Post-processing Strategy Research Articles

Generalized reassigning transform: Algorithm and applications

Time-frequency analysis (TFA) has attracted the attention of many scholars and engineers for analyzing nonstationary signals in the field of condition monitoring of rotating machinery. For complex mechanical equipment, measured signals always contain both harmonic and impulsive components, which presents a challenge for current TFA methods. To concurrently characterize harmonic and impulsive components, a novel TFA algorithm, called generalized reassigning transform (GRT) is developed in this paper. First, the time-frequency fusion extraction criterion (TFFEC), which includes time-frequency data fusion (TFDF) and time-frequency data extraction (TFDE), is designed to calculate time-frequency representation (TFR) from short-time Fourier transform (STFT) results under different window sizes, which improves time-frequency resolution and eliminates noise influence. Furthermore, the chirp rate (CR)-based postprocessing strategy is constructed to characterize nonstationary signals with both harmonic-like and impulsive-like components. Specifically, the CR discrimination criterion is introduced to classify the TFFEC result into two distinct types: harmonic-like component and impulsive-like component, and then, the TFR with high energy concentration and strong readability is obtained by advanced postprocessing TFA including the synchro-reassigning transform (SRT) and horizontal reassigning transform (HRT). The effectiveness of the GRT in condition monitoring and fault diagnosis is validated through numerical and experiment verification.

A novel trajectory planning method for mobile robotic grinding wind turbine blade

As the core component of wind turbine, wind turbine blade requires two grinding processes in the production. However, mobile robotic automation grinding of wind turbine blades is considered to be a challenging task due to the high aspect ratio and compound surface of the wind turbine blade. The trajectories generated by most robotic grinding trajectory planning algorithms are often found to be inferior in grinding large compound surface workpieces, as they are typically designed for robotic machining with a fixed base. In this paper, a novel iso-planar algorithm based on oriented bounding box (OBB) of the workpiece is developed to plan the grinding trajectories by taking into consideration the characteristics of blade. The constant chord length algorithm employing Taylor quadratic expansion is then developed to discretize trajectory into grinding points. Considering the characteristics of compound surface, a post-processing strategy is proposed to eliminate redundant grinding points and generate consistent tool orientations on compound surface. Based on these three steps, a workstation location optimization model for improving robot manipulability is introduced to determine a series of workstation locations. Furthermore, the grinding and movement synchronization strategy based on mobile platform trajectory interpolation is proposed to enhance the efficiency of machining large workpieces. The simulation and experiments demonstrate the effectiveness of the proposed trajectory planning method for mobile robotic grinding wind turbine blade, the rationality of the optimization model and the feasibility of grinding and movement synchronization strategy.

Translucent zirconia dental prosthesis processed by Direct Ink Writing: Updates and challenges

Direct ink writing (DIW) is an extrusion-based additive manufacturing technique, which uses a layer-by-layer deposition of an ink, usually containing high solid loads and low concentration of additives, to provide plasticity and structural stability for the construction of near-net-shaped pieces. In this work, a zirconia stabilized with 5 mol% Y2O3 colloidal ink was developed aiming to obtain dense prototypes, partially translucent, with potential for making dental prostheses. Cylindrical pieces (Ø 20 mm × 2.5 mm), prismatic bars (55 × 5 × 4 mm) and 3-element dental prostheses were printed based on Computer-aided design (CAD) models, using a printing speed of 10 mm.s−1 and a nozzle with Ø 0.25 mm. The printed samples were dried at strictly controlled conditions, debinded using an optimized thermal cycle, and sintered at 1600 °C-2 h. The sintered parts were characterized by X-ray diffraction, scanning electron microscopy, Vickers hardness, fracture toughness and bending strength. Concerning dental prosthesis, a post-processing finishing strategy was performed on pre-sintered bodies to highlight the anatomical details. After sintering, samples presented relative density around 94–95 %, Vickers hardness, fracture toughness and Young Modulus of 12.7 ± 0.3GPa, 5.31 ± 0.21 MPa.m1/2 and 195.4 ± 45 GPa, respectively. A characteristic strength (σ0) of 303 MPa was attained, which is above the requirements for its application as ceramic single-unit dental prostheses. A critical analysis of the technological difficulties of this 3D printing technique for manufacturing dental prostheses is presented, while future directions are addressed.

