Weighted Loss Function Research Articles

Stacking-based ensemble learning for identifying artist signatures on paintings

Identifying artist signatures on paintings is essential for authenticating artworks and advancing digital humanities. An artist’s signature is a consistent element included in each painting that the artist creates, providing a unique identifier for their work. Traditional methods that rely on expert analysis and manual comparison are time-consuming and are prone to human error. Although convolutional neural networks (CNNs) have shown promise in automating this process, existing single-model approaches struggle with the diversity and complexity of artistic styles, leading to limitations in their performance and generalizability. Therefore, this study proposes an ensemble learning approach that integrates the predictive power of multiple CNN-based models. The proposed framework leverages the strengths of three state-of-the-art CNNs: EfficientNetB4, ResNet-50, and Xception. These models were independently trained, and the predictions were combined using a meta-learning strategy. To address class imbalance, data augmentation techniques and weighted loss functions were employed. The experimental results obtained on a dataset of more than 8,000 paintings from 50 artists demonstrate significant improvements over individual CNN architectures and other ensemble methods, thereby effectively capturing complex features and improving generalizability.

Gradient Harmonization in Unsupervised Domain Adaptation.

Unsupervised domain adaptation (UDA) intends to transfer knowledge from a labeled source domain to an unlabeled target domain. Many current methods focus on learning feature representations that are both discriminative for classification and invariant across domains by simultaneously optimizing domain alignment and classification tasks. However, these methods often overlook a crucial challenge: the inherent conflict between these two tasks during gradient-based optimization. In this paper, we delve into this issue and introduce two effective solutions known as Gradient Harmonization, including GH and GH++, to mitigate the conflict between domain alignment and classification tasks. GH operates by altering the gradient angle between different tasks from an obtuse angle to an acute angle, thus resolving the conflict and trade-offing the two tasks in a coordinated manner. Yet, this would cause both tasks to deviate from their original optimization directions. We thus further propose an improved version, GH++, which adjusts the gradient angle between tasks from an obtuse angle to a vertical angle. This not only eliminates the conflict but also minimizes deviation from the original gradient directions. Finally, for optimization convenience and efficiency, we evolve the gradient harmonization strategies into a dynamically weighted loss function using an integral operator on the harmonized gradient. Notably, GH/GH++ are orthogonal to UDA and can be seamlessly integrated into most existing UDA models. Theoretical insights and experimental analyses demonstrate that the proposed approaches not only enhance popular UDA baselines but also improve recent state-of-the-art models.

MaskDGNets: Masked-attention guided dynamic graph aggregation network for event extraction.

Considering that the traditional deep learning event extraction method ignores the correlation between word features and sequence information, it cannot fully explore the hidden associations between events and events and between events and primary attributes. To solve these problems, we developed a new framework for event extraction called the masked attention-guided dynamic graph aggregation network. On the one hand, to obtain effective word representation and sequence representation, an interaction and complementary relationship are established between word vectors and character vectors. At the same time, a squeeze layer is introduced in the bidirectional independent recurrent unit to model the sentence sequence from both positive and negative directions while retaining the local spatial details to the maximum extent and establishing practical long-term dependencies and rich global context representations. On the other hand, the designed masked attention mechanism can effectively balance the word vector features and sequence semantics and refine these features. The designed dynamic graph aggregation module establishes effective connections between events and events, and between events and essential attributes, strengthens the interactivity and association between them, and realizes feature transfer and aggregation on graph nodes in the neighborhood through dynamic strategies to improve the performance of event extraction. We designed a reconstructed weighted loss function to supervise and adjust each module individually to ensure the optimal feature representation. Finally, the proposed MaskDGNets framework is evaluated on two baseline datasets, DuEE and CCKS2020. It demonstrates its robustness and event extraction performance, with F1 of 81.443% and 87.382%, respectively.

