High-quality Training Datasets Research Articles

中国河南省2016–2021年尾矿库目标检测数据集

Tailings ponds are essential infrastructures in mining operations, but they also pose a significant risk source as potential sources of manmade debris flow with high potential energy. The object detection of tailings pond in remote sensing imagery is to accurately recognize and pinpoint the locations of them on the images. Compared with traditional methods, tailings pond detection in remote sensing imagery with the aid of deep learning has seen substantial improvements in accuracy, stability, and efficiency especially when trained on high quality training dataset. Based on years of China high-resolution satellite remote sensing images, through data processing, manual interpretation and annotation, image slicing and other steps, we have constructed a dataset of object detection of tailings ponds in Henan Province, China, available for public access. This dataset has the following characteristics: (1) the domestic high-resolution dataset comprising 1,183 slices and 1,728 object instances; (2) multi temporal dataset containing a total of four different years of sample data in 2016, 2018, 2020 and 2021; (3) objects labeled with oriented bounding box, with less image background interference. The dataset can be used for the technical research on the development of tailings pond detection models with deep learning, as well as for the automatic and intelligent detection of tailings pond, It is of great significance for promoting the development of automatic extraction technology and safety supervision of tailings ponds.

Mid-DeepLabv3+: A Novel Approach for Image Semantic Segmentation Applied to African Food Dietary Assessments.

Recent decades have witnessed the development of vision-based dietary assessment (VBDA) systems. These systems generally consist of three main stages: food image analysis, portion estimation, and nutrient derivation. The effectiveness of the initial step is highly dependent on the use of accurate segmentation and image recognition models and the availability of high-quality training datasets. Food image segmentation still faces various challenges, and most existing research focuses mainly on Asian and Western food images. For this reason, this study is based on food images from sub-Saharan Africa, which pose their own problems, such as inter-class similarity and dishes with mixed-class food. This work focuses on the first stage of VBDAs, where we introduce two notable contributions. Firstly, we propose mid-DeepLabv3+, an enhanced food image segmentation model based on DeepLabv3+ with a ResNet50 backbone. Our approach involves adding a middle layer in the decoder path and SimAM after each extracted backbone feature layer. Secondly, we present CamerFood10, the first food image dataset specifically designed for sub-Saharan African food segmentation. It includes 10 classes of the most consumed food items in Cameroon. On our dataset, mid-DeepLabv3+ outperforms benchmark convolutional neural network models for semantic image segmentation, with an mIoU (mean Intersection over Union) of 65.20%, representing a +10.74% improvement over DeepLabv3+ with the same backbone.

View Synthesis of Dynamic Scenes based on Deep 3D Mask Volume.

Image view synthesis has seen great success in reconstructing photorealistic visuals, thanks to deep learning and various novel representations. The next key step in immersive virtual experiences is view synthesis of dynamic scenes. However, several challenges exist due to the lack of high-quality training datasets, and the additional time dimension for videos of dynamic scenes. To address this issue, we introduce a multi-view video dataset, captured with a custom 10-camera rig in 120FPS. The dataset contains 96 high-quality scenes showing various visual effects and human interactions in outdoor scenes. We develop a new algorithm, Deep 3D Mask Volume, which enables temporally-stable view extrapolation from binocular videos of dynamic scenes, captured by static cameras. Our algorithm addresses the temporal inconsistency of disocclusions by identifying the error-prone areas with a 3D mask volume, and replaces them with static background observed throughout the video. Our method enables manipulation in 3D space as opposed to simple 2D masks, We demonstrate better temporal stability than frame-by-frame static view synthesis methods, or those that use 2D masks. The resulting view synthesis videos show minimal flickering artifacts and allow for larger translational movements.

