Real Data Research Articles

New information measures for linear Diophantine fuzzy sets and their applications with LDF-ARAS on data storage system selection problem

In addition to membership and non-membership degrees, linear Diophantine fuzzy set (LDFS) incorporates two reference parameters in fuzzy set definition to represent human judgment more comprehensively. It also provides flexibility in updating the decision setting by changing the meaning of the reference parameters when required by the decision-maker. Information measures in fuzzy sets have many application areas in the literature. Especially, distance and entropy measures are utilized in multiple attribute decision-making (MADM) field. Distance measures are needed by distance-based methods like TOPSIS and EDAS because their alternative ranking procedure is based on measuring the distances between alternatives and some artificial ideal or average solutions. Entropy measure is a very important tool being used in objective attribute weighting procedures of MADM applications. Thus, new distance and entropy measure propositions for LDFS are the first contribution of the study. Also, Additive Ratio Assessment (ARAS) method is extended into LDFS (LDF-ARAS) for the first time in the literature as a second contribution. In this extension, we selected ARAS as one of the most cited tools because it is found flexible and easily adaptable by researchers who do not have any background in MADM. The proposed entropy measure is integrated with the objective attribute weighting procedure of the novel LDF-ARAS approach. The third contribution is the newly developed decision model proposed for a real-life data storage system (DSS) selection problem. Three DSS options were evaluated with respect to thirteen attributes by five skilled and well-educated experts and LDF-ARAS was performed to rank alternatives. The entropy-based LDF-ARAS approach was validated by comparing the results of six different applications. In the first three, different entropy formulas are used and in the last three, these entropy measures are integrated with LDF-TOPSIS. All six comparisons gave similar alternative rankings.

Minimisation of the Energy Expenditure of Electric Vehicles in Municipal Service Companies, Taking into Account the Uncertainty of Charging Point Operation

This article presents an original method for minimising the energy expenditure of electric vehicles used in municipal service undertakings, taking into account the uncertainty in the functioning of their charging points. The uncertainty of the charging points’ operation was presented as the probability of the occurrence of an emergency situation hindering a point’s operation, e.g., a breakdown or lack of energy supply. The problem is how to calculate the driving routes of electric vehicles so that they will arrive at charging points at times at which there is a minimal probability of breakdowns. The second aspect of this problem to be solved is that the designated routes are supposed to ensure the minimum energy expenditure that is needed for the vehicles to complete the tasks assigned. The developed method is based on two heuristic algorithms, i.e., the ant algorithm and genetic algorithms. These algorithms work in a hybrid combination, i.e., the ant algorithm generates the initial population for the genetic algorithm. An important element of this method is the decision-making model for defining the driving routes of electric vehicles with various restrictions, e.g., their battery capacity or the permissible risk of charging point breakdown along the routes of the vehicles. The criterion function of the model was defined as the minimisation of the energy expenditure needed by the vehicles to perform their transport tasks. The method was verified against real-life data, and its effectiveness was confirmed. The authors presented a method of calibrating the developed optimisation algorithms. Theoretical distributions of the probability of charging point failure were determined based on the Statistica 13 program, while a graphical implementation of the method was carried out using the PTV Visum 23 software.

ViPRA-Haplo: De Novo Reconstruction of Viral Populations Using Paired End Sequencing Data.

We present ViPRA-Haplo, a de novo strain-specific assembly workflow for reconstructing viral haplotypes in a viral population from paired-end next generation sequencing (NGS) data. The proposed Viral Path Reconstruction Algorithm (ViPRA) generates a subset of paths from a De Bruijn graph of reads using the pairing information of reads. The paths generated by ViPRA are an over-estimation of the true contigs. We propose two refinement methods to obtain an optimal set of contigs representing viral haplotypes. The first method clusters paths reconstructed by ViPRA using VSEARCH Deorowicz et al. 2015 based on sequence similarity, while the second method, MLEHaplo, generates a maximum likelihood estimate of viral populations. We evaluated our pipeline on both simulated and real viral quasispecies data from HIV (and real data from SARS-COV-2). Experimental results show that ViPRA-Haplo, although still an overestimation in the number of true contigs, outperforms the existing tool, PEHaplo, providing up to 9% better genome coverage on HIV real data. In addition, ViPRA-Haplo also retains higher diversity of the viral population as demonstrated by the presence of a higher percentage of contigs less than 1000 base pairs (bps), which also contain k-mers with counts less than 100 (representing rarer sequences), which are absent in PEHaplo. For SARS-CoV-2 sequencing data, ViPRA-Haplo reconstructs contigs that cover more than 90% of the reference genome and were able to validate known SARS-CoV-2 strains in the sequencing data.

