Capacity Problem Research Articles

Assessing bias and computational efficiency in vision transformers using early exits

Face recognition with deep learning is generally approached as a problem of capacity. The field has seen progressively deeper, more complex models or larger, more highly variant data sets. The data sets can be problematic, as they are often scraped indiscriminately from the internet. This results in an uncertain, and often heavily unbalanced distribution of race, gender, age and other aspects of the subjects, which is then manifested in the decisions of the models trained on them. The carbon footprint of machine learning is a concern. A real push is developing to reduce the energy consumption of machine learning as we strive for a more eco-friendly society. In addition, due to many instances of misuse by law enforcement and other agencies, unbiased models for face recognition are now fundamental to the practical application of the field. We present an approach using the state of the art Vision Transformer and Early Exits for reducing compute budget without significantly affecting performance. We develop a system for face recognition and identification with a closed-set gallery and show that with a small reduction in performance, a reasonable reduction in compute cost can be obtained using our method. Second, we investigate how these early Exits interact with the bias model through a robust evaluation of matching scores on a racially balanced data set. We show that matching scores vary heavily between cohorts, and these variations are magnified at the earlier exits.

Rehabilitative short-term care (REKUP): acceptance and practicability of anew care concept

Rehabilitative short-term care (REKUP) is aimed at geriatric patients in need of rehabilitation but not (yet) capable of rehabilitation after acute inpatient hospitalization with apositive rehabilitation prognosis. It was tested at two geriatric rehabilitation clinics from October 2020 to March 2022. Qualitative process evaluation with respect to acceptance (service providers, health insurance), adoption (implementation at the start of the project), practicability (processes, contents, structures) and observed patient satisfaction. Afocus group and 14semi-structured interviews with atotal of 18employees of both rehabilitation clinics and ahealth insurance provider were recorded, transcribed and qualitatively analyzed for content. The survey of service providers (acute care, model clinics) and the health insurance provider revealed ahigh level of acceptance and good practicability (processes, contents, structures). Due to the high level of acceptance and the familiar service content in the geriatric rehabilitation clinics, the implementation was feasible and patient satisfaction was rated as very good. The acceptance of REKUP was high because the basic idea was easy to understand. Nevertheless, acareful patient selection is crucial. There is arisk that REKUP could be used as aquick discharge option to avoid capacity problems. The success of treatment perceived by the service provider is based on early exercises, psychosocial care and alonger period of time before rehabilitation, which avoids excessive demands and strengthens self-efficacy. Admission to the long-term nursing home sector is avoided, which has apositive psychological and motivational effect on patients. Atransfer of REKUP to other geriatric rehabilitation clinics seems to be easily possible.

Modelling of shallow groundwater levels and interaction with drainage system and streams under the impact of climate change

ABSTRACT Under the impact of climate change, groundwater levels are anticipated to increase in temperate climatological zones with increased precipitation and sea level rise. Urban areas located at low elevations are especially at risk, and increased shallow groundwater levels can lead to flooding, water seepage into basements, capacity problems in water infrastructure, etc. Furthermore, old sewer systems can act as drains by allowing groundwater to seep into the sewer system. This process can artificially maintain the groundwater level low, and therefore if the sewer system is rehabilitated, the problems are even increased further if necessary solutions are not included in the planning process. In this article, we analyze an urban catchment in Aalborg, Denmark, which is currently facing problems with high groundwater levels. By simulating the groundwater interaction with sewer systems and watercourses, we show that the groundwater table can increase up to 1 m under the impact of sewer rehabilitation and climate change. This increase will lead to significant flood risk in the residential area. We propose multiple technical solution scenarios and evaluate their effectiveness, demonstrating how the impact of increasing groundwater levels can be mitigated through implementation of drains during sewer reconstruction, but not completely avoided.