Enhancing mechanical performance of hydroxyapatite-based bone implants via citric acid post-processing in binder jetting additive manufacturing

Binder jetting is a promising technology in the additive manufacturing of bone implants, particularly for printing brittle bioceramics that are susceptible to thermal residual stresses. However, challenges in this field include low strength and undesirable size changes due to post-sintering treatments, as well as the absence of necessary organic matter like Glycosaminoglycans, citric acid, etc. To address these issues, a novel approach was introduced using citric acid (CA) as a post-processing agent to enhance the mechanical performance of green samples and add organic matter, with boric acid (BA) as a control. A hydroxyapatite (HA) based powder mixed with 25 wt.% high-viscosity polyvinyl alcohol (PVA) was prepared and printed using a self-made printer with deionized water as the binder. The post-processing effects were analyzed in terms of mechanical properties and microstructure. The application of 5 wt.% CA solution increased the thickness of the PVA film between HA particles by 320.0%, leading to an increase in compressive strength (7.37 ± 0.28 MPa) and modulus (102.81 ± 6.74 MPa) by 840.7% and 1571.3%, respectively, achieving the mechanical standards for human trabecular bone. This work presents a simple and rapid room-temperature post-processing strategy for enhancing the mechanical properties of bone implants produced by binder jetting additive manufacturing.

Comparison of acoustic, optical, and heat release rate based flame transfer functions for a lean-burn injector under engine-like conditions

Determining the flame transfer function plays a crucial role during the development phase of lean-burn injectors to predict the overall stability of the combustion system. This study develops an enhanced acoustic post-processing strategy using acoustic network modeling and compressible large-eddy simulation for a lean-burn aero-engine injector in a realistic test rig. Results show that optical flame transfer functions experimentally obtained align with those derived from large-eddy simulation based on heat release rate. However, discrepancies, especially in gain, are observed when using the standard acoustic post-processing method. By employing an acoustic network model, it is shown that comparable flame transfer functions can be achieved through refined post-processing strategies. This study highlights the limitations of conventional acoustic post-processing and underscores the necessity of explicit acoustic modeling in complex setups to accurately obtain the flame response. A generalized formulation applicable to a broader range of setups, extending beyond simple configurations, is presented as an alternative to the conventional acoustic post-processing method.

SAR Image Despeckling Based on Denoising Diffusion Probabilistic Model and Swin Transformer

The speckle noise inherent in synthetic aperture radar (SAR) imaging has long posed a challenge for SAR data processing, significantly affecting image interpretation and recognition. Recently, deep learning-based SAR speckle removal algorithms have shown promising results. However, most existing algorithms rely on convolutional neural networks (CNN), which may struggle to effectively capture global image information and lead to texture loss. Besides, due to the different characteristics of optical images and synthetic aperture radar (SAR) images, the results of training with simulated SAR data may bring instability to the real-world SAR data denoising. To address these limitations, we propose an innovative approach that integrates swin transformer blocks into the prediction noise network of the denoising diffusion probabilistic model (DDPM). By harnessing DDPM’s robust generative capabilities and the Swin Transformer’s proficiency in extracting global features, our approach aims to suppress speckle while preserving image details and enhancing authenticity. Additionally, we employ a post-processing strategy known as pixel-shuffle down-sampling (PD) refinement to mitigate the adverse effects of training data and the training process, which rely on spatially uncorrelated noise, thereby improving its adaptability to real-world SAR image despeckling scenarios. We conducted experiments using both simulated SAR image datasets and real SAR image datasets, evaluating our algorithm from subjective and objective perspectives. The visual results demonstrate significant improvements in noise suppression and image detail restoration. The objective results demonstrate that our method obtains state-of-the-art performance, which outperforms the second-best method by an average peak signal-to-noise ratio (PSNR) of 0.93 dB and Structural Similarity Index (SSIM) of 0.03, affirming the effectiveness of our approach.

Enhancing exciton-to-Mn2+ energy transfer and emission efficiency in Mn2+-doped CsPbCl3 perovskite nanocrystals via CaCl2 post-treatment

Mn2+-doped CsPbCl3 perovskite nanocrystals (NCs) exhibit a significant Stokes shift and emit bright orange-red light, making them promising candidates for optoelectronic devices. However, these NCs are prone to surface defects during multiple purification steps, which hinder the transfer of energy from excitons to Mn2+ and reduce luminescence efficiency. Herein, we introduce a post-treatment method using a CaCl2 ethanol solution to passivate surface defects in Mn2+-doped CsPbCl3 NCs. This treatment enhances the energy transfer from excitons to Mn2+ and boosts luminescence performance, achieving a photoluminescence quantum yield of up to 96 % and maintaining stability through several washing cycles. In this process, Ca2+ ions, as a Lewis acid, replace oleic acid molecules on NC surfaces, leading to stronger binding. The Cl− from the CaCl2 solution fill vacancies on NC surfaces, effectively passivating surface defects. This reduces non-radiative recombination centers, increases energy transfer efficiency from 83 % to 87 %, and accelerates the radiative recombination rates of excitons and Mn2+. The post-treated NCs also retain their PL against continuous heat and ultraviolet irradiation. The passivation of surface defects on perovskite NCs achieved through this post-processing strategy enhances the energy transfer from excitons to doped ions, providing guidance for achieving efficient and stable doped perovskite NCs.