Unravelling the last-mile conundrum: A comparative study of autonomous delivery robots, delivery bicycles, and light commercial vehicles in 14 varied European landscapes

In this study, the growing need for efficient delivery services in the expanding e-commerce sector is addressed, with a focus on real-life consumption data. A comprehensive modelling framework is proposed to evaluate the efficiency of various transportation modes, including Light Commercial Vehicles (LCVs), cargo bicycles, and Autonomous Delivery Robots (ADRs). Utilizing the Google API, delivery destinations are identified, origin-destination matrices are created, and routes are optimized using Google OR-Tools and a capacitated vehicle routing problem solver. The study's robustness is further enhanced by incorporating real-life consumption data, considering diverse European contexts, varying urban scales, traffic patterns, and topographical factors, thus assessing their impact on transportation efficiency. The findings reveal that ADRs are efficient in pedestrian-focused, traffic-limited areas, while bicycles are more effective in dense city centres. This research highlights the necessity of tailoring transportation mode choice to specific urban characteristics for optimal efficiency and consumer satisfaction.Overall, the present study offers valuable insights into optimizing delivery services in different urban settings, providing a significant model for improving last-mile delivery systems. It contributes to understanding how different transportation modes can be effectively integrated into urban logistics, addressing environmental sustainability, operational efficiency, and real-life consumer demands.

Evaluation of the clinical and economic value of sofosbuvir/velpatasvir (SOF/VEL) in patients with chronic hepatitis C in Spain during the last 5 years

ObjectiveDirect-acting antivirals for the treatment of hepatitis C virus (HCV) represented a paradigm shift. In 2017, sofosbuvir/velpatasvir (SOF/VEL-Epclusa®) was approved, which showed a high cure rate in all patient, contributing to HCV elimination. The analysis aimed to quantify the clinical and economic value of SOF/VEL in HCV chronic patients since its approval in Spain. MethodsAn economic evaluation was elaborated adapting a Markov model that simulated the lifetime disease progression in of all HCV chronic patients treated with SOF/VEL (30,488 patients) since its launch (5-years), compared to previous therapies. Patients entered the model and were distributed between the fibrosis states (F0-to-F4) in treated and untreated. All patients (100%) were treated with SOF/VEL regardless of their fibrosis, and 49% with previous therapies in ≥F2. The average sustained viral response (SVR) rates 98.9% SOF/VEL versus 61.0% previous therapies. All parameters for the analysis were obtained from real-life data and literature. Only direct healthcare costs associated with disease management were included. The SOF/VEL value was measured as the number of hepatic complications avoided and their associated cost, and hepatic mortality compared to previous therapies. National Health System perspective and a 3% discount rate was applied. ResultsSOF/VEL decreased the number of liver complications, avoiding 92% decompensated cirrhosis, 80% hepatocellular carcinomas, and 87% liver transplants, as well as 85% liver-related mortality. Their cost associated was reduced, amounting to savings of 197M€. ConclusionSOF/VEL adds relevant value to the HCV treatment, reducing the clinical and economic disease burden and contributing to HCV elimination in Spain.

Trastuzumab deruxtecan in previously treated HER2-positive metastatic or unresectable breast cancer: Real-life data from the temporary use authorization program in France.