Review of Space Debris Modeling Methods and Development Direction of the Korean Space Debris Models

Space debris poses significant threats to spacecraft and human activities in space. Accurate modeling of space debris is crucial for understanding and mitigating these risks, ensuring the sustainability of the space environment. This paper discusses the importance of space debris modeling in the space environment, highlighting its critical role in safeguarding assets in orbit. Two primary methods of space debris modeling, namely the 1D and 3D approaches, are discussed in detail, and their respective strengths and limitations are elucidated. Furthermore, a comprehensive review of existing models, including the space debris evolutionary model (MOCAT, SOLEM, DAMAGE, LEODEEM & GEODEEM, DELTA, and LEGEND) and engineering models (MOCAT-MC, NEODEEM, MASTER, ORDEM), are presented. These models offer valuable insights into the dynamics and characteristics of space debris populations, aiding in formulating effective debris mitigation strategies and orbital capacity problems for reducing the possibilities of Kessler’s syndrome. Additionally, the paper provides insights into the ongoing development of the Korean space debris model, focusing on its methodology and space debris cataloging techniques for modeling space debris environments.

Optimalisasi Kapasitas Stasiun Kerja dengan Metode Theory of Constraints (TOC) di PT. XYZ

PT. XYZ is a company that produces rims cars. One of the products produced is rims with ABC type. The production process carried out is make to order repetitive. PT. XYZ has a production capacity problem, causing an unfulfilled order. The production floor has a bottleneck work station at work stations 1 and 2. The accumulation occurs because of the standard time difference and the maximum capacity. This study uses the Theory of Constraints (TOC) method to deal with bottlenecked work stations. Calculating the maximum capacity of each work station results in work station s 1 and 2 having less capacity than the other work station. To solve the bottleneck problem at work stations 1 and 2, there are two alternative solutions, overtime and the addition of a work shift. After analyzing the existing problem, to solve bottlenecks at work station s 1 and 2, add the shift of work from 2 shifts to 3 shifts with an overtime of 1 hour per shift.

Flexible DC Optimization Control Technology Based on Zonal Interconnection of Urban Power Grids

Aiming at the problems of insufficient power supply capacity, unbalanced load, and poor safety and reliability of the urban power grid, a flexible DC optimization control strategy is proposed to support the interconnection of power grid zones. Firstly, the steady-state and transient optimal control objectives are established to meet the safety and economic operation requirements of the grid; secondly, comprehensive evaluation indexes are constructed, covering load balancing, network loss, and voltage deviation, and a step-by-step approximation method is used to quickly solve the optimal power of the flexible DC, so as to realize the optimal control of the urban grid in the steady state. Meanwhile, a multi-terminal flexible DC adaptive sag control strategy based on the power margin of the converter station is designed to reasonably allocate the emergency power support in each partition after a fault and reduce the frequency deviation of the heavily loaded nodes. Taking the urban grid as an example, a simulation model of the grid structure with a 750 KV outer ring and a 220 KV inner ring is constructed, and the simulation results verify the effectiveness of the proposed method and its engineering practicability.

OLM-SLAM: online lifelong memory system for simultaneous localization and mapping

Abstract Simultaneous Localization and Mapping is a fundamental task for robots in unknown environments. However, the poor generalization ability of learning-based algorithms in unknown environments hinders their widespread adoption. Additionally, artificial neural networks are subject to catastrophic forgetting. We propose a lifelong SLAM framework called OLM-SLAM that effectively solves the neural network catastrophic forgetting problem. To ensure the generalization of the neural network, this paper proposes a method for the sensitivity analysis of the network weight parameters. Meanwhile, inspired by human memory storage mechanisms, we design a dual memory storages mechanism that retains dynamic memory and static memory. A novel memory filtering mechanism is proposed to maximize image diversity within a fixed-size memory storage area addressing the problem of limited storage capacity of embedded devices in real-world situations. We have extensively evaluated the model on a variety of real-world datasets. Compared with CL-SLAM, the overall translation error of the test sequence is improved by 44.9%. The translation and rotation errors of Retention Ability (RA) were improved by 111.6% and 66.7%, respectively. The results demonstrate that OLM-SLAM can outperform previous methods of the same type, and OLM-SLAM has high RA when facing different sequences of the same type of environment.