Application of deep learning algorithms for identifying deterioration in the ushnisha (Head Bun) of the Leshan Giant Buddha

AbstractThe Leshan Giant Buddha’s ushnisha (Head Bun) has suffered from loss of lime plaster, cracks, and biological damage, compromising its structural integrity and reducing the effectiveness of the drainage system in the Buddha's head. The infiltration of moisture has led to water damage within the statue, significantly accelerating its weathering. This situation urgently requires protection and reinforcement measures. Detecting deterioration in the ushnisha is a crucial step in the preservation process. In this study, we utilized two deep learning models for pixel-level semantic segmentation of the damage. Due to the small size of the cracks, a weighted loss function was applied to improve both the training speed of the model and the efficiency of crack identification. This weighting strategy proved effective for both models. The weighted K-Net model achieved a mean accuracy (mAcc) of 90.23% and a mean intersection-over-union (mIoU) of 69.55%, with a damage segmentation speed of 7 images per second, which is 1309 times faster than manual segmentation. By applying the trained deep learning models to re-examine the ushnisha, we successfully identified damage that had been overlooked during manual annotation. Using the model’s enhanced results, we conducted a comprehensive quantification of the damage across all ushnisha and identified the most severely affected areas. Additionally, we performed a model interpretability analysis to explain the decision-making process and principles of the deep learning models. This research provides significant practical value for detecting and quantifying damage in the Leshan Giant Buddha.

Association between self-reported weight loss and new long-term care insurance certifications: A 9-year Japanese older adult cohort study.

This cohort study aimed to assess weight loss associated with new long-term care insurance (LTCI) certifications over a 9-year period, accounting for the competing risk of death. We analyzed data from 3749 Japanese individuals aged ≥65 years in Kami Town, Hyogo Prefecture, Japan. Weight loss was assessed using the Kihon Checklist during the baseline survey. Data regarding LTCI certifications were collected until March 2022. Cox proportional hazards models were used to calculate adjusted hazard ratios (HRs) of 9-year LTCI certification because of weight loss, adjusted for confounding factors. To exclude the effect of competing risks, Fine-Gray regression was used to estimate subdistribution HRs. Subgroup analyses were carried out after the examination of potential interactions between subjective cognitive function, body mass index categories and weight loss. The incidence rate of new LTCI certifications was 5.16 per 100 person-years overall - broken down into 7.02 for those with weight loss and 4.97 for those without. The adjusted HR for weight loss to new LTCI certifications was 1.35 (95% CI 1.15-1.59). Considering mortality as a competing risk, the adjusted subdistribution HR was 1.37 (95% CI 1.16-1.61). Conversely, no interaction was observed between weight loss and subjective cognitive function or body mass index categories. Excluding the effect of mortality, weight loss was identified as a risk factor for new LTCI certifications. However, no interaction was observed between weight loss and subjective cognitive function or body mass index categories. Geriatr Gerontol Int 2024; ••: ••-••.

Efficient Ensemble Adversarial Attack for a Deep Neural Network (DNN)-Based Unmanned Aerial Vehicle (UAV) Vision System

In recent years, unmanned aerial vehicles (UAVs) vision systems based on deep neural networks (DNNs) have made remarkable advancements, demonstrating impressive performance. However, due to the inherent characteristics of DNNs, these systems have become increasingly vulnerable to adversarial attacks. Traditional black-box attack methods typically require a large number of queries to generate adversarial samples successfully. In this paper, we propose a novel adversarial attack technique designed to achieve efficient black-box attacks with a minimal number of queries. We define a perturbation generator that first decomposes the image into four frequency bands using wavelet decomposition and then searches for adversarial perturbations across these bands by minimizing a weighted loss function on a set of fixed surrogate models. For the target victim model, the perturbation images generated by the perturbation generator are used to query and update the weights in the loss function, as well as the weights for different frequency bands. Experimental results show that, compared to state-of-the-art methods on various image classifiers trained on ImageNet (such as VGG-19, DenseNet-121, and ResNext-50), our method achieves a success rate over 98% for targeted attacks and nearly a 100% success rate for non-targeted attacks with only 1–2 queries per image.

A novel label-aware global graph construction method and spiking-coded graph neural network for intelligent process fault diagnosis

Fault diagnosis plays a crucial role in ensuring the safety and efficiency of industrial processes. However, traditional techniques often face difficulties in handling large-scale data characterized by complex structures and relationships. To efficiently represent industrial data as graphs and develop a low-energy-cost feature extraction model, a novel label-aware global graph construction method and a spiking graph convolutional network (SGCN) are proposed in this study to achieve intelligent process fault diagnosis. The label-aware method enhances graph data representation by capturing intrinsic correlations and global features. The SGCN integrates graph convolutional layers with spiking encoding, enabling effective feature extraction while offering computational efficiency advantages. A weighted loss function is introduced to mitigate data imbalance issues. Experiments on the Tennessee Eastman process, the Three-phase Flow Facility, and the de-propanizer distillation process demonstrate SGCN’s superior performance over baseline models in various fault scenarios, while significantly reducing computational costs. The proposed method offers promising potential for reliable and efficient fault diagnosis in complex real-world industrial environments.