Early access program (formerly cohort Temporary Authorization for Use) was granted for trastuzumab deruxtecan (T-DXd) in France based on DESTINY-Breast01 trial which demonstrated its efficacy and safety in HER2-positive metastatic/unresectable breast cancer after ≥2 anti-HER2-based regimens received at metastatic stage. This multicenter real-world early access program included HER2-positive metastatic/unresectable breast patients pretreated with at least two lines of anti-HER2 regimens who received T-DXd 5.4 mg/kg intravenously in monotherapy every 3 weeks. Four hundred and fifty-nine patients (median age, 58 years; hormone receptor-positive, 67%; brain metastases, 28.1%) received T-DXd. Before inclusion, 81.7% of patients had radiation therapy and 76.5% had undergone surgery. Median number of prior metastatic treatment lines was four (range, 2-22); 99.8% patients had received trastuzumab, 94.8% trastuzumab emtansine and 79.3% pertuzumab. Follow-up was performed from September 30, 2020 to March 30, 2021; when the early access program stopped, the median duration of T-DXd treatment was 3.4 (range, 0-7.8) months. In 160 patients with available tumor assessment, objective response rate was 56.7% and 12.1% had progression. In 57 patients with available brain tumor assessment, complete or partial intracranial response was reported for 35.7% patients and 5.4% had progression. A total of 17 (3.7%) patients with interstitial lung disease (ILD) was reported with no cases of ILD-related death. In this early access program in patients with heavily pretreated HER2-positive metastatic/unresectable breast cancer, T-DXd had antitumor activity with a similar response to that reported in previous clinical studies. T-DXd was well tolerated and no new safety signals were observed.

EPIC: Emotion Perception by Spatio-Temporal Interaction Context of Gait.

Recently, psychophysiological computing has received considerable attention. Due to easy acquisition at a distance and less conscious initiation, gait-based emotion recognition is considered as a valuable research branch in the field of psychophysiological computing. However, most existing methods rarely explore the spatio-temporal context of gait, which limits the ability to capture the higher-order relationship between emotion and gait. In this paper, we utilize a range of research, including psychophysiological computing and artificial intelligence, to propose an integrated emotion perception framework called EPIC, which can find novel joint topology and generate thousands of synthetic gaits by spatio-temporal interaction context. First, we analyze the joint coupling among non-adjacent joints by calculating Phase Lag Index (PLI), which can discover the latent connection among body joints. Second, to synthesize more sophisticated and accurate gait sequences, we explore the effect of spatio-temporal constraints, and propose a new loss function that utilizes the Dynamic Time Warping (DTW) algorithm and pseudo-velocity curve to constrain the output of Gated Recurrent Units (GRU). Finally, Spatial Temporal Graph Convolution Networks (ST-GCN) is used to classify emotions using the generation and the real data. Experimental results demonstrate our approach achieves the accuracy of 89.66%, and outperforms the state-of-the-art methods on Emotion-Gait dataset.

3D-DGGAN: A Data-Guided Generative Adversarial Network for High Fidelity in Medical Image Generation.

Three-dimensional images are frequently used in medical imaging research for classification, segmentation, and detection. However, the limited availability of 3D images hinders research progress due to network training difficulties. Generative methods have been proposed to create medical images using AI techniques. Nevertheless, 2D approaches have difficulty dealing with 3D anatomical structures, which can result in discontinuities between slices. To mitigate these discontinuities, several 3D generative networks have been proposed. However, the scarcity of available 3D images makes training these networks with limited samples inadequate for producing high-fidelity 3D images. We propose a data-guided generative adversarial network to provide high fidelity in 3D image generation. The generator creates fake images with noise using reference code obtained by extracting features from real images. The generator also creates decoded images using reference code without noise. These decoded images are compared to the real images to evaluate fidelity in the reference code. This generation process can create high-fidelity 3D images from only a small amount of real training data. Additionally, our method employs three types of discriminator: volume (evaluates all the slices), slab (evaluates a set of consecutive slices), and slice (evaluates randomly selected slices). The proposed discriminator enhances fidelity by differentiating between real and fake images based on detailed characteristics. Results from our method are compared with existing methods by using quantitative analysis such as Fréchet inception distance and maximum mean discrepancy. The results demonstrate that our method produces more realistic 3D images than existing methods.

Physical Modelling Of A Centralized Controlled Array Of Five Wecfarm Wave Energy Converters

Point absorber wave energy converters (WECs) closely placed in array hydrodynamically interact through wave radiation and diffraction. The power absorption by these WECs is optimised by altering the WEC dynamics through the control of the Power Take-Off (PTO). As the WEC dynamics are changed, the hydrodynamic interactions change as well. Therefore, the PTO control should be optimized taking all these interactions into account, referred to as centralized control. The ‘WECfarm’ project has been initiated to study WEC arrays, and to address the research gap on available realistic and reliable data on WEC array tests to validate numerical models (Vervaet et al., 2022). This work discusses the experimental design, implementation and testing of a centralized controlled array of five ‘WECfarm’ heaving point absorber WECs, tested at the Coastal and Ocean Basin Ostend.