Research on the mechanical and thermal properties of potting adhesive with different fillers of h-BN and MPCM

Using packaging materials to reduce contact thermal resistance has become a promising method to solve the problem of insufficient heat dissipation capacity of electronic components. The purpose of this work is to optimize the mechanical and thermodynamic performance of potting adhesive using phase change microcapsules (MPCM) and hexagonal boron nitride nano-powder (h-BN) as thermal conductive fillers. The experimental results indicated that h-BN has a positive effect on the tensile strength of the potting adhesive, with a 7.1 % increase in tensile strength at a mass fraction of 30 %. However, the addition of MPCM will weaken the tensile strength of the potting adhesive. Adding MPCM and h-BN can both effectively improve the thermal conductivity of the potting adhesive: when the filler mass fraction is lower than 20 %, the potting adhesive with MPCM filler exhibits more strengthening capability than h-BN type; while with the continuous increase of filler mass fraction, the thermal conductivity of the potting adhesive with h-BN filler is better. The thermal buffering capacity of the potting adhesive significantly increases with the mass fraction of MPCM, while the effect of h-BN on thermal buffering capacity is not significant. In addition, the addition of h-BN and MPCM significantly improves the temperature uniformity of the potting adhesive.

How should mode competitiveness and profit be balanced in maritime transport? – Vessel speed optimization approach including Northern Sea Route

ABSTRACT Transport between East Asia and Europe heavily relies on maritime transport via the Suez Canal route (SCR), with the Northern Sea Route (NSR) through the Arctic Sea emerging as a new route due to the continuously melting polar ice. However, the dominance of maritime transport faces a threat from railways and airlines, especially for high-sensitivity commodities. This study mainly maximizes the profit of a shipping company from both NSR and SCR services with vessel speed optimization, focusing on the competition between four alternatives quantified by a logit model. The sensitivity analysis is also conducted with expanding ice-classed vessels and fluctuating bunker prices. Our findings show that maintaining a speed of approximately 20kn can capture an additional 10% of the market share for SCR; however, SCR faces challenges if future advancements enable more frequent and cost-effective railway transport. Moreover, besides expanding ice-classed vessel size, tackling the problem of limited capacity may position NSR as a promising competitor. This study contributes to providing an effective and innovative speed-optimization scheme combining economic viabilities and mode choice. It is aimed to (1) provide feasible political insights regarding mode competitions and (2) promote the commercial economic use of various modes, including NSR.

Carbon-Encapsulated Bimetallic Fe-W-Based Selenides with Abundant Heterogeneous Conductive Network for Superior Potassium Ion Storage.

Potassium-ion batteries (KIBs) are one of the most promising alternatives to lithium-ion batteries (LIBs) due to their high ion mobility, abundant resources, and low cost. However, the problems of low capacity and poor cycling stability of KIBs remain unsolved; therefore, it is crucial to explore alternative electrode materials suitable for reversible insertion and extraction of K+. Herein, carbon-encapsulated Fe3Se4/WSe2 microsphere material assembled from nanosheets with abundant heterogeneous interfaces and expanded interlayer spacing (denoted as FWSC) is designed as an anode material for superior potassium ion storage. The microspheres assembled from nanosheets, the outer carbon coating, and the expanded layer spacing can expose more active sites and improve the electronic conductivity and structural stability of the material. The abundant heterogeneous conductive network not only improves the conductivity of the material but also enhances the reversibility of the electrochemical reaction. As a result, the FWSC exhibits outstanding cycling stability, excellent rate performance (the capacity retention is 53% when the current density is from 0.5 to 1.0 A g-1), and stable long-term cycle performance (a capacity of 209 mAh g-1 at 1.0 A g-1 over 1000 cycles).