Simultaneous denoising and reconstruction of distributed acoustic sensing seismic data via a multicascade deep-learning method

Distributed acoustic sensing (DAS) has been moderately used in seismic exploration as a consequence of its advantages, such as low cost, high resolution, high sensitivity, and strong resistance to high temperatures. Nonetheless, seismic data recorded by the DAS system often are seriously contaminated by a massive amount of background noise and recording gaps, which also decrease the quality of DAS seismic data. Building on its success in numerous fields of data processing, convolutional neural networks (CNNs) also have been applied to seismic data denoising as well as reconstruction and achieved good performance compared with traditional methods. However, relatively few CNN-based studies have been done on the simultaneous denoising and reconstruction of seismic data, especially for the complex DAS seismic data with a low signal-to-noise ratio. In this study, we have developed a multicascade network structure in combination with a channel attention mechanism (MCA-Net) to achieve the simultaneous denoising and reconstruction of DAS seismic data by only using a unified trained model. Particularly, our MCA-Net consists of three branches of cascade networks that can extract abundant features from DAS seismic data with different resolutions. Furthermore, we deploy some channel attention modules and aggressive blocks to MCA-Net to fuse these aforementioned multiresolution features and thus improving the feature extraction ability of the whole network. Meanwhile, a high-quality training data set composed of paired clean complete and noisy incomplete patches is used to train MCA-Net in a supervised fashion. Some synthetic and real DAS seismic data are used to demonstrate the effectiveness of MCA-Net, and our method exhibits better denoising and reconstruction performance than some traditional methods and two existing CNN architectures.

Image-Based Ship Detection Using Deep Variational Information Bottleneck.

Image-based ship detection is a critical function in maritime security. However, lacking high-quality training datasets makes it challenging to train a robust supervision deep learning model. Conventional methods use data augmentation to increase training samples. This approach is not robust because the data augmentation may not present a complex background or occlusion well. This paper proposes to use an information bottleneck and a reparameterization trick to address the challenge. The information bottleneck learns features that focus only on the object and eliminate all backgrounds. It helps to avoid background variance. In addition, the reparameterization introduces uncertainty during the training phase. It helps to learn more robust detectors. Comprehensive experiments show that the proposed method outperforms conventional methods on Seaship datasets, especially when the number of training samples is small. In addition, this paper discusses how to integrate the information bottleneck and the reparameterization into well-known object detection frameworks efficiently.

Embracing limited and imperfect training datasets: opportunities and challenges in plant disease recognition using deep learning.

Recent advancements in deep learning have brought significant improvements to plant disease recognition. However, achieving satisfactory performance often requires high-quality training datasets, which are challenging and expensive to collect. Consequently, the practical application of current deep learning-based methods in real-world scenarios is hindered by the scarcity of high-quality datasets. In this paper, we argue that embracing poor datasets is viable and aims to explicitly define the challenges associated with using these datasets. To delve into this topic, we analyze the characteristics of high-quality datasets, namely, large-scale images and desired annotation, and contrast them with the limited and imperfect nature of poor datasets. Challenges arise when the training datasets deviate from these characteristics. To provide a comprehensive understanding, we propose a novel and informative taxonomy that categorizes these challenges. Furthermore, we offer a brief overview of existing studies and approaches that address these challenges. We point out that our paper sheds light on the importance of embracing poor datasets, enhances the understanding of the associated challenges, and contributes to the ambitious objective of deploying deep learning in real-world applications. To facilitate the progress, we finally describe several outstanding questions and point out potential future directions. Although our primary focus is on plant disease recognition, we emphasize that the principles of embracing and analyzing poor datasets are applicable to a wider range of domains, including agriculture. Our project is public available at https://github.com/xml94/EmbracingLimitedImperfectTrainingDatasets.

A quantile-regression physics-informed deep learning for car-following model

Car-following models have been extensively studied using either theory-driven models or data-driven models. On the other hand, theory-driven models have low accuracy, whereas data-driven models have poor interpretability and require a high-quality training dataset. Furthermore, the majority of studies focused on predicting short-term trajectory based on a single vehicle. This paper aims to combine physics knowledge with deep learning and stochasticity to completely capture the genuine regularities of car-following behaviors, leveraging the benefits of both physics-based (interpretability) and deep learning-based (data-efficient and generalizability) models. We design a quantile-regression physics-informed deep learning car-following model (QRPIDL-CF). The physics-informed computational graph is integrated with a long short-term memory (LSTM) architecture (abbreviated as PIDL). By incorporating the quantile regression to the loss function of the PIDL model, stochastic behaviors can be produced without prior assumptions of drivers. To validate the proposed model, data from a 25-vehicle platoon experiment was used. The prediction results prove that given the first vehicle’s trajectory and other vehicles’ initial states, it can preserve the high precision in learning car-following behaviors and introduce small disturbances into car-following in a vehicle-platoon just the same as the real driving process. The trajectory prediction of the whole platoon shows that the proposed hybrid QRPIDL-CF model can better perform the macro-traffic behavior of car-following than every single model from the hybrid QRPIDL-CF model.