Single-Image 3-D Reconstruction: Rethinking Point Cloud Deformation.

Single-image 3-D reconstruction has long been a challenging problem. Recent deep learning approaches have been introduced to this 3-D area, but the ability to generate point clouds still remains limited due to inefficient and expensive 3-D representations, the dependency between the output and the number of model parameters, or the lack of a suitable computing operation. In this article, we present a novel deep-learning-based method to reconstruct a point cloud of an object from a single still image. The proposed method can be decomposed into two steps: feature fusion and deformation. The first step extracts both global and point-specific shape features from a 2-D object image, and then injects them into a randomly generated point cloud. In the second step, which is deformation, we introduce a new layer termed as GraphX that considers the interrelationship between points like common graph convolutions but operates on unordered sets. The framework can be applicable to realistic image data with background as we optionally learn a mask branch to segment objects from input images. To complement the quality of point clouds, we further propose an objective function to control the point uniformity. In addition, we introduce different variants of GraphX that cover from best performance to best memory budget. Moreover, the proposed model can generate an arbitrary-sized point cloud, which is the first deep method to do so. Extensive experiments demonstrate that we outperform the existing models and set a new height for different performance metrics in single-image 3-D reconstruction.

Analysis of Time-to-Lane-Change-Initiation Using Realistic Driving Data

Analysis of Time-to-Lane-Change-Initiation Using Realistic Driving Data

Towards Energy-Balance Closure with a Model of Dispersive Heat Fluxes

In the last decades the energy-balance-closure problem has been thoroughly investigated from different angles, resulting in approaches to reduce but not completely close the surface energy balance gap. Energy transport through secondary circulations has been identified as a major cause of the remaining energy imbalance, as it is not captured by eddy covariance measurements and can only be measured additionally with great effort. Several models have already been developed to close the energy balance gap that account for factors affecting the magnitude of the energy transport by secondary circulations. However, to our knowledge, there is currently no model that accounts for thermal surface heterogeneity and that can predict the transport of both sensible and latent energy. Using a machine-learning approach, we developed a new model of energy transport by secondary circulations based on a large data set of idealized large-eddy simulations covering a wide range of unstable atmospheric conditions and surface-heterogeneity scales. In this paper, we present the development of the model and show first results of the application on more realistic LES data and field measurements from the CHEESEHEAD19 project to get an impression of the performance of the model and how the application can be implemented on field measurements. A strength of the model is that it can be applied without additional measurements and, thus, can retroactively be applied to other eddy covariance measurements to model energy transport through secondary circulations. Our work provides a promising mechanistic energy balance closure approach to 30-min flux measurements.

Optimization of Switching Algorithms for a Thermal Switch-PCM-Switch Stack in a Building Wall for Energy Savings

Thermally switchable building envelopes can provide energy savings compared to their traditional static counterparts. For example, in the summertime cooling season the envelope can switch into a thermally conductive state at night to exploit the free cooling of the nighttime air, while switching back into a highly insulating state during the warm daytime. Further benefits can be realized by incorporating phase change materials (PCMs) into the same building envelope between two thermal switches, because the PCM can store that nighttime cooling for later use during the day. Here we use analytical and numerical modelling to investigate the advantages of a switch-PCM-switch building envelope and specifically look to optimize the switching algorithms for the two switches. We frame the problem using dimensionless groups in order to identify the key parameter groupings that determine the response. We then build a lightweight lumped element model with runtime speeds of <0.5 seconds per scenario, which enables a brute force search of over 300,000 potential switching algorithms to identify the one with the highest total and net energy savings. We compare the results of the brute force approach with an algorithm that was proposed previously based on physical intuition [1]. We then conduct a similar brute force investigation of switching algorithms optimizing for average daily cost of energy, focusing on scenarios where time-of-day energy pricing is important. Using the resulting optimal switching algorithms, we present cost and energy savings for several real cities based on their realistic temperature data and time-of-day pricing.

