion behavior. Experimental results show that the proposed method achieves better performance than some state-of-the-art works, especially in terms of computational cost.

Framework for optimizing AR waveguide in-coupler architectures.

Waveguide displays have been shown to exhibit multiple interactions of light at the in-coupler diffractive surface, leading to light loss. Any losses at the in-coupler set a fundamental upper limit on the full-system efficiency. Furthermore, these losses vary spatially across the beam for each field, significantly decreasing the displayed image quality. We present a framework for alleviating the losses based on irradiance, efficiency, and MTF maps. We then derive and quantify the innate tradeoff between the in-coupling efficiency and the achievable modulation transfer function (MTF) characterizing image quality. Applying the framework, we show a new in-coupler architecture that mitigates the efficiency vs image quality tradeoff. In the example architecture, we demonstrate a computation speed that is 2,000 times faster than that of a commercial non-sequential ray tracer, enabling faster optimization and more thorough exploration of the parameter space. Results show that with this architecture, the in-coupling efficiency still meets the fundamental limit, while the MTF achieves the diffraction limit up to and including 30 cycles/deg, equivalent to 20/20 vision.

FRAMM: Fair ranking with missing modalities for clinical trial site selection

The underrepresentation of gender, racial, and ethnic minorities in clinical trials is a problem undermining the efficacy of treatments on minorities and preventing precise estimates of the effects within these subgroups. We propose FRAMM, a deep reinforcement learning framework for fair trial site selection to help address this problem. We focus on two real-world challenges: the data modalities used to guide selection are often incomplete for many potential trial sites, and the site selection needs to simultaneously optimize for both enrollment and diversity. To address the missing data challenge, FRAMM has a modality encoder with a masked cross-attention mechanism for bypassing missing data. To make efficient trade-offs, FRAMM uses deep reinforcement learning with a reward function designed to simultaneously optimize for both enrollment and fairness. We evaluate FRAMM using real-world historical clinical trials and show that it outperforms the leading baseline in enrollment-only settings while also greatly improving diversity.

Real time Semantic Segmentation for Human-Labeled Data: A comparative Study Between CNN and Transformer

Convolutional Neural Networks have rapidly developed in the field of computer vision, notably in image classification, semantic segmentation, and object detection. These networks efficiently extract image features through local receptive fields and shared weights. Despite their effectiveness in various applications, CNNs face limitations, such as challenges in managing large-scale parameters and a tendency to overfit, especially in complex scenarios requiring contextual understanding. On the other hand, Transformer-based models, originally designed for natural language processing, have recently gained prominence in computer vision. They are particularly adept at capturing long-range dependencies, a critical aspect for interpreting complex visual scenes. Their scalability and adaptability open up new avenues for innovation. However, these models also come with drawbacks, including a need for extensive training data and higher computational costs. Their complex structures make optimization particularly challenging in resource-constrained environments. In this paper, our focus is on real-time model comparisons between CNN and Transformer architectures. We represent CNNs with the STDC model and Transformer-based models with the SegFormer. Our analysis revealed that the STDC model significantly outperforms in inference speed, achieving around 97 frames per second (fps), which is notably faster than the 50 fps of the SegFormer B0. However, when it comes to accuracy, the SegFormer B0 demonstrates superiority with a mean Intersection over Union (mIoU) of 86.78, which is more favorable compared to the 82.9 mIoU of the STDC. This study underscores the efficiency and accuracy trade-offs between these two architectures, highlighting the strengths of both CNNs and Transformer-based models in real-time applications.

The Blue Problem: OLED Stability and Degradation Mechanisms.

OLED technology has revolutionized the display industry and is promising for lighting. Despite its maturity, there remain outstanding device and materials challenges to address. Particularly, achieving stable and highly efficient blue OLEDs is still proving to be difficult; the vast array of degradation mechanisms at play, coupled with the precise balance of device parameters needed for blue high-performance OLEDs, creates a unique set of challenges in the quest for a suitably stable yet high-performance device. Here, we discuss recent progress in the understanding of device degradation pathways and provide an overview of possible strategies to increase device lifetimes without a significant efficiency trade-off. Only careful consideration of all variables that go into OLED development, from the choice of materials to a deep understanding of which degradation mechanisms need to be suppressed for the particular structure, can lead to a meaningful positive change toward commercializable blue devices.

Shifting the mechanical strength-healing efficiency trade-off of polyurethanes by incorporating boronic ester bonds and hydrogen bonding

It is challenging to design and synthesize elastomers with both good mechanical performance and high healing efficiency because their mechanisms are mutually exclusive. This work aims to shift the trade-off between them by incorporating dynamic non-covalent bonds via 2-ureido-4[1H]-pyrimidione (UPy) and dynamic covalent bonds via boronic ester bonds into polyurethane. The quadruple hydrogen bonds between UPy dimers serve as physical crosslinks, improving the overall mechanical properties of the polyurethane, while the boronic esters facilitate its healing. The resulting polyurethane can achieve an ultimate stress of 27.4 MPa and a healing efficiency of 90 %. The effect of healing time on the healing efficiency is modeled and discussed.

Trade-offs in Encapsulation Efficiency and Drug Release: A Multi-Objective Optimization Approach

Trade-offs in Encapsulation Efficiency and Drug Release: A Multi-Objective Optimization Approach