Meta-learning Algorithm Research Articles

Enhanced landslide susceptibility mapping in data-scarce regions via unsupervised few-shot learning

Given the critical need to assess landslide hazards, producing landslide susceptibility map (LSM) in regions with scarce historical landslide inventories poses significant challenges. This study introduces a novel landslide susceptibility assessment framework that combines unsupervised learning strategies with few-shot learning methods to increase the accuracy of LSM in these areas. The framework has been practically validated in a representative geological disaster-prone area along the West-East Gas Pipeline in Shaanxi Province, China. We employed three advanced few-shot learning models: a support vector machine, meta-learning, and transfer learning. These models implement feature representation learning for weakly correlated influencing factors through an unsupervised approach, thereby constructing an effective landslide susceptibility assessment model. We compared traditional learning methods and used the receiver operating characteristic (ROC) curve and SHAP values to quantify the effectiveness of the models. The results indicate that the meta-learning algorithm outperforms both the SVM and transfer learning in areas with limited landslide data. The integration of unsupervised strategies significantly improves performance, achieving area under the curve (AUC) values of 0.9385 and 0.9861, respectively. Compared with using meta-learning alone, incorporating unsupervised learning strategies increased the AUC by 4.76%, enhancing both the predictive power of the model and the interpretability of the features. Meta-learning under unsupervised conditions effectively mitigates the evaluation difficulties caused by insufficient landslide records, providing a viable path and empirical evidence for performance improvement in similar data- scarce regions worldwide.

A Transferable Meta-Learning Phase Prediction Model for High-Entropy Alloys Based on Adaptive Migration Walrus Optimizer

The phases of high-entropy alloys (HEAs) are crucial to their material properties. Although meta-learning can recommend a desirable algorithm for materials designers, it does not utilize the optimal solution information of similar historical problems in the HEA field. To address this issue, a transferable meta-learning model (MTL-AMWO) based on an adaptive migration walrus optimizer is proposed to predict the phases of HEAs. Firstly, a transferable meta-learning algorithm frame is proposed, which consists of meta-learning based on adaptive migration walrus optimizer, balanced-relative density peaks clustering, and transfer strategy. Secondly, an adaptive migration walrus optimizer model is proposed, which adaptively migrates walruses according to the changes in the average fitness value of the population over multiple iterations. Thirdly, balanced-relative density peaks clustering is proposed to cluster the samples in the source and target domains into several clusters with similar distributions, respectively. Finally, the transfer strategy adopts the maximum mean discrepancy to find the most matching historical problem and transfer its optimal solution information to the target domain. The effectiveness of MTL-AMWO is validated on 986 samples from six datasets, including 323 quinary HEAs, 366 senary HEAs, and 297 septenary HEAs. The experimental results show that the MTL-AMWO achieves better performance than other algorithms.

Graph neural processes for molecules: an evaluation on docking scores and strategies to improve generalization

Neural processes (NPs) are models for meta-learning which output uncertainty estimates. So far, most studies of NPs have focused on low-dimensional datasets of highly-correlated tasks. While these homogeneous datasets are useful for benchmarking, they may not be representative of realistic transfer learning. In particular, applications in scientific research may prove especially challenging due to the potential novelty of meta-testing tasks. Molecular property prediction is one such research area that is characterized by sparse datasets of many functions on a shared molecular space. In this paper, we study the application of graph NPs to molecular property prediction with DOCKSTRING, a diverse dataset of docking scores. Graph NPs show competitive performance in few-shot learning tasks relative to supervised learning baselines common in chemoinformatics, as well as alternative techniques for transfer learning and meta-learning. In order to increase meta-generalization to divergent test functions, we propose fine-tuning strategies that adapt the parameters of NPs. We find that adaptation can substantially increase NPs' regression performance while maintaining good calibration of uncertainty estimates. Finally, we present a Bayesian optimization experiment which showcases the potential advantages of NPs over Gaussian processes in iterative screening. Overall, our results suggest that NPs on molecular graphs hold great potential for molecular property prediction in the low-data setting.Scientific contributionNeural processes are a family of meta-learning algorithms which deal with data scarcity by transferring information across tasks and making probabilistic predictions. We evaluate their performance on regression and optimization molecular tasks using docking scores, finding them to outperform classical single-task and transfer-learning models. We examine the issue of generalization to divergent test tasks, which is a general concern of meta-learning algorithms in science, and propose strategies to alleviate it.