A Grid-Based Method for Removing Overlaps of Dimensionality Reduction Scatterplot Layouts.

Dimensionality Reduction (DR) scatterplot layouts have become a ubiquitous visualization tool for analyzing multidimensional datasets. Despite their popularity, such scatterplots suffer from occlusion, especially when informative glyphs are used to represent data instances, potentially obfuscating critical information for the analysis under execution. Different strategies have been devised to address this issue, either producing overlap-free layouts that lack the powerful capabilities of contemporary DR techniques in uncovering interesting data patterns or eliminating overlaps as a post-processing strategy. Despite the good results of post-processing techniques, most of the best methods typically expand or distort the scatterplot area, thus reducing glyphs' size (sometimes) to unreadable dimensions, defeating the purpose of removing overlaps. This article presents Distance Grid (DGrid), a novel post-processing strategy to remove overlaps from DR layouts that faithfully preserves the original layout's characteristics and bounds the minimum glyph sizes. We show that DGrid surpasses the state-of-the-art in overlap removal (through an extensive comparative evaluation considering multiple different metrics) while also being one of the fastest techniques, especially for large datasets. A user study with 51 participants also shows that DGrid is consistently ranked among the top techniques for preserving the original scatterplots' visual characteristics and the aesthetics of the final results.

A greedy post-processing strategy for multi-objective performance optimization of general single-server finite queueing networks

A greedy post-processing strategy for multi-objective performance optimization of general single-server finite queueing networks

Effects of homogenization and deep cryogenic treatments on microstructure and mechanical property of D2 tool steel fabricated by laser direct energy deposition

The integration of laser direct energy deposition (L-DED) for manufacturing high-hardness D2 tool steel components presents a promising avenue for addressing the increasing complexity of product geometries demanded by modern industry. However, the inherent variability in microstructure and mechanical properties induced by L-DED across the build height necessitates the employment of post-processing techniques to achieve material homogeneity. This study explores the synergistic application of homogenization and deep cryogenic treatment (DCT) as a novel post-processing strategy aimed at mitigating microstructural inhomogeneities and enhancing the mechanical properties of L-DED-fabricated D2 tool steel. Initial analysis of the as-built samples revealed a microstructure characterized by a fine network of eutectic carbides enveloping an FCC matrix, which imposed constraints on martensite transformation during DCT. Homogenization was found to effectively alleviate these constraints, promoting compositional and microstructural uniformity and enabling a more consistent transformation to martensite across the samples upon subsequent DCT. The strategic combination of homogenization and DCT resulted in achieving uniformly distributed high hardness levels exceeding 750 HV throughout the entire sample, regardless of their initial positions within the build. This study demonstrates the potential of combining homogenization with DCT as an effective approach to enhance the structural integrity and mechanical performance of L-DED-fabricated D2 tool steel, offering valuable insights for the development of advanced manufacturing processes for alloyed materials.

Imperfect refractive index matching in scanning laser optical tomography and a method for digital correction.

Scanning laser optical tomography (SLOT) is a volumetric multi-modal imaging technique that is comparable to optical projection tomography and computer tomography. Image quality is crucially dependent on matching the refractive indexes (RIs) of the sample and surrounding medium, but RI matching often requires some effort and is never perfect. Reducing the burden of RI matching between the immersion medium and sample in biomedical imaging is a challenging and interesting task. We aim at implementing a post processing strategy for correcting SLOT measurements that have errors caused by RI mismatch. To better understand the problems with poorly matched Ris, simulated SLOT measurements with imperfect RI matching of the sample and medium are performed and presented here. A method to correct distorted measurements was developed and is presented and evaluated. This method is then applied to a sample containing fluorescent polystyrene beads and a sample made of olydimethylsiloxane with embedded fluorescent nanoparticles. From the simulations, it is evident that measurements with an RI mismatch larger than 0.02 and no correction yield considerably worse results compared to perfectly matched measurements. RI mismatches larger than 0.05 make it almost impossible to resolve finer details and structures. By contrast, the simulations imply that a measurement with an RI mismatch of up to 0.1 can still yield reasonable results if the presented correction method is applied. The experiments validate the simulated results for an RI mismatch of about 0.09. The method significantly improves the SLOT image quality for samples with imperfectly matched Ris. Although the absolutely best imaging quality will be achieved with perfect RI matching, these results pave the way for imaging in SLOT with RI mismatches while maintaining high image quality.