Unveiling the neural mechanisms of acute aerobic exercise on inhibitory control among young adults with obesity: Insights from an ERP study

Obesity has become a prominent public health concern worldwide and is associated with adverse cognitive function. Exercise, particularly aerobic exercise, is known to benefit for weight loss and cognitive function. However, whether acute aerobic exercise could yield benefits to obese individuals and the precise brain mechanisms of action remain poorly understood. The study aimed to investigate whether acute aerobic exercise could improve inhibitory control among obese individuals and what neuroelectric mechanisms are implicated. A 3 (session: control, low-intensity exercise, moderate-intensity exercise) × 2 (congruency: congruent, incongruent) within-subject design was conducted. 18 obese young male adults underwent three sessions of 30-min interventions in a counterbalanced order seperated by five days: moderate-intensity aerobic exercise (MIE), low-intensity aerobic exercise (LIE) and a control session (a sedentary period of seated rest). The Flanker task and EEG recordings (N2 and P3 amplitude) were investigated following exercise and the control treatment. Results showed that the N2 amplitude following MIE was larger than the control session, whereas a larger N2 and reduced congruent P3 amplitude was observed following MIE than LIE. However, no main effect of the session was found for reaction time and accuracy, but a significant main effect of congruency was observed. These findings suggest acute moderate-intensity aerobic exercise may modulate brain activity through enhanced recruitment of attentional resources for cognitive control and conflict monitoring in adults with obesity.

Porphyry-type mineral prospectivity mapping with imbalanced data via prior geological transfer learning

Mineral prospectivity mapping is crucial for identifying areas with economically valuable minerals. Therefore, several methods based on machine learning have been applied to predict the likelihood of mineral occurrences, especially deep learning (DL), which provides a flexible and precise approach to the use of continuous data. It allows the approximation of predictive variables with probability values related to new ore targets. However, in the early stages of mineral exploration, DL-based methods face a challenge related to class and sampling imbalance due to scarce mineral deposits, resulting in a lack of enough samples to train, limiting the model’s predictive ability. This work proposed a detailed and systematic framework to address imbalanced data issues with prior geological transfer learning and a weighted loss function. We exploited the abundant pixel information of input variables to develop a pretext geological classification and a feature data extraction task as an initializer for the trainable variables of the neural network. The proposed workflow was tested in a porphyry-rich Yukon (Canada) region and overperformed other state-of-the-art classification algorithms such as random forest, support vector machines, and logistic regression. Moreover, our results were contrasted against different geological reports, where our mineral prospectivity map was coherent with regional and local potential assessments of porphyry-type mineral occurrences. The quantitative metrics with a validation dataset suggested that the proposed method can effectively predict mineral prospective areas in different imbalanced data scenarios.

Sensitivity Matrix Update Method for Electrical Resistance Tomography Based on Error-Constrained Cross Fusion Residual Attention Network

Abstract Electrical resistance tomography (ERT) is an imaging technique of conductivity distribution. The image reconstruction algorithm based on the empty field sensitivity matrix is widely used because it provides an approximate solution to the complex ERT inverse problems. However, due to the soft field effect, the sensitivity matrix changes with the media distribution, leading to significant errors in image reconstruction. To address this issue, an error-constrained deep learning scheme for bubble flow, namely cross fusion residual attention network (CFRA-Net) is designed to update the sensitivity matrix during gas-liquid-solid fluidized bed operation. CFRA-Net employs a multi-head self-attention mechanism to enhance bubble features in boundary measurements, a multilevel cross fusion module to fuse empty field strength and boundary measurement information, and a novel serial-channel residual spatial attention block to boost the feature extraction capability of the model. A weighted loss function is designed to give more weight to the gas phase distribution region. Additionally, this paper establishes a dataset for gas-liquid-solid three-phase flow, considering background field conductivity variations. Validation and reliability of the proposed method are assessed through ablation, comparison, and noise experiments. The experimental results demonstrate that CFRA-Net achieves rapid updates of the sensitivity matrix, high imaging accuracy, and robust noise immunity.