Development of taxonomy for classifying defect patterns on wafer bin map using Bin2Vec and clustering methods

A wafer consists of several chips, and serial electrical tests are conducted for each chip to investigate whether the chip is defective. A bin indicates the test results for each chip with information on which tests the chip failed. A wafer bin map (WBM) shows the locations and bins of the defects on the wafer. WBMs showing spatial patterns of defects usually result from assignable causes in the wafer fabrication process; hence, they should be classified in advance. The existing defect-pattern taxonomies do not consider bins, although useful information can be obtained from them. We propose a taxonomy that consists of the shape, size, location, and bin dimensions. The bin dimension is developed using Bin2Vec method, which determines RGB (red-green-blue) code for each bin according to the spatial similarity between bins. Three levels of the bin dimension are defined by analyzing a large number of WBMs using Bin2Vec and clustering methods. Compared with the existing taxonomies, the proposed taxonomy has the advantage of identifying major bins of defect patterns, new defect patterns, and non-critical defect patterns. A high-quality training dataset was obtained using the proposed taxonomy; consequently, a defect pattern classification model with satisfactory classification performance could be obtained.

Terabyte-scale supervised 3D training and benchmarking dataset of the mouse kidney

The performance of machine learning algorithms, when used for segmenting 3D biomedical images, does not reach the level expected based on results achieved with 2D photos. This may be explained by the comparative lack of high-volume, high-quality training datasets, which require state-of-the-art imaging facilities, domain experts for annotation and large computational and personal resources. The HR-Kidney dataset presented in this work bridges this gap by providing 1.7 TB of artefact-corrected synchrotron radiation-based X-ray phase-contrast microtomography images of whole mouse kidneys and validated segmentations of 33 729 glomeruli, which corresponds to a one to two orders of magnitude increase over currently available biomedical datasets. The image sets also contain the underlying raw data, threshold- and morphology-based semi-automatic segmentations of renal vasculature and uriniferous tubules, as well as true 3D manual annotations. We therewith provide a broad basis for the scientific community to build upon and expand in the fields of image processing, data augmentation and machine learning, in particular unsupervised and semi-supervised learning investigations, as well as transfer learning and generative adversarial networks.

An optimized GATK4 pipeline for Plasmodium falciparum whole genome sequencing variant calling and analysis

BackgroundAccurate variant calls from whole genome sequencing (WGS) of Plasmodium falciparum infections are crucial in malaria population genomics. Here a falciparum variant calling pipeline based on GATK version 4 (GATK4) was optimized and applied to 6626 public Illumina WGS samples.MethodsControl WGS and accurate PacBio assemblies of 10 laboratory strains were leveraged to optimize parameters that control the heterozygosity, local assembly region size, ploidy, mapping and base quality in both GATK HaplotypeCaller and GenotypeGVCFs. From these controls, a high-quality training dataset was generated to recalibrate the raw variant data.ResultsOn current high-quality samples (read length = 250 bp, insert size = 405–524 bp), the optimized pipeline shows improved sensitivity (86.6 ± 1.7% for SNPs and 82.2 ± 5.9% for indels) compared to the default GATK4 pipeline (77.7 ± 1.3% for SNPs; and 73.1 ± 5.1% for indels, adjusted P < 0.001) and previous variant calling with GATK version 3 (GATK3, 70.3 ± 3.0% for SNPs and 59.7 ± 5.8% for indels, adjusted P < 0.001). Its sensitivity on simulated mixed infection samples (80.8 ± 6.1% for SNPs and 78.3 ± 5.1% for indels) was again improved relative to default GATK4 (68.8 ± 6.0% for SNPs and 38.9 ± 0.7% for indels, adjusted, adjusted P < 0.001). Precision was high and comparable across all pipelines on each type of data tested. The resulting combination of high-quality SNPs and indels increases the resolution of local population population structure detection in sub-Saharan Africa. Finally, increasing ploidy improves the detection of drug resistance mutations and estimation of complexity of infection.ConclusionsOverall, this study provides an optimized falciparum GATK4 pipeline resource for variant calling which should help improve genomic studies of malaria.