EVALUATION OF THE EFFICACY AND TOLERABILITY OF A FIXED DOSE COMBINATION OF AMLODIPINE AND INDAPAMIDE IN PATIENTS OLDER THAN 55 YEARS

Objective: There is very limited real-life data on hypertension management in Bangladeshi population. Considering this fact, this study evaluated the effectiveness of treatment with an indapamide/amlodipine single-pill combination (SPC) in Bangladeshi patients over the age of 55 years, in an OPD setting with uncontrolled systolic hypertension in real-life clinical practice. Methods: This was a 3-month, multicenter, observational, open-label study conducted in 3 divisional cities of Bangladesh among patients with grade I or II hypertension who were either uncontrolled on previous antihypertensive treatment or treatment-naïve. The effectiveness of indapamide/amlodipine SPC was assessed by the change in office systolic blood pressure (SBP) and the rate of target SBP (&lt; 140 mmHg) achievement at 2 weeks, 1 month and 3 months, in four age groups: 55-59 years, 60-69 years, 70-79 years, and 80 years or older. Design and method: The COMBINE study recruited 213 patients, of whom 185 took indapamide/amlodipine 1.5/5 mg SPC fora full three-month course of therapy. Mean age was 62.4±7.4 years, 76 men [41.1%] and 109 women [58.9%]. The absolute decrease of SBP during 3 months of taking indapamide/amlodipine single-pill combination (SPC) was 29.3 mm Hg, DBP - 12.9 mm Hg. After 3 months of treatment significant SBP decreases from baseline were observed in each age group: ? 31.8 mmHg (from 156.7 to 124.9), ? 26.4 mmHg (from 152.7 to 126.3), ? 29.5 mmHg (from 152.8 to 123.3), and ? 25.8 mmHg (from 151.7 to 125.9) in the 55-59, 60-69, 70-79, and 80 years and older age groups, respectively. Leg edema was observed in 3 patients (1.62%), electrolyte imbalance was observed in 4 patients (2.16%) and 1 patients observed sudden fall of blood pressure. Conclusion: This COMBINE study results showed that indapamide/amlodipine SPC was associated with significant and rapid reductions in BP even in a different age range of Bangladeshi patients in routine clinical practice. This study result is consistent with already achieved data of this SPC globally.

The single and double inverse Pareto–Burr XII distribution: Properties, estimation methods, and real-life data applications

Researchers consistently improve statistical tools to provide a flexible method to fit or predict. The solution needs to be more suitable since the approach is sometimes complex and complicated. Generally, statistical tools are generated from ordinary least squares and wavelet methods. However, the reliability tool provides a good fit comparable to state-of-the-art methods, including the complex posterior setting of hyperparameters. In this article, a flexible extension of the Burr XII model, which is the odd inverse Pareto–Burr XII (OIPBXII) distribution, is studied and investigated using statistical tools such as Markov chain Monte Carlo methods. The mathematical properties of the probability density distribution of OIPBXII are derived, and it shows a behavior shape, decreasing, increasing, J-shaped, reversed-J-shaped, bathtub, upside-down bathtub, decreasing–increasing–decreasing hazard rates, and right-skewed, symmetrical, and concave down densities. The OIPBXII parameters are investigated by using seven classical approaches to estimation. Extensive simulation results are presented to explore the performance of these methods for small and large samples. An application to two real-life sets of data from engineering and medicine is analyzed, showing the flexibility of the OIPBXII distribution as compared to existing Burr XII distributions.

Topp-Leone Exponentiated Burr XII Distribution: Theory and Application to Real-Life Data Sets

This paper introduces a novel compound distribution, named the Topp Leone Exponentiated Burr XII (TLEBXII) Distribution, aimed at generating a more flexible family of probability distributions. By taking the Exponentiated Burr XII distribution as the base and the Topp Leone exponentiated G family as the generator, this new distribution seeks to enhance modeling capabilities in various statistical applications. We present general expressions for the density and distribution functions of the TLEBXII model and explore its fundamental properties, including survival and hazard rate functions, moments, moment-generating functions, and entropy. Additionally, we discuss the methodology for estimating the model parameters through maximum likelihood estimation. The utility and superior performance of the TLEBXII distribution are demonstrated through its application to two real datasets, where it shows significant improvements over existing models. This paper underscores the potential of the TLEBXII distribution in providing a robust tool for statistical modeling and analysis.

A new generalization of Samade distribution with properties and its application to lung cancer data

Abstract In this study, a new generalization of Samade distribution has been proposed. The weighted Samade distribution (WSD) is a new distribution designed for modeling real-life data. The survival function, hazard function, moments, moment-generating function, order statistics, and entropies of this distribution have all been determined. The parameters of the suggested model are estimated using the maximum likelihood estimation technique. Finally, the helpfulness and application of the distribution have been demonstrated by lung cancer data.