Enhancing real-time PM2.5 forecasts: A hybrid approach of WRF-CMAQ model and CNN algorithm

As fine particulate matter (PM2.5) poses significant environmental and human health risks, there is an urgent need for accurate forecasting systems. In Taiwan, the current air quality forecasting (AQF) system based on the Weather Research and Forecasting meteorological model and Community Multiscale Air Quality model provides essential predictions but is limited by biases and computational complexities. This study introduces a convolutional neural network (CNN)-based PM2.5 forecasting model to enhance prediction accuracy. The CNN model incorporates hourly PM2.5 concentrations from surface observations and the AQF system, along with synoptic weather patterns (SWPs), to predict PM2.5 levels up to 72 h in advance. Three CNN models were developed: CNN-BASE (without SWPs), CNN-SWP (with SWPs), and CNN-SWPW (with SWPs and a weighted loss function). Performance assessment reveals a significant reduction in the mean RMSE of 72-h PM2.5 prediction, from 10.48 μg/m3 with the AQF system to 6.88 μg/m3 with the CNN-BASE model. However, CNN-BASE showed the lowest prediction accuracy for high PM2.5 concentrations (only 26.2%) due to a small subset of samples. Including SWPs improves the model's ability to capture meteorological influences, enhancing predictions of high PM2.5 concentrations. Furthermore, CNN-SWPW incorporates a weighted loss function to address imbalanced sample size distributions, further enhancing the accuracy of high PM2.5 predictions. This study demonstrates the potential of CNNs in operational air quality forecasting.

UCFTNet: multi-class dtection of city critical disaster information based on U-shaped cross fusion transformer

ABSTRACT With the continuous enhancement of remote sensing data acquisition capabilities, the demand for efficient extraction of multi-class critical urban disaster information has become increasingly urgent. However, current deep learning-based models primarily focus on single disaster objects and struggle with accuracy in complex post-disaster scenarios. To address this, we propose UCFTNet, a semantic segmentation model for urban post-disaster damaged roads and buildings. UCFTNet employs a U-shaped encoder-decoder structure and ConvNext modules, integrating spatial and multi-scale information of damaged buildings and roads from different scales and channels. Additionally, the decoder uses the CCT module and CFF module to merge multi-scale features and eliminate irrelevant background interference. Finally, a weighted loss function is designed to enhance the model's focus on disaster category pixels. Experimental results demonstrate that UCFTNet improves the accuracy of extracting damaged buildings and roads by at least 5% across three urban disaster datasets, confirming its effectiveness in post-disaster damage information extraction.

A2ST-GCM: An adaptive spatio-temporal aware graph convolutional model for predicting pathological complete response in neoadjuvant therapy

Accurate preoperative prediction of pathological complete response (pCR) following neoadjuvant immunochemotherapy is crucial for refining and customising perioperative treatment decisions. However, the challenge persists in developing reliable, interpretable, and intelligent imaging markers using early-stage computed tomography (CT). Considering the dynamic evolution of tumours during treatment can offer additional discriminative insights, modelling the spatial and temporal relationships in pre- and post-neoadjuvant treatment CT images may help address these challenges. This study proposes an adaptive spatio-temporal aware graph inference framework (A2ST-GCM) to learn potential correlations between tumour regions in the same period and across periods, generating an adaptive spatio-temporal topology, and effectively aggregating features. Specifically, the model first utilises the adaptive spatial graph convolution module (AS-GConv) to capture irregular spatial dependencies within tumour regions before and after treatment. Subsequently, the adaptive spatio-temporal graph convolution module (AST-GConv) is employed to model dynamic dependencies over time, effectively exploring the complementarity and correlation mechanisms among multiple spatio-temporal features, ultimately obtaining a spatio-temporal graph representation containing tumour heterogeneity information. Furthermore, this paper introduces a novel weighted loss function, effectively alleviating the class imbalance issue in predicting pCR. Quantitative experimental results demonstrate the outstanding performance of our model in pCR prediction. To the best of our knowledge, this paper represents the first attempt to apply graph networks to predicting pathological response in neoadjuvant immunochemotherapy. It provides a potential auxiliary tool to evaluate pathological response, aiming to identify individuals who could benefit from surgery avoidance, thereby offering significant clinical benefits for patients undergoing personalised organ-preserving treatment.

Anterior Cruciate Ligament Tear Detection Based on T-Distribution Slice Attention Framework with Penalty Weight Loss Optimisation.