Efficient data-driven nonlinear system identification for structural health monitoring: A proof-of-principle study

Structural health monitoring (SHM) is the process of detecting damage in structures. Typically, the damage-sensitive feature identification is based on fitting some models to the measured system response data. The parameters of these models or the prediction errors associated with these models then become the desired damage-sensitive features. Real-world structures usually exhibit significant nonlinear behaviors. Data-driven nonlinear system identification (NSI) is essential to obtain accurate models of structural dynamics for reliable damage detection. However, on the one hand, the data acquisition of real-world structures for damage detection is usually difficult or scarce. On the other hand, NSI generally needs to solve a nonlinear optimization problem that requires more computational time for convergence than linear system identification. Data efficiency and computational efficiency become two critical challenges in data-driven NSI. To address these challenges, this work presents a proof-of-principle study on developing a meta learning-based method for efficient data-driven NSI. Specifically, the meta learning algorithm leverages a previously collected database of similar but different systems to learn the meta-knowledge about how to identify a new system, enabling a fast identification/modeling with limited data. Numerical simulations are performed to validate the method on fundamental nonlinear dynamical systems widely seen in civil engineering. It is observed that the presented meta learning-based method outperforms the conventional method in terms of both data and computational efficiency. Limitations of this work are further discussed.

Style Variable and Irrelevant Learning for Generalizable Person Re-identification

Domain generalization person re-identification (DG-ReID) has gained much attention recently due to the poor performance of supervised re-identification on unseen domains. The goal of domain generalization is to develop a model that is insensitive to domain bias and can perform well across different domains. In this article, We conduct experiments to verify the importance of style factors in domain bias. Specifically, the experiments are to affirm that style bias across different domains significantly contributes to domain bias. Based on this observation, we propose style variable and irrelevant learning (SVIL) to eliminate the influence of style factors on the model. Specifically, we employ a style jitter module (SJM) that enhances the style diversity of a specific source domain and reduces the style differences among various source domains. This allows the model to focus on identity-relevant information and be robust to style changes. We also integrate the SJM module with a meta-learning algorithm to further enhance the model’s generalization ability. Notably, our SJM module is easy to implement and does not add any inference cost. Our extensive experiments demonstrate the effectiveness of our approach, which outperforms existing methods on DG-ReID benchmarks.

A Meta-Learning Approach for Classifying Multimodal Retinal Images of Retinal Vein Occlusion With Limited Data.

To propose and validate a meta-learning approach for detecting retinal vein occlusion (RVO) from multimodal images with only a few samples. In this cross-sectional study, we formulate the problem as meta-learning. The meta-training dataset consists of 1254 color fundus (CF) images from 39 different fundus diseases. Two meta-testing datasets include a public domain dataset and an independent dataset from Kandze Prefecture People's Hospital. The proposed meta-learning models comprise two modules: the feature extraction networks and the prototypical networks (PNs). We use two deep learning models (the ResNet and the Contrastive Language-Image Pre-Training networks [CLIP]) for feature extraction. We evaluate the performance of the algorithms using accuracy, area under the receiver operating characteristic curve (AUCROC), F1-score, and recall. CLIP-based PNs outperform across all meta-testing datasets. For the public APTOS dataset, meta-learning algorithms achieve good results with an accuracy of 86.06% and an AUCROC of 0.87 with only 16 training images. In the hospital datasets, meta-learning algorithms show excellent diagnostic capability for detecting RVO with a very low number of shots (AUCROC above 0.99 for n = 4, 8, and 16, respectively). Notably, even though the meta-training dataset does not include fluorescein angiography (FA) images, meta-learning algorithms also have excellent diagnostic capability for detecting RVO from images with a different modality (AUCROC above 0.93 for n = 4, 8, and 16, respectively). The proposed meta-learning models excel in detecting RVO, not only on CF images but also on FA images from a different imaging modality. The proposed meta-learning models could be useful in automatically detecting RVO on CF and FA images.