Spectral–Spatial transformer-based semantic segmentation for large-scale mapping of individual date palm trees using very high-resolution satellite data

Date palm plantations in the United Arab Emirates (UAE) are under threat from soil salinity, drought, and date palm weevils. Accordingly, monitoring and conserving date palms are crucial to preserving a vital component of the country’s agricultural heritage, economy, food security, and ecological balance. Previous studies have effectively identified date palm trees using RGB-based aerial and UAV imagery utilizing diverse deep learning methods. However, the utilization of very high-resolution satellite data for delineating individual date palm crowns remains unexplored due to the limited spatial resolution capabilities of existing satellite systems. This study primarily aimed to achieve precise and comprehensive mapping of date palm trees using WorldView-3 (WV-3) satellite data by leveraging the high representational power of the state-of-the-art vision transformers (ViT) in capturing global information from the input data. First, an in-depth analysis assessment of the various transformer-based semantic segmentation architectures, including UperNet with vision transformer and Swin transformer, SegFormer, Mask2Former, and UniFormer, was conducted. Second, the integration of spectral data on the performance of ViTs was evaluated. Moreover, the models’ generalizability and complexity effect on the segmentation effectiveness were assessed. Accordingly, a postprocessing strategy was developed to aid in delineating and counting date palm trees from semantic segmentation outputs. Results demonstrated that integration of WV-3 spectral data into the analysis resulted in a marked improvement in segmentation quality. The UniFormer, UperNet-Swin, and Mask2Former models demonstrated considerable improvements in multispectral data analysis, with increases in mean intersection over union (mIoU) of 2.17% (77.88% mIoU, 86.01% mean F-score [mF-score]), 2% (78.10% mIoU, 86.18% mF-score), and 1.15% (77.36% mIoU, 85.59% mF-score), respectively, compared with their RGB-based results. Evaluations of model transferability also indicated that Mask2Former, UniFormer, and UperNet-Swin transformers efficiently adapted to multispectral data in the Dibba region. These models achieved mIoU scores of 84.36%, 84.25%, and 83.17% and mF-scores of 90.95%, 90.87%, and 90.13%, highlighting their effectiveness and potential for broader regional application. This research highlights the efficacy and feasibility of using ViTs with WV-3 multispectral data for accurate and comprehensive surveying of date palm plantations, enabling the development of palm tree inventories and continuously updating geospatial databases.

Time–Frequency-Multisqueezing Transform

In the real world, most signals encountered are nonstationary. It is essential to extract a time-frequency (TF) characteristics in such signals for an accurate description. Two parameters are usually applied to quantify the TF characteristics of a nonstationary signal, i.e., instantaneous frequency (IF) and group delay (GD). A post-processing strategy was adopted by two recently developed techniques, the synchrosqueezing transform (SST) and the time-reassigned SST (TSST) to accurately capture the change rules of IF and GD respectively. However, due to the diversity of modes in complex nonstationary signals, no existing technique has been used to effectively estimate both IF and GD simultaneously. To solve this problem, a post-processing analysis technique termed as time-frequency-multisqueezing transform (TFMST) is proposed in this paper where a so-called chirp rate (CR) discrimination criterion is established by considering the Gaussian window in the short-time Fourier transform. The proposed method can accurately categorize nonstationary signals containing harmonic- and impulsive-like components to achieve a concurrent description and ensure the recovery of original signals. The proposed method is validated by numerical simulation and real signal analyses.