Anterior cruciate ligament (ACL) plays an important role in stabilising the knee joint, prevents excessive anterior translation of the tibia, and provides rotational stability. ACL injuries commonly occur as a result of rapid deceleration, sudden change in direction, or direct impact to the knee during sports activities. Although several deep learning techniques have recently been applied in the detection of ACL tears, challenges such as effective slice filtering and the nuanced relationship between varying tear grades still remain underexplored. This study used an advanced deep learning model that integrated a T-distribution-based slice attention filtering mechanism with a penalty weight loss function to improve the performance for detection of ACL tears. A T-distribution slice attention module was effectively utilised to develop a robust slice filtering system of the deep learning model. By incorporating class relationships and substituting the conventional cross-entropy loss with a penalty weight loss function, the classification accuracy of our model is markedly increased. The combination of slice filtering and penalty weight loss shows significant improvements in diagnostic performance across six different backbone networks. In particular, the VGG-Slice-Weight model provided an area score of 0.9590 under the receiver operating characteristic curve (AUC). The deep learning framework used in this study offers an effective diagnostic tool that supports better ACL injury detection in clinical diagnosis practice.

Contrast-Enhancing Lesion Segmentation in Multiple Sclerosis: A Deep Learning Approach Validated in a Multicentric Cohort.

The detection of contrast-enhancing lesions (CELs) is fundamental for the diagnosis and monitoring of patients with multiple sclerosis (MS). This task is time-consuming and suffers from high intra- and inter-rater variability in clinical practice. However, only a few studies proposed automatic approaches for CEL detection. This study aimed to develop a deep learning model that automatically detects and segments CELs in clinical Magnetic Resonance Imaging (MRI) scans. A 3D UNet-based network was trained with clinical MRI from the Swiss Multiple Sclerosis Cohort. The dataset comprised 372 scans from 280 MS patients: 162 showed at least one CEL, while 118 showed no CELs. The input dataset consisted of T1-weighted before and after gadolinium injection, and FLuid Attenuated Inversion Recovery images. The sampling strategy was based on a white matter lesion mask to confirm the existence of real contrast-enhancing lesions. To overcome the dataset imbalance, a weighted loss function was implemented. The Dice Score Coefficient and True Positive and False Positive Rates were 0.76, 0.93, and 0.02, respectively. Based on these results, the model developed in this study might well be considered for clinical decision support.

Stacked autoencoder with weighted loss function for intrusion detection in IoT application

Stacked autoencoder with weighted loss function for intrusion detection in IoT application

Improving 3D dose prediction for breast radiotherapy using novel glowing masks and gradient-weighted loss functions.

The quality of treatment plans for breast cancer can vary greatly. This variation could be reduced by using dose prediction to automate treatment planning. Our work investigates novel methods for training deep-learning models that are capable of producing high-quality dose predictions for breast cancer treatment planning. The goal of this work was to compare the performance impact of two novel techniques for deep learning dose prediction models for tangent field treatments for breast cancer. The first technique, a "glowing" mask algorithm, encodes the distance from a contour into each voxel in a mask. The second, a gradient-weighted mean squared error (MSE) loss function, emphasizes the error in high-dose gradient regions in the predicted image. Four 3D U-Net deep learning models were trained using the planning CT and contours of the heart, lung, and tumor bed as inputs. The dataset consisted of 305 treatment plans split into 213/46/46 training/validation/test sets using a 70/15/15% split. We compared the impact of novel "glowing" anatomical mask inputs and a novel gradient-weighted MSE loss function to their standard counterparts, binary anatomical masks, and MSE loss, using an ablation study methodology. To assess performance, we examined the mean error and mean absolute error (ME/MAE) in dose across all within-body voxels, the error in mean dose to heart, ipsilateral lung, and tumor bed, dice similarity coefficient (DSC) across isodose volumes defined by 0%-100% prescribed dose thresholds, and gamma analysis (3%/3mm). The combination of novel glowing masks and gradient weighted loss function yielded the best-performing model in this study. This model resulted in a mean ME of 0.40%, MAE of 2.70%, an error in mean dose to heart and lung of -0.10 and 0.01Gy, and an error in mean dose to the tumor bed of -0.01%. The median DSC at 50/95/100% isodose levels were 0.91/0.87/0.82. The mean 3D gamma pass rate (3%/3mm) was 93%. This study found the combination of novel anatomical mask inputs and loss function for dose prediction resulted in superior performance to their standard counterparts. These results have important implications for the field of radiotherapy dose prediction, as the methods used here can be easily incorporated into many other dose prediction models for other treatment sites. Additionally, this dose prediction model for breast radiotherapy has sufficient performance to be used in an automated planning pipeline for tangent field radiotherapy and has the major benefit of not requiring a PTV for accurate dose prediction.