Model optimization and acceleration method based on meta-learning and model pruning for laser vision weld tracking system

Purpose This paper aims to propose a lightweight, high-accuracy object detection model designed to enhance seam tracking quality under strong arcs and splashes condition. Simultaneously, the model aims to reduce computational costs. Design/methodology/approach The lightweight model is constructed based on Single Shot Multibox Detector (SSD). First, a neural architecture search method based on meta-learning and genetic algorithm is introduced to optimize pruning strategy, reducing human intervention and improving efficiency. Additionally, the Alternating Direction Method of Multipliers (ADMM) is used to perform structural pruning on SSD, effectively compressing the model with minimal loss of accuracy. Findings Compared to state-of-the-art models, this method better balances feature extraction accuracy and inference speed. Furthermore, seam tracking experiments on this welding robot experimental platform demonstrate that the proposed method exhibits excellent accuracy and robustness in practical applications. Originality/value This paper presents an innovative approach that combines ADMM structural pruning and meta-learning-based neural architecture search to significantly enhance the efficiency and performance of the SSD network. This method reduces computational cost while ensuring high detection accuracy, providing a reliable solution for welding robot laser vision systems in practical applications.

A novel multi-modal fusion method based on uncertainty-guided meta-learning

Multi-modal data fusion for effective feature representation in machine learning is challenging due to intrinsic biases present within and across different modalities. Existing multi-modal data fusion methods often face difficulties in learning generic features due to diverse noise patterns and variations in feature dynamics across different modalities. In this paper, we present a novel method called Uncertainty-guided Meta-Learning Multi-modal Fusion and Classification (UMLMC) to address these challenges. UMLMC dynamically transforms multi-modal feature spaces at both the pre- and post-fusion levels by incorporating uncertainty estimates from an auxiliary network. Our model is optimized using a meta-learning algorithm to enhance its generalization capabilities. Extensive experiments on multi-modal data from diverse domains, along with comparisons to state-of-the-art methods, demonstrate the effectiveness of UMLMC in improving classification performance. These results confirm that UMLMC, with its innovative uncertainty estimation and meta-learning framework, effectively learns informative intra- and inter-modal features, leading to superior classification outcomes.

Machine learning and explainable artificial intelligence for the prevention of waterborne cryptosporidiosis and giardiosis

Cryptosporidium and Giardia are important parasitic protozoa due to their zoonotic potential and impact on human health, and have often caused waterborne outbreaks of disease. Detection of (oo)cysts in water matrices is challenging and extremely costly, thus only few countries have legislated for regular monitoring of drinking water for their presence. Several attempts have been made trying to investigate the association between the presence of such (oo)cysts in waters with other biotic or abiotic factors, with inconclusive findings. In this regard, the aim of this study was the development of an holistic approach leveraging Machine Learning (ML) and eXplainable Artificial Intelligence (XAI) techniques, in order to provide empirical evidence related to the presence and prediction of Cryptosporidium oocysts and Giardia cysts in water samples. To meet this objective, we initially modelled the complex relationship between Cryptosporidium and Giardia (oo)cysts and a set of parasitological, microbiological, physicochemical and meteorological parameters via a model-agnostic meta-learner algorithm that provides flexibility regarding the selection of the ML model executing the fitting task. Based on this generic approach, a set of four well-known ML candidates were, empirically, evaluated in terms of their predictive capabilities. Then, the best-performed algorithms, were further examined through XAI techniques for gaining meaningful insights related to the explainability and interpretability of the derived solutions. The findings reveal that the Random Forest achieves the highest prediction performance when the objective is the prediction of both contamination and contamination intensity with Cryptosporidium oocysts in a given water sample, with meteorological/physicochemical and microbiological markers being informative, respectively. For the prediction of contamination with Giardia, the eXtreme Gradient Boosting with physicochemical parameters was the most efficient algorithm, while, the Support Vector Regression that takes into consideration both microbiological and meteorological markers was more efficient for evaluating the contamination intensity with cysts. The results of the study designate that the adoption of ML and XAI approaches can be considered as a valuable tool for unveiling the complicated correlation of the presence and contamination intensity with these zoonotic parasites that could constitute, in turn, a basis for the development of monitoring platforms and early warning systems for the prevention of waterborne disease outbreaks.