Average-Case Analysis of Greedy Matching for Large-Scale D2D Resource Sharing

Given the proximity of many wireless users and their diversity in consuming local resources (e.g., data-plans, computation and energy resources), device-to-device (D2D) resource sharing is a promising approach towards realizing a sharing economy. This paper adopts an easy-to-implement greedy matching algorithm with distributed fashion and only sub-linear <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$O(\log n)$</tex-math></inline-formula> parallel complexity (in user number <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$n$</tex-math></inline-formula> ) for large-scale D2D sharing. Practical cases indicate that the greedy matching's average performance is far better than the worst-case approximation ratio 50% as compared to the optimum. However, there is no rigorous average-case analysis in the literature to back up such encouraging findings and this paper is the first to present such analysis for multiple representative classes of graphs. For 1D linear networks, we prove that our greedy algorithm performs better than 86.5% of the optimum. For 2D grids, though dynamic programming cannot be directly applied, we still prove this average performance ratio to be above 76%. For the more challenging Erdos-Rényi random graphs, we equivalently reduce to the asymptotic analysis of random trees and successfully prove a ratio up to 79%. Finally, we conduct experiments using real data to simulate realistic D2D networks, and show that our analytical performance measure approximates well practical cases.

Enhanced Security and Privacy via Fragmented Federated Learning.

In federated learning (FL), a set of participants share updates computed on their local data with an aggregator server that combines updates into a global model. However, reconciling accuracy with privacy and security is a challenge to FL. On the one hand, good updates sent by honest participants may reveal their private local information, whereas poisoned updates sent by malicious participants may compromise the model's availability and/or integrity. On the other hand, enhancing privacy via update distortion damages accuracy, whereas doing so via update aggregation damages security because it does not allow the server to filter out individual poisoned updates. To tackle the accuracy-privacy-security conflict, we propose fragmented FL (FFL), in which participants randomly exchange and mix fragments of their updates before sending them to the server. To achieve privacy, we design a lightweight protocol that allows participants to privately exchange and mix encrypted fragments of their updates so that the server can neither obtain individual updates nor link them to their originators. To achieve security, we design a reputation-based defense tailored for FFL that builds trust in participants and their mixed updates based on the quality of the fragments they exchange and the mixed updates they send. Since the exchanged fragments' parameters keep their original coordinates and attackers can be neutralized, the server can correctly reconstruct a global model from the received mixed updates without accuracy loss. Experiments on four real data sets show that FFL can prevent semi-honest servers from mounting privacy attacks, can effectively counter-poisoning attacks, and can keep the accuracy of the global model.

Learning Bilateral Cost Volume for Rolling Shutter Temporal Super-Resolution.

Rolling shutter temporal super-resolution (RSSR), which aims to synthesize intermediate global shutter (GS) video frames between two consecutive rolling shutter (RS) frames, has made remarkable progress with the development of deep convolutional neural networks over the past years. Existing methods cascade multiple separated networks to sequentially estimate intermediate motion fields and synthesize target GS frames. Nevertheless, they are typically complex, do not facilitate the interaction of complementary motion and appearance information, and suffer from problems such as pixel aliasing or poor interpretation. In this paper, we derive the uniform bilateral motion fields for RS-aware backward warping, which endows our network a more explicit geometric meaning by injecting spatio-temporal consistency information through time-offset embedding. More importantly, we develop a unified, single-stage RSSR pipeline to recover the latent GS video in a coarse-to-fine manner. It first extracts pyramid features from given inputs, and then refines the bilateral motion fields together with the anchor frame until generating the desired output. With the help of our proposed bilateral cost volume, which uses the anchor frame as a common reference to model the correlation with two RS frames, the gradually refined anchor frames not only facilitate intermediate motion estimation, but also compensate for contextual details, making additional frame synthesis or refinement networks unnecessary. Meanwhile, an asymmetric bilateral motion model built on top of the symmetric bilateral motion model further improves the generality and adaptability, yielding better GS video reconstruction performance. Extensive quantitative and qualitative experiments on synthetic and real data demonstrate that our method achieves new state-of-the-art results.