Coronagraphic Data Post-processing Using Projections on Instrumental Modes

Directly observing exoplanets with coronagraphs is impeded by the presence of speckles from aberrations in the optical path, which can be mitigated in hardware with wave front control, as well as in post-processing. This work explores using an instrument model in post-processing to separate astrophysical signals from residual aberrations in coronagraphic data. The effect of wave front error (WFE) on the coronagraphic intensity consists of a linear contribution and a quadratic contribution. When either of the terms is much larger than the other, the instrument response can be approximated by a transfer matrix mapping WFE to detector plane intensity. From this transfer matrix, a useful projection onto instrumental modes that removes the dominant error modes can be derived. We apply this approach to synthetically generated Roman Space Telescope hybrid Lyot coronagraph data to extract “robust observables,” which can be used instead of raw data for applications such as detection testing. The projection improves planet flux ratio detection limits by about 28% in the linear regime and by over a factor of 2 in the quadratic regime, illustrating that robust observables can increase sensitivity to astrophysical signals and improve the scientific yield from coronagraphic data. While this approach does not require additional information such as observations of reference stars or modulations of a deformable mirror, it can and should be combined with these other techniques, acting as a model-informed prior in an overall post-processing strategy.

An adaptable rotated bounding box method for automatic detection of arbitrary-oriented cracks

Concrete crack detection is a crucial task for the safety and durability of engineering structures. Extensive research has been conducted on deep-learning methods employing horizontal bounding boxes (HBBs) for crack detection. However, due to the inherently random distribution of concrete cracks, HBB-based methods often produce excessive overlaps and encompass extensive background regions, obstructing the effective interpretation and adaptation of the detection results. To address this issue and achieve efficient utilization of bounding box space for detecting cracks at any orientation, a rotated bounding box (RBB)-based method, that is, Rotated Faster R-CNN with a post processing strategy (RFR-P), was proposed. To realize this method, an RBB-based crack annotation strategy was introduced to standardize the annotation baseline for the evolutionarily established RBB-based crack detection dataset. Then, an RBB-based post-processing strategy was inventively developed to quantify the patterns of cracks with their corresponding rotation angles encompassing longitudinal cracks, transverse cracks, and diagonal cracks. Subsequently, experimental results showed that the RFR-P method provides more reasonable and elaborate detection results in terms of crack distribution patterns when compared to HBB-based methods. Based on the comprehensive consideration of evaluation metrics and detected results, it can be concluded that the RFR-P is aptly designed for detecting cracks at any rotation angle with relatively high accuracy. Finally, an RBB-based concrete crack detection platform was established to automatically detect in situ concrete bridge cracks for real-world applications. The proposed RFR-P model introduces a new perspective on crack detection methods and offers practical references for structural condition evaluation.

Valorization and biorefinery of local agricultural and textile wastes through mycelium composites for structural applications

Abstract Manufacturing of mycelium-based composites is an emerging biorefinery technology toward the development of environmentally positive materials within the circular economy: it benefits from waste and industrial by-products upcycling while excelling in biodegradability. This study investigates the compressive behavior of materials repurposed from local agricultural wastes (tree nuts and crop wastes in California’s Central Valley), using the fungal mycelium of Pleurotus ostreatus and Ganoderma lucidum, well-known edible and medicinal species. We also explore the hybridization of these mycelium-based composites with local textile waste fibers as reinforcements. Following guidelines from several ASTM standards, the compressive behavior of these composites is analyzed to determine the impact of biomass processing, composition, fungal species used, and post-processing strategy. We propose a post-processing strategy based on a short exposure to sodium chloride solutions in ambient conditions, to de-activate mycelium and prevent its fruiting, replacing the established energy-intensive heat-based post-processing. This work aims at contributing to the decarbonization of the built environment and the construction industry in particular, through materials designed with upcycled waste (agricultural and textile), fungal mycelium and low-carbon footprint processes.

Estimating Per-Class Statistics for Label Noise Learning.

Real-world data may contain a considerable amount of noisily labeled examples, which usually mislead the training algorithm and result in degraded classification performance on test data. Therefore, Label Noise Learning (LNL) was proposed, of which one popular research trend focused on estimating the critical statistics (e.g., sample mean and sample covariance), to recover the clean data distribution. However, existing methods may suffer from the unreliable sample selection process or can hardly be applied to multi-class cases. Inspired by the centroid estimation theory, we propose Per-Class Statistic Estimation (PCSE), which establishes the quantitative relationship between the clean (first-order and second-order) statistics and the corresponding noisy statistics for every class. This relationship is further utilized to induce a generative classifier for model inference. Unlike existing methods, our approach does not require sample selection from the instance level. Moreover, our PCSE can serve as a general post-processing strategy applicable to various popular networks pre-trained on the noisy dataset for boosting their classification performance. Theoretically, we prove that the estimated statistics converge to their ground-truth values as the sample size increases, even if the label transition matrix is biased. Empirically, we conducted intensive experiments on various binary and multi-class datasets, and the results demonstrate that PCSE achieves more precise statistic estimation as well as higher classification accuracy when compared with state-of-the-art methods in LNL.