Machine learning-based clustering analysis of the banking green credit for the renewable energy industry

Credit policies for clean and renewable energy businesses play a crucial role in supporting carbon neutrality efforts to combat climate change. Clustering the credit capacity of these companies to prioritize lending is essential given the limited capital available. Support Vector Machine (SVM) and Artificial Neural Network (ANN) are two robust machine learning algorithms for addressing complex clustering problems. Additionally, hyperparameter selection within these models is effectively enhanced through the support of a robust heuristic optimization algorithm, Particle Swarm Optimization (PSO). To leverage the strength of these advanced machine learning techniques, this paper aims to develop SVM and ANN models, optimized with the PSO, for the clustering problem of green credit capacity in the renewable energy industry. The results show low Mean Square Error (MSE) values for both models, indicating high clustering accuracy. The credit capabilities of wind energy, clean fuel, and biomass pellet companies are illustrated in quadrant charts, providing stakeholders with a clear view to adjust their credit strategies. This helps ensure the efficient operation of banking green credit policies.

The influence of land use planning and exposure to flooding on urban growth in San José, Costa Rica

We estimate areas of potential flooding exposure within the San José Metropolitan Region with morphometric methods. Based on this potential hazard zoning, we use a spatial-statistical model to explore the difference introduced by land use regulations (municipal and regional) on urban growth, controlling for potential exposure to hazards and other sources of spatial heterogeneity. The potential flood zoning derived coincided with upstream rivers and flat areas between them (likely susceptible to urban drainage capacity problems). We estimated regional land use regulations significantly constrain urban development (with urban growth reductions of about 6 %), but the constraint does not increase in areas exposed to potential flooding; municipal environmental protection zoning was estimated to decrease urban development by 8 %, with an additional but very small 0.1–0.9 % reduction in potential flooding areas. Regional instruments fundamentally applied to the steeper (and hence less exposed to flooding) periphery whereas municipal land use instruments did constrain land from development (especially along rivers) within urban areas.

Design and application of deployable heat radiator

Abstract Limited by the envelope size of the launch vehicle, comparing with the body-mounted radiator of the traditional spacecraft, the thermal control scheme based on single-phase fluid loop is proposed in order to improve the system-level heat dissipation. The thermal analysis model of the deployable heat radiator, including the spacecraft, is established. Through the comparative analysis of the fin parameter, infrared radiation effect, and working fluid selection, the heat dissipation efficiency of the radiator is improved, and the first flight application of the deployable heat radiator is realized. The research results show that in view of the GEO spacecraft’s bad external heat flow condition, the heat dissipation of a single radiator can reach 250W/m2 above; the method of using multiple deployable heat radiators can effectively solve the problem of insufficient heat dissipation capacity on the spacecraft.

„Traps” of Institutional Capacity of the Ukrainian State: Causes of Emergence and Ways to Overcome

In the article, based on the use of international indices (governance effectiveness, corruption perception, democracy, rule of law, and others), an assessment of the institutional capacity of the Ukrainian state is given, particularly during the Great War. The article also analyzes "capacity traps" as gaps in legislation that undermine the effectiveness of the state governance system, and as a result, devalue the institution of political accountability and deteriorate the state’s institutional capacity. The author also explains, based on sociological surveys regarding trust in state institutions, why the decline in trust towards government bodies (with the exception of the Armed Forces and the President) increases the risks of delegitimizing the government, as this translates into a reluctance of many citizens to fulfill their civic duties, particularly with respect to mobilization, tax payments, etc. The author identifies the inefficiency of the system of checks and balances as the key problem of the institutional capacity of the Ukrainian state, which leads to duplication of powers between branches of government, ineffective functioning of the parliament and government, excessive concentration of power in the hands of the president, and the nullification of the role of the government as the center of executive power. The author thoroughly analyzes the existing in Ukraine "capacity traps" (what has been done for the first time in domestic political science), emphasizes their cumulative negative effect, outlines the negative consequences if they persist, and proposes effective instruments of non-electoral legitimacy of power, which are extremely necessary during wartime. Particular attention is paid to the need to reform Ukraine’s constitutional system to eliminate these institutional traps. Several ways to enhance institutional capacity are outlined, including the self-legitimization of power through information policy under conditions of forced restrictions on rights and freedoms; amending the constitution and legislation to eliminate the "capacity traps"; and the use of non-electoral instruments of legitimacy of power as a temporary measure that can strengthen the state's institutional capacity, but only during the war.

