Shorter Training Time Research Articles

Research on extraction and identification of automotive wind noise based on LSTM combined road and drum experiments

Wind noise is a significant source of noise when vehicles are driven at high speeds. While wind tunnel experiments are commonly method to study real automotive wind noise, anechoic wind tunnel laboratory are prohibitive for automakers on account of their high costs and limited experimental period. Therefore, this paper proposes to extract automotive wind noise through combining road experiment and drum experiment, define significant wind noise prime regions, and verify its effectiveness. Subsequently, the Long Short-Term Memory Neural Network algorithm (LSTM) is employed to reveal the complex nonlinear relationship between wind noise and its impact areas, in order to establish an identification model for automotive wind noise. The performance of the model is then compared with Backpropagation Neural Networks (BPNN) and Support Vector Regression (SVR) models. The results indicate that the LSTM wind noise identification model exhibits higher accuracy, shorter training time, and stronger generalization ability, thus demonstrating the superiority of this model.

Rapid and accurate bacteria identification through deep-learning-based two-dimensional Raman spectroscopy

Surface-enhanced Raman spectroscopy (SERS) offers a distinctive vibrational fingerprint of the molecules and has led to widespread applications in medical diagnosis, biochemistry, and virology. With the rapid development of artificial intelligence (AI) technology, AI-enabled Raman spectroscopic techniques, as a promising avenue for biosensing applications, have significantly boosted bacteria identification. By converting spectra into images, the dataset is enriched with more detailed information, allowing AI to identify bacterial isolates with enhanced precision. However, previous studies usually suffer from a trade-off between high-resolution spectrograms for high-accuracy identification and short training time for data processing. Here, we present an efficient bacteria identification strategy that combines deep learning models with a spectrogram encoding algorithm based on wavelet packet transform and Gramian angular field techniques. In contrast to the direct analysis of raw Raman spectra, our approach utilizes wavelet packet transform techniques to compress the spectra by a factor of 1/15, while concurrently maintaining state-of-the-art accuracy by amplifying the subtle differences via Gramian angular field techniques. The results demonstrate that our approach can achieve a 99.64 % and a 90.55 % identification accuracy for two types of bacterial isolates and thirty types of bacterial isolates, respectively, while a 90 % reduction in training time compared to the conventional methods. To verify the model's stability, Gaussian noises were superimposed on the testing dataset, showing a specific generalization ability and superior performance. This algorithm has the potential for integration into on-site testing protocols and is readily updatable with new bacterial isolates. This study provides profound insights and contributes to the current understanding of spectroscopy, paving the way for accurate and rapid bacteria identification in diverse applications of environment monitoring, food safety, microbiology, and public health.

W1-Net: a highly scalable ptychography convolutional neural network

X-ray ptychography is a coherent diffraction imaging technique that allows for the quantitative retrieval of both the amplitude and phase information of a sample in diffraction-limited resolution. However, traditional reconstruction algorithms require a large number of iterations to obtain phase and amplitude images exactly, and the expensive computation precludes real-time imaging. To solve the inverse problem of ptychography data, PtychoNN uses deep convolutional neural networks for real-time imaging. However, its model is relatively simple, and its accuracy is limited by the size of the training dataset, resulting in lower robustness. To address this problem, a series of W-Net neural network models have been proposed which can robustly reconstruct the object phase information from the raw data. Numerical experiments demonstrate that our neural network exhibits better robustness, superior reconstruction capabilities and shorter training time with high-precision ptychography imaging.

A multimodal multistream multilevel fusion network for finger joint angle estimation with hybrid sEMG and FMG sensing

Finger joint angle (FJA) estimation, as a dynamic and fine-grained decoding mode, can support intuitive and natural human–machine interactions. This study is pioneering work dedicated to achieving accurate and clinically friendly FJA estimation. In this context, a novel surface electromyography and force myography-based multimodal multistream multilevel fusion network (MMMFNet) is proposed. MMMFNet utilises gate-recurrent units (GRU) networks to capture spatial dependencies among features and adopts a multilevel fusion strategy to extract more comprehensive representations from information sources than a single-level fusion. A cross-stream interaction (CSI) block is added between the GRU networks to focus on important information in the prediction and optimise the learning process and representation capability by facilitating information interaction among branch networks. The experimental results suggest that the MMMFNet has higher estimation accuracy (RMSE: 6.225±0.276 vs. 6.953±0.282, R2: 0.879±0.019 vs 0.846±0.021), shorter training time (4.146±0.074 min vs. 16.536±0.430 min), and lower computational complexity (5.4×106 FLOPs vs 8.8×106 FLOPs) than state-of-the-art fusion methods. Moreover, MMMFNet, which quantifies feature importance using attention weights, is the first explainable multimodal model for hand movement intention decoding. In summary, compared with state-of-the-art methods, MMMFNet can not only estimate FJAs more accurately but is also more clinically friendly.