DNA Binding Protein Prediction based on Multi-feature Deep Metatransfer Learning

Background: In recent years, the rapid development of deep learning technology has had a significant impact on the prediction of DNA-binding proteins. Deep neural networks can automatically learn complex features in protein and DNA sequences, improving prediction accuracy and generalization capabilities. Objective: This article mainly establishes a meta-migration model and combines it with a deep learning model to predict DNA-binding proteins. Methods: This study introduces a meta-learning algorithm based on transfer learning, which helps achieve rapid learning and adaptation to new tasks. In addition, normalized Moreau-Broto autocorrelation attributes (NMBAC), position-specific scoring matrix-discrete cosine transform (PSSMDCT), and position-specific scoring matrix-discrete wavelet transform (PSSM-DWT) are also used for feature extraction. Finally, the prediction of DBP is achieved through the deep neural network model based on the attention mechanism. Results: This paper first establishes the basis of deep meta-transfer learning and uses the PDB186 data set as the benchmark to extract features using NMBAC, PSSM-DCT, and PSSM-DWT, respectively, and compare the fused features in pairs, and finally obtain the fused feature process. Through deep learning processing, it is concluded that the fused feature prediction effect is the best. At the same time, compared with the currently popular models, there are obvious improvements in the ACC, MCC, SN and Spec evaluation indicators. Conclusion: Finally, it was concluded that the method used in this article can effectively predict DNA-binding proteins and show more significant performance.

Beamforming optimization of meta-learning algorithm based on gradient descent in RIS-assisted ISAC system

In this paper, we study an integrated sensing and communication (ISAC) system assisted by reconfigurable intelligent surface (RIS). Our goal is to minimize the transmit power of the BS by jointly optimizing the active beamforming and the passive beamforming under the constraints of the signal-to-interference-plus-noise ratio (SINR) of users, the Cramer–Rao bound (CRB) of sensing and the unit-modulus of RIS reflection elements. The optimization problem is difficult to solve because the objective function is non-convex and the variables are coupled. In addition, a large number of base station (BS) antennas lead to high-dimensional transmit beamforming vectors. We first prove that the optimization variables in the optimization problem are low-dimensional coefficient matrices and passive beamforming, rather than active and passive beamforming, which reduces the complexity compared with the original problem. In addition, when the sensing channel space belongs to the communication channel space, the dedicated sensing signal is not needed. Inspired by the meta-learning technology that can solve non-convex problems, we propose a meta-learning algorithm based on gradient descent (MLAGD) to design beamforming, which tends to learn dynamic optimization strategies and iteratively update optimization variables. Specifically, we first introduce Lagrangian multipliers to transform the constrained optimization problem into the Lagrangian dual domain, and the manifold-based optimization is employed to handle the unit-modulus constraint. The meta-learning neural networks are constructed to update optimization variables through local loss function, and then network parameters are updated through global loss function. The numerical results show the effectiveness of the algorithm.