A hybrid approach to investigating major management factors for effective highway preventive maintenance

To resolve the problem of the low management capacity of highway preventive maintenance (HPM), this paper identified and evaluated the major HPM management factors to improve management effectiveness and achieve sustainable highway development. The study conducted a literature review and exploratory factor analysis (EFA) to identify the major HPM management factors. Social network analysis (SNA) was used to distinguish the degree of importance of these factors. A system dynamics (SD) model was developed to explore their patterns of influence. The research identified six dimensions of HPM management, including the management system, management resources, management cognition, management decisions, management technology, and external conditions, along with 26 major management factors. Moreover, information acquisition, system perfection, etc., are key factors; system execution, manager capability, etc., are hub factors; and route selection, machinery allocation, etc., are non-key factors. These factors have a positive impact on HPM management, leading to an upward trend in management effectiveness. The main innovation provided a hybrid and comprehensive approach to identify and evaluate the major management factors for effective HPM. This study can guide managers in developing effective HPM plans, allocating resources more efficiently, improving the overall quality of highway maintenance and forming a sustainable transportation system.

Research on mobile energy storage scheduling strategy for emergency power supply protection of post-disaster isolated loads

Aiming at the problem of insufficient power supply capacity of isolated loads in oceanic islands, a concept based on mobile energy storage and power conservation is proposed in this paper. Firstly, the energy flow and traffic flow characteristics of mobile energy storage are quantitatively analyzed, and an accurate mathematical model is established. Furthermore, the time-space distribution characteristics of load are carefully considered, and the Voronoi diagram and Tyson polygon method are introduced to determine cell differentiation. On this basis, combined with the power demand of load nodes and the energy storage characteristics of mobile energy storage vehicles, the evaluation indicators of cell energy and relative surplus energy of cells are designed, and the agent-cellular automata algorithm is used to schedule the mobile energy storage. The simulation results show that the power supply mode based on mobile energy storage can effectively improve the reliability of isolated loads. This paper provides a new perspective for solving the reliable power supply problem of isolated loads.

Research and Application of Drilling and Completion Technologies on Large-size Injection-production Well

Abstract The large-size injection-production well can increase the injection-production capacity to more than 3 times, compared with the conventional-size one. However, it had not been popularized and promoted for many years since it had the technical difficulties of long construction period, high engineering risk and large investment in a single well. The Liaohe double-6 gas storage planned to deploy 3 large-size horizontal wells, aiming at further improving injection-production capacity to satisfy peak shaving demand. To overcome the technical difficulties of large-size well, the drilling scheme was optimized according to geological characteristics of double-6 block in the paper. It included wellbore structure design, wellbore diameter control technology and hole cleaning technology. The optimized drilling scheme can not only improve the wellbore integrity, but also reduce the risk of engineering accidents and shorten the drilling period greatly. Besides, the cementing technology was optimized as well from the aspects of hierarchical cementing and cement slurry design, which can be used to solve the problems of high pressure at the bottom of well and large leaking risk of cement. On the basis, the above drilling and cementing technologies were applied to develop three large-size wells, and good practical results were achieved. The average completion period of the three well was 104 days, 36 days shorter than the planned period, which realized the applying effects that large-size wells can be designed, constructed, and produced in the same year. Moreover, the ultrasonic evaluation results showed that the average qualified rates of the three wells in technical casing was 92.8%, which was much higher than the industry standard. Besides, the test results showed that the injection and production volume of each large-size well in the paper can exceed 3 million cubic meters, solving the problem of insufficient injection-production capacity in the months with high peak shaving demand. Above all, the drilling and cementing technologies of large-size well had been optimized in the paper, which successfully reduced the well completion period, and improve the cementing quality and the safety of drilling construction. The researches can not only ensure the project quality of large-size well in Liaohe double-6 gas storage, but also provide technical support for promoting the drilling and completion engineering of large-size well in other domestic and foreign gas storages.