A novel post-processing strategy to improve the accuracy of complete-arch intraoral scanning for implants: an in vitro study

ObjectivesTo develop a new post-processing strategy that utilizes an auxiliary device to adjust intraoral scans and improve the accuracy of 3D models of complete-arch dental implants. Materials and methodsAn edentulous resin model with 6 dental implants was prepared. An auxiliary device, consisting of an opaque base and artificial landmarks, was fabricated and mounted onto the resin model. Twenty intraoral scans (raw scans) were taken using this setup. A new post-processing strategy was proposed to adjust the raw scans using reverse engineering software (verified group). Additionally, ten conventional gypsum casts were duplicated and digitized using a laboratory scanner. The linear and angular trueness and precision of the models were evaluated and compared. The effect of the proposed strategy on the accuracy of complete-arch intraoral scans was analyzed using one-way ANOVA. ResultsThe linear trueness (29.7 µm) and precision (24.8 µm) of the verified group were significantly better than the raw scans (46.6 µm, 44.7 µm) and conventional casts (51.3 µm, 36.5 µm), particularly in cross-arch sites. However, the angular trueness (0.114°) and precision (0.085°) of the conventional casts were significantly better than both the verified models (0.298°, 0.168°) and the raw scans (0.288°, 0.202°). ConclusionsThe novel post-processing strategy is effective in enhancing the linear accuracy of complete-arch implant IO scans, especially in cross-arch sites. However, further improvement is needed to eliminate the angular deviations. Clinical significanceErrors generated from intraoral scanning in complete edentulous arches exceed the clinical threshold. The elimination of stitching errors in the raw scans particularly in the cross-arch sites, through the proposed post-processing strategy would enhance the accuracy of complete-arch implant prostheses.

Enhanced strength of additively manufactured Inconel 718 by means of a simplified heat treatment strategy

This study simplified the heat treatment route and reduced the post-processing burden for laser powder bed fusion IN718 (a nickel-based superalloy). The tailored route retained advantageous microstructures and improved tensile strength when compared to the conventional post-processing strategy. The implementation of a single pre-aging step (combining the stress relief, HIP, and solution steps into one) and a single-step aging treatment (as opposed to two-step aging) resulted in a total heat treatment time of 15 h, which is far shorter than the conventional five-step strategy (42 h). The grain structure, pore structure, and room-temperature tensile properties were measured for three material conditions: the as-built condition, the baseline condition (a conventional five-step heat treatment route), and the HIP1020RQSA condition (novel high pressure rapid quenching HIP treatment followed by a single-step age). In the HIP1020RQSA condition, the high pressure (200 MPa), short hold time at a carefully selected temperature (0.5 h at 1020 °C), and rapid quenching (2150 °C/min) were implemented. Unique findings in the microstructure include generation of localized micro-scale recrystallization throughout the retained columnar grain and sub-grain structures, internal pore closure with a shortened HIP soak time, incomplete dissolution of Laves phase, and the absence of δ phase. The single step aging treatment also generated γ′ and γ′′ precipitates. The HIP1020RQSA treatment improved the yield strength of the baseline condition from 1112 MPa to 1209 MPa, decreased total elongation from 24.9% to 23.2%, and still exceeds the minimum Z direction (build direction) tensile properties outlined in ASTM F3055–14a.

FASTER: A Dynamic Fairness-assurance Strategy for Session-based Recommender Systems

When only users’ preferences and interests are considered by a recommendation algorithm, it will lead to the severe long-tail problem over items. Therefore, the unfair exposure phenomenon of recommended items caused by this problem has attracted widespread attention in recent years. For the first time, we reveal the fact that there is a more serious unfair exposure problem in session-based recommender systems (SRSs), which learn the short-term and dynamic preferences of users from anonymous sessions. Considering the fact that in SRSs, recommendations are provided multiple times and item exposures are accumulated over interactions in a session, we define new metrics both for the fairness of item exposure and recommendation quality among sessions. Moreover, we design a dynamic F airness- A ssurance ST rategy for s E ssion-based R ecommender systems ( FASTER ). FASTER is a post-processing strategy that tries to keep a balance between item exposure fairness and recommendation quality. It can also maintain the fairness of recommendation quality among sessions. The effectiveness of FASTER is verified on three real-world datasets and five original algorithms. The experiment results show that FASTER can generally reduce the unfair exposure of different session-based recommendation algorithms while still ensuring a high level of recommendation quality.