Automated Pavement Crack Detection with Deep Learning Methods: What Are the Main Factors and How to Improve the Performance?

Recently, deep learning (DL) methods have been intensively studied in pavement distress detection research. To build an effective yet robust DL-based distress detection model, both the training data and DL model structure need to be carefully prepared. This paper discusses the main influencing factors that affect the performance of DL models in pavement distress detection and aims to explore possible ways to improve the model performance. An open-source pavement image dataset and three promising object detection methods were selected for model development and evaluation. A series of experiments were conducted with different settings of model structures and data manipulation. According to the experiment results, selection of detection methods, data size, and annotation consistency have significant impact on detection performance. DL models YOLO and Faster R-CNN were found to yield the same level of performance, while CenterNet notably underperformed on the test dataset. Distress classes with a larger number of instances were found to be able to yield better performances. With the new annotation data carefully marked up and verified by experienced pavement raters, the accuracy score using the F1 index was found to be 0.84, which was a very sound improvement compared with the 0.7 of the original annotation. Data augmentation methods were found to be able to significantly improve the detection accuracy. Based on the experiment results, we believe the focus should be placed on the development of high-quality training datasets and thin object detection methods in the future.

Machine learning for predicting the viscosity of binary liquid mixtures

Viscosity is an important parameter in process engineering and is a key design objective for application areas including the coatings, lubricants, personal care, and pharmaceutical industries. The lack of reliable and general methods for predicting the viscosities of mixtures creates a barrier for modern process engineering and product design. In this work, we developed a graph-based neural network architecture and applied it to the problem of predicting the viscosity of binary liquid mixtures as a function of composition and temperature. To obtain a high-quality training dataset, we also developed an automated curation pipeline and applied it to a large dataset collected from the literature by the National Institute of Standards and Technology (NIST) to be used as training data. The resulting model predicts viscosity with an MAE of 0.043 and an RMSE of 0.080 in log cP units (base 10). To improve the dependability of the model, we developed a classifier that evaluated the reliability of a prediction based on the variance between an ensemble of models. Using this approach, the model had an effective MAE of 0.029 and RMSE of 0.047 for predictions that were assessed as reliable (80% of the test set). Overall, this work provides 1) a large set of curated viscosity data that can be used for future machine learning efforts, 2) a new, graph-based deep learning approach for predicting the viscosity of binary mixtures, and 3) an illustrative case study for how deep learning can be used for accurate and reliable property prediction.

SCMS-Net: Self-Supervised Clustering-Based 3D Meshes Segmentation Network

The superior performance of deep learning in different domains has sparked significant interest in its applicability to 3D computer graphics. Deep learning has become the dominant technical architecture in current 3D mesh segmentation. However, learning-based 3D segmentation methods usually rely on high-quality training datasets, which are not readily available in practical applications. How to segment 3D meshes without exhaustive label annotations remains a challenging problem, especially in the context of deep learning. As a subset of unsupervised learning methods, self-supervised learning offers a promising learning paradigm for unlabeled 3D mesh segmentation. In this paper, we introduce a self-supervised clustering-Based network specifically for the segmentation of label-free 3D meshes. Our self-supervised clustering-based 3D mesh segmentation network (SCMS-Net) employs a two-branch architecture to learn effective feature representation. The two branches are unified into an end-to-end framework using a self-supervised strategy. Finally, the label predictions of the parts are generated by iterative clustering. We conducted ablation studies and comparative experiments on a standard benchmark to demonstrate the effectiveness of our approach.

Machine learning-assisted processing workflow for multi-fiber DAS microseismic data

In recent years, Distributed Acoustic Sensing (DAS) deployed in deviated wells has been increasingly used for microseismic monitoring. DAS can provide observations of microseismic wavefields with high spatial resolution and wide aperture, at the cost of unusually large data volumes compared with conventional downhole microseismic monitoring. To tackle this big-data challenge, we have developed key elements of a processing workflow that is assisted by machine learning techniques. We trained a convolutional neural network (CNN) for event detection and a U-Net model for both P- and S-wave arrival time picking. The workflow was applied to two multiwell DAS datasets acquired during hydraulic fracturing completions in western Canada. These datasets also include co-located 3C borehole geophone arrays that enable further comparison between catalogs from both sensor types. Compared with a traditional short-term average/long-term average (STA/LTA) method for event detection, our results indicate that the CNN method has a lower false-trigger rate and increases the event catalog size by a factor of 2.6–5.6. U-Net yields arrival-time picks with relatively small errors, high efficiency, and minimal user intervention, providing hypocenter location and focal depth that is arguably more accurate than the geophone catalog. While the proposed automated workflow requires substantial effort to build high-quality and large training datasets, it enables the use of DAS for real-time seismicity monitoring and risk management after the training stage. Although the DAS system detected fewer events than the geophone catalog and missed smaller magnitude events, our results indicate that fiber-optic sensors provide enough sensitivity to detect and locate sufficient events to characterize the treatment stages. DAS also captured induced events located at a hypocentral distance of &amp;gt;1 km, which are possibly indicative of reactivation of structural features.