Improving the performance of deep learning models in predicting and classifying gamma passing rates with discriminative features and a class balancing technique: a retrospective cohort study

BackgroundThe purpose of this study was to improve the deep learning (DL) model performance in predicting and classifying IMRT gamma passing rate (GPR) by using input features related to machine parameters and a class balancing technique.MethodsA total of 2348 fields from 204 IMRT plans for patients with nasopharyngeal carcinoma were retrospectively collected to form a dataset. Input feature maps, including fluence, leaf gap, leaf speed of both banks, and corresponding errors, were constructed from the dynamic log files. The SHAP framework was employed to compute the impact of each feature on the model output for recursive feature elimination. A series of UNet++ based models were trained on the obtained eight feature sets with three fine-tuning methods including the standard mean squared error (MSE) loss, a re-sampling technique, and a proposed weighted MSE loss (WMSE). Differences in mean absolute error, area under the receiver operating characteristic curve (AUC), sensitivity, and specificity were compared between the different models.ResultsThe models trained with feature sets including leaf speed and leaf gap features predicted GPR for failed fields more accurately than the other models (F(7, 147) = 5.378, p < 0.001). The WMSE loss had the highest accuracy in predicting GPR for failed fields among the three fine-tuning methods (F(2, 42) = 14.149, p < 0.001), while an opposite trend was observed in predicting GPR for passed fields (F(2, 730) = 9.907, p < 0.001). The WMSE_FS5 model achieved a superior AUC (0.92) and more balanced sensitivity (0.77) and specificity (0.89) compared to the other models.ConclusionsMachine parameters can provide discriminative input features for GPR prediction in DL. The novel weighted loss function demonstrates the ability to balance the prediction and classification accuracy between the passed and failed fields. The proposed approach is able to improve the DL model performance in predicting and classifying GPR, and can potentially be integrated into the plan optimization process to generate higher deliverability plans.Trial registration: This clinical trial was registered in the Chinese Clinical Trial Registry on March 26th, 2020 (registration number: ChiCTR2000031276). https://clinicaltrials.gov/ct2/show/ChiCTR2000031276

Thermal Transformation of Kaolinite Clays: Analyzing Dehydroxylation and Amorphization for Improved Pozzolanic Performance

This study explores the impact of heat treatment parameters on the dehydroxylation and amorphization processes of kaolinite-based materials, in particular natural kaolin clays. It develops a quantitative method for estimating the amorphous phase in heat-treated kaolinite, using differential thermal analysis/thermogravimetry (DTA/TGA), mass spectrometry, X-ray diffraction (XRD) and Fourier transform infrared spectroscopy (FTIR). Kaolin clays, subjected to controlled heat treatments between 600°C and 800°C, undergo significant structural transformations. The study identifies the degree of dehydroxylation (DTG) as a crucial measure of kaolinite performance after heat treatment, calculated as a function of sample weight loss. The transition from a crystalline to an amorphous phase strongly influences pozzolanic activity, characterized by its ability to react with portlandite (Ca(OH)₂) in water-rich environments.The materials used, including clay from the Benslimane region of Morocco, were finely ground and subjected to various thermal treatments. Thermal analyses revealed characteristic patterns of mass loss and phase transformation. XRD and FTIR analyses provided detailed information on structural changes, confirming the amorphization of kaolinite after dehydro.

Locally sparse and robust partial least squares in scalar-on-function regression

We present a novel approach for estimating a scalar-on-function regression model, leveraging a functional partial least squares methodology. Our proposed method involves computing the functional partial least squares components through sparse partial robust M regression, facilitating robust and locally sparse estimations of the regression coefficient function. This strategy delivers a robust decomposition for the functional predictor and regression coefficient functions. After the decomposition, model parameters are estimated using a weighted loss function, incorporating robustness through iterative reweighting of the partial least squares components. The robust decomposition feature of our proposed method enables the robust estimation of model parameters in the scalar-on-function regression model, ensuring reliable predictions in the presence of outliers and leverage points. Moreover, it accurately identifies zero and nonzero sub-regions where the slope function is estimated, even in the presence of outliers and leverage points. We assess our proposed method’s estimation and predictive performance through a series of Monte Carlo experiments and an empirical dataset—that is, data collected in relation to oriented strand board. Compared to existing methods our proposed method performs favorably. Notably, our robust procedure exhibits superior performance in the presence of outliers while maintaining competitiveness in their absence. Our method has been implemented in the robsfplsr package in .