Association between dietary total antioxidant capacity and sleep problems and depressive symptoms among U.S. adults.

In this study, we aim to examine the impact of dietary total antioxidant capacity (TAC) on sleep problems and depressive symptoms (DS); besides, we seek to elucidate the potential mediating effect of dietary TAC on the relationships between sleep problems and DS. Weighted Kruskal-Wallis tests for continuous variables and Chi-square tests for categorical variables were employed to discriminate between DS and non-DS participants. Multivariable logistic regression and restricted cubic spline analysis were applied to evaluate the associations of TAC with DS and sleep problems. Among the 21,805 participants, 1,947 participants suffered from DS. Weighted multivariable logistical regression indicated that shorter sleep hours were linked to an increased likelihood of risk of DS even after complete adjustments. Restricted cubic spline cure displayed that TAC was almost non-linearly correlated with DS and sleep problems. Mediation analysis indicated that sleep duration slightly mediated the association between TAC and DS (proportion of mediation: 3.12%, p < 0.001). This study illustrated the inverse association between TAC value and sleep problems and DS. Furthermore, TAC slightly mediated the effect of sleep duration on the DS, and there was a nearly non-linear relationship between TAC and DS, and TAC and sleep problems.

Encoding innate ability through a genomic bottleneck

Animals are born with extensive innate behavioral capabilities, which arise from neural circuits encoded in the genome. However, the information capacity of the genome is orders of magnitude smaller than that needed to specify the connectivity of an arbitrary brain circuit, indicating that the rules encoding circuit formation must fit through a "genomic bottleneck" as they pass from one generation to the next. Here, we formulate the problem of innate behavioral capacity in the context of artificial neural networks in terms of lossy compression of the weight matrix. We find that several standard network architectures can be compressed by several orders of magnitude, yielding pretraining performance that can approach that of the fully trained network. Interestingly, for complex but not for simple test problems, the genomic bottleneck algorithm also captures essential features of the circuit, leading to enhanced transfer learning to novel tasks and datasets. Our results suggest that compressing a neural circuit through the genomic bottleneck serves as a regularizer, enabling evolution to select simple circuits that can be readily adapted to important real-world tasks. The genomic bottleneck also suggests how innate priors can complement conventional approaches to learning in designing algorithms for AI.

A Semi-Supervised Active Learning Method for Structured Data Enhancement with Small Samples

In order to solve the problems of the small capacity of structured data and uneven distribution among classes in machine learning tasks, a supervised generation method for structured data called WAGAN and a cyclic sampling method named SACS (Semi-supervised and Active-learning Cyclic Sampling), based on semi-supervised active learning, are proposed. The loss function and neural network structure are optimized, and the quantity and quality of the small sample set are enhanced. To enhance the reliability of generating pseudo-labels, a Semi-supervised Active learning Framework (SAF) is designed. This framework redistributes class labels to samples, which not only enhances the reliability of generated samples but also reduces the influence of noise and uncertainty on the generation of false labels. To mine the diversity information of generated samples, an uncertain sampling strategy based on spatial overlap is designed. This strategy incorporates the idea of spatial overlap and uses global and local sampling methods to calculate the information content of generated samples. Experimental results show that the proposed method performs better than other data enhancement methods on three different datasets. Compared to the original data, the average F1macro value of the classification model is improved by 11.5%, 16.1%, and 19.6% relative to compared methods.