Joint physics-based and data-driven time-lapse seismic inversion: Mitigating data scarcity

In carbon capture and sequestration, developing rapid and effective imaging techniques is crucial for real-time monitoring of the spatial and temporal dynamics of [Formula: see text] propagation during and after injection. With continuing improvements in computational power and data storage, data-driven techniques based on machine learning (ML) have been effectively applied to seismic inverse problems. In particular, ML helps alleviate the ill-posedness and high computational cost of full-waveform inversion (FWI). However, such data-driven inversion techniques require massive high-quality training data sets to ensure prediction accuracy, which hinders their application to time-lapse monitoring of [Formula: see text] sequestration. We develop an efficient “hybrid” time-lapse workflow that combines physics-based FWI and data-driven ML inversion. The scarcity of the available training data is addressed by developing a new data-generation technique with physics constraints. The method is validated using a synthetic [Formula: see text]-sequestration model based on the Kimberlina storage reservoir in California. Our approach is shown to synthesize a large volume of high-quality, physically realistic training data, which is critically important in accurately characterizing the [Formula: see text] movement in the reservoir. The developed hybrid methodology can also simultaneously predict the variations in velocity and saturation and achieve high spatial resolution in the presence of realistic noise in the data.

Assisted Cement Log Interpretation Using Machine Learning

Summary The Assisted Cement Log Interpretation Project has used machine learning (ML) to create a tool that interprets cement logs by predicting a predefined set of annular condition codes used in the cement log interpretation process. The development of a cement log interpretation tool speeds up the log interpretation process and enables expert knowledge to be efficiently shared when training new professionals. By using high-quality and consistent training data sets, the project has trained a model that will support unbiased and consistent interpretations over time. The tool consists of a training and a prediction tool integrated with cased-hole logging interpretation software. By containerizing the code using an “API First” design principle (API: application programming interface), the applicability of this add-on tool is broad. The ML model is trained using selected and engineered features from cement logs, and the tool predicts an annular condition code according to the cement classification system for each depth segment in the log. The interpreters can easily fetch a complete cement log interpretation prediction for the log and use that as a template for their final interpretation. The ML model can easily be retrained with new data sets to improve accuracy even further. To improve cement log interpretation consistency in the industry, the code will be made available as open source.

Transfer learning of long short-term memory analysis in significant wave height prediction off the coast of western Tohoku, Japan

A large amount of high quality training datasets is usually required for machine learning to predict wave conditions with high precision. However, it is not always easy to obtain such sufficient training datasets, especially on offshore new construction locations. This study evaluated significant wave height (SWH) prediction by applying transfer learning into long short-term memory (LSTM) with one month datasets and pre-trained layers from the nearest ports. SWH was predicted for lead times of 6-, 12-, and 24-h at two locations off the coast of the western Tohoku, Japan. The results showed that the application of transfer learning could improve the accuracy of SWH prediction. For example, the coefficient of determination (R2) of the 6-h lead time SWH prediction off the coasts of Akita and Yamagata improved from 0.685 to 0.807 and from 0.710 to 0.850, respectively. Its accuracy was nearly equivalent to that of previous studies employing LSTM networks trained by almost ten years datasets. Furthermore, the results showed that transferring all layers improved the accuracy compared to transferring a part of the hidden or output layers. We believe the results of this study demonstrate that transfer learning has a superior SWH prediction ability with limited training data.