Open DGML: Intrusion Detection Based on Open-Domain Generation Meta-Learning

Network security is crucial for national infrastructure, but the increasing number of network intrusions poses significant challenges. To address this issue, we propose Open DGML, a framework based on open-domain generalization meta-learning for intrusion detection. Our approach incorporates flow imaging, data augmentation, and open-domain generalization meta-learning algorithms. Experimental results on the ISCX2012, NDSec-1, CICIDS2017, and CICIDS2018 datasets demonstrate the effectiveness of Open DGML. Compared to state-of-the-art models (HAST-IDS, CLAIRE, FC-Net), Open DGML achieves higher accuracy and detection rates. In closed-domain settings, it achieves an average accuracy of 96.52% and a detection rate of 97.04%. In open-domain settings, it achieves an average accuracy of 68.73% and a detection rate of 61.49%. These results highlight the superior performance of Open DGML, particularly in open-domain scenarios, for effective identification of various network attacks.

Meta-Embedded Clustering (MEC): A new method for improving clustering quality in unlabeled bird sound datasets

In recent years, ecoacoustics has offered an alternative to traditional biodiversity monitoring techniques with the development of passive acoustic monitoring (PAM) systems allowing, among others, to detect and identify species that are difficult to detect by human observers, automatically. PAM systems typically generate large audio datasets, but using these monitoring techniques to infer ecologically meaningful information remains challenging. In most cases, several thousand hours of recordings need to be manually labeled by experts limiting the operability of the systems. Based on recent developments of meta-learning algorithms and unsupervised learning techniques, we propose here Meta-Embedded Clustering (MEC), a new method with high potential for improving clustering quality in unlabeled bird sound datasets. MEC method is organized in two main steps, with: (a) fine-tuning of a pretrained convolutional neural network (CNN) backbone with different meta-learning algorithms using pseudo-labeled data, and (b) clustering of manually-labeled bird sounds in the latent space based on vector embeddings extracted from the fine-tuned CNN. The MEC method significantly enhanced average clustering performance from less than 1% to more than 80%, greatly outperforming the traditional approach of relying solely on CNN features extracted from a general neotropical audio database. However, this enhanced performance came with the cost of excluding a portion of the data categorized as noise. By improving the quality of clustering in unlabeled bird sound datasets, the MEC method should facilitate the work of ecoacousticians in managing acoustic units of bird song/call clustered according to their similarities, and in identifying potential clusters of species undetected using traditional approaches.

Predicting mixing degree of supersonic flow by a few target data using meta-learning

Predicting the mixing degree of supersonic flow is essential for designing compressible flow devices, such as scramjets, chemical lasers and ejectors. However, building appropriate predictive models of supersonic flow mixing degrees utilizing deep learning is challenging because the samples are expensive from computation or experiments. Herein, a meta-learning method is proposed to reduce the amount of target data required for training a mixing degree estimation by improving the pre-training performance. The datasets of multiple other inlet pressures are used as the pre-training datasets to allow the neural network to generalize to the target inlet pressure. A deep neural network pertained by a model-agnostic meta-learning algorithm estimates mixing degrees from the input information of a jet angle and an inlet temperature. The estimation accuracy is verified using the simulated dataset. Using 5 target data and 10 gradient steps, the meta-learning method achieves the averaged coefficient of determination of 0.95. Moreover, the meta-learning framework achieves mixing degree estimation more quickly and accurately than transfer learning and merged learning. This work provides a way to alleviate the required data usually insufficient for predicting the mixing performance for high-speed flow applications.

Few-Shot Learning for Medical Image Segmentation Using 3D U-Net and Model-Agnostic Meta-Learning (MAML).