Метод опорних моделей синтезу інтелектуальних систем ідентифікації нелінійних динамічних об’єктів

The paper is devoted to resolving the contradiction between the accuracy of modeling nonlinear dynamics and the speed of model construction under conditions of limited computing resources. The aim of the work is to reduce the time for building models of nonlinear dynamic objects with continuous characteristics in the form of neural networks while ensuring a given modeling accuracy. This goal is achieved by developing a new method for synthesizing intelligent systems based on the superposition of pre-trained (support) models in the form of neural networks that reflect the basic properties of the object. The scientific novelty of the work lies in the development of a method for identifying nonlinear dynamic objects in the form of neural networks with time delays based on a set of pre-trained neural network models that reflect the basic properties of the subject area. In contrast to the traditional approach to pre-training, the developed method allows building models of lower complexity and with shorter training time while ensuring the required accuracy. To determine the initial parameters of the model, expressions based on the superposition of reference models in the form of neural networks are proposed. The practical utility of the work lies in the development of an algorithm for the method of reference models for training neural networks with time delays, which can significantly reduce the training time of neural networks without losing model accuracy

Method of reference models for synthesis of intellectual systems of nonlinear dynamic objects identification

The paper is devoted to resolving the contradiction between the accuracy of modeling nonlinear dynamic objects and the speed of models building under conditions of limited computing resources. The purpose of the work is to reduce the time for building models of nonlinear dynamic objects with continuous characteristics while ensuring a given modeling accuracy. This goal is achieved by further developing the method of synthesing intelligent systems based on the superposition of pre-trained reference models in the form of neural networks reflecting the basic properties of the object. The scientific novelty of the work novelty consists in the development of a method for identifying nonlinear dynamic objects in the form of neural networks with time delays based on a set of pre-trained neural network models that reflect the basic properties of the subject area. In contrast to the traditional approach based on pre-trained neural networks the developed method allows building models of lower complexity and with shorter training time while ensuring the required accuracy. To determine the initial parameters of the model, expressions based on the superposition of reference models in the form of neural networks are proposed. The practical usefullness of the work consists in the development of an algorithm for the method of reference models for training neural networks with time delays in the tasks of identifying nonlinear dynamic objects with continuous characteristics, which can significantly reduce the training time of neural networks without losing the accuracy of the model. The value of the study lies in determining the area of effective use of the proposed method, namely, in the availability of a sufficient amount of qualitative data for the building of reference models. Insufficient data or poor data quality can significantly reduce the accuracy of reference models and, as a result, significantly reduce the training time of the target model.

Nondestructive detection of saline-alkali stress in wheat (Triticum aestivum L.) seedlings via fusion technology

BackgroundWheat (Triticum aestivum L.) is an important grain crops in the world, and its growth and development in different stages is seriously affected by saline-alkali stress, especially in seedling stage. Therefore, nondestructive detection of wheat seedlings under saline-alkali stress can provide more comprehensive technical support for wheat breeding, cultivation and management.ResultsThis research focused on moisture signal prediction and classification of saline-alkali stress in wheat seedlings using fusion techniques. After collecting and analyzing transverse relaxation time and Multispectral imaging (MSI) information of wheat seedlings, four regression models were used to predict the moisture signal. K-Nearest Neighbor (KNN) and Gaussian-Naïve Bayes (GNB) models were combined with fivefold cross validation to classify the prediction of wheat seedling stress. The results showed that wheat seedlings would increase the bound water content through a certain mechanism to enhance their saline-alkali stress. Under the same Na concentration, the effect of alkali stress on moisture, growth and spectrum of wheat seedlings is stronger than salt stress. The Gradient Boosting Decision Regression Tree model performs the best in predicting wheat moisture signals, with a coefficient of determination (R2P) of 0.98 and a root mean square error of 109.60. It also had a short training time (1.48 s) and an efficient prediction speed (1300 obs/s). The KNN and GNB demonstrated significantly enhanced predictive performance when classifying the fused dataset, compared to using single datasets individually. In particular, the GNB model performing best on the fused dataset, with Precision, Recall, Accuracy, and F1-score of 90.30, 88.89%, 88.90%, and 0.90, respectively.ConclusionsUnder the same Na concentration, the effects of alkali stress on water content, spectrum, and growth of wheat were stronger than that of salt stress, which was more unfavorable to the growth of wheat. The fusion of low-field nuclear magnetic resonance and MSI technology can improve the classification of wheat stress, and provide an effective technical method for rapid and accurate monitoring of wheat seedlings under saline-alkali stress.Graphical

Analyzing electroencephalogram signals with machine learning to comprehend online learning media