PET Image Reconstruction Incorporating Deep Image Prior and a Forward Projection Model

Convolutional neural networks (CNNs) have recently achieved remarkable performance in positron emission tomography (PET) image reconstruction. In particular, CNN-based direct PET image reconstruction, which directly generates the reconstructed image from the sinogram, has potential applicability to PET image enhancements because it does not require image reconstruction algorithms, which often produce some artifacts. However, these deep learning-based, direct PET image reconstruction algorithms have the disadvantage that they require a large number of high-quality training datasets. In this study, we propose an unsupervised direct PET image reconstruction method that incorporates a deep image prior framework. Our proposed method incorporates a forward projection model with a loss function to achieve unsupervised direct PET image reconstruction from sinograms. To compare our proposed direct reconstruction method with the filtered back projection (FBP) and maximum likelihood expectation maximization (ML-EM) algorithms, we evaluated using Monte Carlo simulation data of brain [$^{18}$F]FDG PET scans. The results demonstrate that our proposed direct reconstruction quantitatively and qualitatively outperforms the FBP and ML-EM algorithms with respect to peak signal-to-noise ratio and structural similarity index.

A full resolution deep learning network for paddy rice mapping using Landsat data

Rice is the most important food crop in the developing world, and more than half of the global population consumes it as a staple food. Mapping the area of rice cultivation in a timely and accurate manner is essential to ensure food safety and evaluate its environmental impact. Deep learning performs very well in high-resolution remote sensing classification; however, due to lack of high-quality training datasets and spatial semantic information of Landsat data, paddy rice mapping based on deep learning and Landsat data has received less attention. In this study, we constructed the first large-scale training dataset and a deep learning network, named full resolution network (FR-Net), for mapping paddy rice based on Landsat 8 OLI data. The pixel-wise annotated dataset is composed of 64 Landsat 8 OLI scenes covering the main areas producing rice in northeast China. To overcome the coarse segmentation borders resulting from other deep learning models, and especially with the low-resolution Landsat data, a new multi-resolution fusion unit (MRFU) was proposed to fuse different resolution streams and maintain the high-resolution streams of the model. In comparison with other models, the FR-Net acquired the highest accuracy, with MCC of 0.893 and F1 score of 0.898, and in particular, the results of different band combinations showed that the FR-Net perform better for feature extraction than other models. Without using the sensitive bands, i.e., short-wave infrared 1 and 2, the MCC of FR-Net was slightly decreased by 4.14%, smaller than other models, e.g., 8.51% for U-Net and 11.76% for RS-Net etc. The comparison showed the proposed FR-Net presented highest spatial precision among all these models, by which a fine boundary could be obtained for Landsat data to achieve high quality mapping result. The analysis of the performance in different temporal showed that FR-Net performed better at the beginning or the ending of growth stages, and the omission or false alarm main occurred in small size patches or the boundaries. In all, low-resolution characteristics of Landsat data should be more extensively investigated when developing and using deep learning in classification.

NMR-guided directed evolution.

Directed evolution is a powerful tool for improving existing properties and imparting completely new functionalities to proteins1-4. Nonetheless, its potential ineven small proteins is inherently limited by the astronomical number of possible amino acid sequences. Sampling the complete sequence space of a 100-residue protein would require testing of 20100 combinations, which is beyond any existing experimental approach. In practice, selective modification of relatively few residues is sufficient for efficient improvement, functional enhancement and repurposing of existing proteins5. Moreover, computational methods have been developed to predict the locations and, in certain cases, identities of potentially productive mutations6-9. Importantly, all current approaches for prediction of hot spots and productive mutations rely heavily on structural information and/or bioinformatics, which is not always available for proteins of interest. Moreover, they offer a limited ability to identify beneficial mutations far from the active site, even though such changes may markedly improve the catalytic properties of an enzyme10. Machine learning methods have recently showed promise in predicting productive mutations11, but they frequently require large, high-quality training datasets, which are difficult to obtain in directed evolution experiments. Here we show that mutagenic hot spots in enzymes can be identified using NMR spectroscopy. In a proof-of-concept study, we converted myoglobin, a non-enzymatic oxygen storage protein, into a highly efficient Kemp eliminase using only three mutations. The observed levels of catalytic efficiency exceed those of proteins designed using current approaches and are similar with those of natural enzymes for the reactions that they are evolved to catalyse. Given the simplicity of this experimental approach, which requires no a priori structural or bioinformatic knowledge, we expect it to be widely applicable and to enable the full potential of directed enzyme evolution.