Deep learning has attained state-of-the-art results in general image segmentation problems; however, it requires a substantial number of annotated images to achieve the desired outcomes. In the medical field, the availability of annotated images is often limited. To address this challenge, few-shot learning techniques have been successfully adapted to rapidly generalize to new tasks with only a few samples, leveraging prior knowledge. In this paper, we employ a gradient-based method known as Model-Agnostic Meta-Learning (MAML) for medical image segmentation. MAML is a meta-learning algorithm that quickly adapts to new tasks by updating a model's parameters based on a limited set of training samples. Additionally, we use an enhanced 3D U-Net as the foundational network for our models. The enhanced 3D U-Net is a convolutional neural network specifically designed for medical image segmentation. We evaluate our approach on the TotalSegmentator dataset, considering a few annotated images for four tasks: liver, spleen, right kidney, and left kidney. The results demonstrate that our approach facilitates rapid adaptation to new tasks using only a few annotated images. In 10-shot settings, our approach achieved mean dice coefficients of 93.70%, 85.98%, 81.20%, and 89.58% for liver, spleen, right kidney, and left kidney segmentation, respectively. In five-shot sittings, the approach attained mean Dice coefficients of 90.27%, 83.89%, 77.53%, and 87.01% for liver, spleen, right kidney, and left kidney segmentation, respectively. Finally, we assess the effectiveness of our proposed approach on a dataset collected from a local hospital. Employing five-shot sittings, we achieve mean Dice coefficients of 90.62%, 79.86%, 79.87%, and 78.21% for liver, spleen, right kidney, and left kidney segmentation, respectively.

Learning to Diagnose: Meta-Learning for Efficient Adaptation in Few-Shot AIOps Scenarios

With the advancement of technologies like 5G, cloud computing, and microservices, the complexity of network management systems and the variety of technical components have greatly increased. This rise in complexity has rendered traditional operations and maintenance methods inadequate for current monitoring and maintenance demands. Consequently, artificial intelligence for IT operations (AIOps), which harnesses AI and big data technologies, has emerged as a solution. AIOps plays a crucial role in enhancing service quality and customer satisfaction, boosting engineering productivity, and reducing operational costs. This article delves into the primary tasks involved in AIOps, such as anomaly detection, and log fault analysis and classification. A significant challenge identified in many AIOps tasks is the scarcity of fault sample data, indicating a natural alignment of these tasks with few-shot learning. Inspired by model-agnostic meta-learning (MAML), we propose a new anomaly detector, MAML-KAD, for application in various AIOps tasks. Observations confirm that meta-learning algorithms effectively enhance AIOps tasks, showcasing the wide-ranging application prospects of meta-learning algorithms in the field of AIOps. Moreover, we introduced an AIOps platform that embeds meta-learning within its diagnostic core and features streamlined log collection, caching, and alerting to automate the AIOps workflow.

A Meta-Learning Framework for Tuning Parameters of Protection Mechanisms in Trustworthy Federated Learning

Trustworthy federated learning typically leverages protection mechanisms to guarantee privacy. However, protection mechanisms inevitably introduce utility loss or efficiency reduction while protecting data privacy. Therefore, protection mechanisms and their parameters should be carefully chosen to strike an optimal tradeoff among privacy leakage , utility loss , and efficiency reduction . To this end, federated learning practitioners need tools to measure the three factors and optimize the tradeoff between them to choose the protection mechanism that is most appropriate to the application at hand. Motivated by this requirement, we propose a framework that (1) formulates trustworthy federated learning as a problem of finding a protection mechanism to optimize the tradeoff among privacy leakage, utility loss, and efficiency reduction and (2) formally defines bounded measurements of the three factors. We then propose a meta-learning algorithm to approximate this optimization problem and find optimal protection parameters for representative protection mechanisms, including randomization, homomorphic encryption, secret sharing, and compression. We further design estimation algorithms to quantify these found optimal protection parameters in a practical horizontal federated learning setting and provide a theoretical analysis of the estimation error.

KinomeMETA: a web platform for kinome-wide polypharmacology profiling with meta-learning.

Kinase-targeted inhibitors hold promise for new therapeutic options, with multi-target inhibitors offering the potential for broader efficacy while minimizing polypharmacology risks. However, comprehensive experimental profiling of kinome-wide activity is expensive, and existing computational approaches often lack scalability or accuracy for understudied kinases. We introduce KinomeMETA, an artificial intelligence (AI)-powered web platform that significantly expands the predictive range with scalability for predicting the polypharmacological effects of small molecules across the kinome. By leveraging a novel meta-learning algorithm, KinomeMETA efficiently utilizes sparse activity data, enabling rapid generalization to new kinase tasks even with limited information. This significantly expands the repertoire of accurately predictable kinases to 661 wild-type and clinically-relevant mutant kinases, far exceeding existing methods. Additionally, KinomeMETA empowers users to customize models with their proprietary data for specific research needs. Case studies demonstrate its ability to discover new active compounds by quickly adapting to small dataset. Overall, KinomeMETA offers enhanced kinome virtual profiling capabilities and is positioned as a powerful tool for developing new kinase inhibitors and advancing kinase research. The KinomeMETA server is freely accessible without registration at https://kinomemeta.alphama.com.cn/.

Rapid in-situ calibration of computational micro-spectrometer with few-shot meta-learning

Computational micro-spectrometers comprised of detector arrays and encoding structure arrays, such as on-chip Fabry-Perot (FP) cavity filters, have great potential in many in-situ applications owing to their compact size and snapshot imaging ability. Given manufacturing deviation and environmental influence are inevitable, easy and effective calibration for spectrometer is necessary, especially for in-situ applications. Currently calibration strategies based on iterative algorithms or neural networks require accurate measurements of pixel-level (spectral) encoding functions through monochromator or large amounts of standard samples. These procedures are time-consuming and expensive, thereby impeding in-situ applications. Meta-learning algorithms with few-shot learning ability can address this challenge by incorporating the prior knowledge in the simulated dataset. In this work, we propose a meta-learning algorithm free of measuring encoding function or large amounts of standard samples to calibrate a micro-spectrometer with manufacturing deviation effectively. Our micro-spectrometer comprises 16 types of FP filters covering a wavelength range of 550-720 nm. The center wavelength of each filter type deviates from the design up to 6 nm. After calibration with 15 different color data, the average reconstruction error on the test dataset decreased from 7.2 × 10 − 3 to 1.2 × 10 − 3, and further decreased to 9.4 × 10 − 4 when the calibration data increased to 24. The performance is comparable to algorithms trained with measured encoding function both in reconstruction error and generalization ability. We estimated that the cost of in-situ calibration through reflectance measurements of color chart decreased to one percent of the cost through monochromator measurements. By exploiting prior deviation information in simulation data with meta-learning, the efficiency and cost of calibration are significantly improved, thereby facilitating the large-scale production and in-situ application of micro-spectrometers.

Meta-learning of feature distribution alignment for enhanced feature sharing

Meta-learning aims to learn common feature representations between tasks to quickly adapt and perform well when facing new tasks. However, current meta-learning algorithms ignore the differences in feature distributions among tasks. The significant differences in task features may lead to an over-fitting of the model to specific features of certain tasks during the learning process, thereby causing a decrease in adaptability to other tasks. In this paper, we propose a novel method for feature distribution alignment. Specifically, we employ knowledge distillation (KD) to minimize the Kullback–Leibler (KL) divergence between task features, thus promoting greater similarity in the extracted feature distributions. This incentivizes the model to prioritize learning shared parameters and general features across tasks during the training process, rather than task-specific features. Furthermore, we also use an adaptive dropout strategy, which adjusts the training process for the next batch based on the feature distribution of the current task. This method reduces the impact of feature distribution differences during the training iterations. Extensive experiments and analyses demonstrate that our method improves the performance of various meta-learning models in few-shot classification, few-shot fine-grained classification, cross-domain few-shot classification, and non-mutually exclusive few-shot classification.