<p>In E-learning, evaluating students' comprehension of lecture video content is significant. The surge in online platform usage due to the pandemic has been remarkable, but the pressing issue is that learning outcomes still need to match the growth. Addressing this, a scientific system that gauges the comprehensibility of lecture videos becomes crucial for the effective design of future courses. This research paper is based on a cognitive approach utilizing EEG signals to determine student's level of comprehension. The study involves the design, evaluation, and comparison of multiple machines learning models, aiming to contribute to developing an efficient learning system. Fifteen distinct machine learning (ML) classifiers were implemented, among them AdaBoost (ADA), gradient boosting (GBC), extreme gradient boosting (XGboost), extra trees (ET), random forest (RF), light gradient boosting machine (light gum), and decision tree (DT) algorithms standouts. The DT exhibited exceptional performance across metrics such as area under the curve (AUC), accuracy, recall, F1 score, Kappa, precision, and matthews correlation coefficient (MCC). It achieved nearly 1.0 in these metrics while taking a short training time of only 1.7 seconds. This reveals its potential as an efficient classifier for electroencephalography (EEG) datasets and highlights its viability for practical implementation.</p>

EPMITS: An efficient prediction method incorporating trends and shapes features for chemical process variables

With the transformation of industrial production digitization and automation, process monitoring has been an indispensable technical method to realize the safe and efficient production of chemical process. Accurate prediction of process variables in chemical process can indicate the possible system change to reduce the probability of abnormal conditions. Current popular deep learning prediction methods trained with MSE or its variants may exhibit limitations in extracting shape features of chemical process data. In this paper, we proposed an efficient prediction method incorporating trends and shapes features (EPMITS) for chemical process variables. Specifically, we introduced a novel differentiable loss function Efficient Shape Error (ESE) to quantify shape differences between two time series of equal length in chemical process data. Then we trained deep learning models with MSE and ESE as loss function by two steps in training stage, to effectively acquire both trend and shape features of chemical process data. The proposed method was evaluated by the Tennessee Eastman process datasets and a real fluid catalytic cracking dataset from a petrochemical company. The results indicate that EPMITS models exhibit high prediction accuracy and short model training time across various time scales. These findings demonstrate the considerable feasibility and significant potential of EPMITS for future fault prognosis applications.

Deep learning for mango leaf disease identification: A vision transformer perspective

Over the last decade, the use of machine learning in smart agriculture has surged in popularity. Deep learning, particularly Convolutional Neural Networks (CNNs), has been useful in identifying diseases in plants at an early stage. Recently, Vision Transformers (ViTs) have proven to be effective in image classification tasks. These architectures often outperform most state-of-the-art CNN models. However, the adoption of vision transformers in agriculture is still in its infancy. In this paper, we evaluated the performance of vision transformers in identification of mango leaf diseases and compare them with popular CNNs. We proposed an optimized model based on a pretrained Data-efficient Image Transformer (DeiT) architecture that achieves 99.75% accuracy, better than many popular CNNs including SqueezeNet, ShuffleNet, EfficientNet, DenseNet121, and MobileNet. We also demonstrated that vision transformers can have a shorter training time than CNNs, as they require fewer epochs to achieve optimal results. We also proposed a mobile app that uses the model as a backend to identify mango leaf diseases in real-time.

Comparison of deep learning algorithms for site detection of false data injection attacks in smart grids

False Data Injection Attacks (FDIA) pose a significant threat to the stability of smart grids. Traditional Bad Data Detection (BDD) algorithms, deployed to remove low-quality data, can easily be bypassed by these attacks which require minimal knowledge about the parameters of the power bus systems. This makes it essential to develop defence approaches that are generic and scalable to all types of power systems. Deep learning algorithms provide state-of-the-art detection for FDIA while requiring no knowledge about system parameters. However, there are very few works in the literature that evaluate these models for FDIA detection at the level of an individual node in the power system. In this paper, we compare several recent deep learning-based model that proven their high performance and accuracy in detecting the exact location of the attack node, which are convolutional neural networks (CNN), Long Short-Term Memory (LSTM), attention-based bidirectional LSTM, and hybrid models. We, then, compare their performance with baseline multi-layer perceptron (MLP)., All the models are evaluated on IEEE-14 and IEEE-118 bus systems in terms of row accuracy (RACC), computational time, and memory space required for training the deep learning model. Each model was further investigated through a manual grid search to determine the optimal architecture of the deep learning model, including the number of layers and neurons in each layer. Based on the results, CNN model exhibited consistently high performance in very short training time. LSTM achieved the second highest accuracy; however, it had required an averagely higher training time. The attention-based LSTM model achieved a high accuracy of 94.53 during hyperparameter tuning, while the CNN model achieved a moderately lower accuracy with only one-fourth of the training time. Finally, the performance of each model was quantified on different variants of the dataset—which varied in their l2\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${\ ext{l}}_{2}$$\\end{document}-norm. Based on the results, LSTM, CNN obtained the highest accuracy followed by CNN-LSTM and lastly MLP.

MKELM based multi-classification model for foreign accent identification

The automatic identification of foreign accents can play a crucial role in various speech systems, including speaker identification, e-learning, telephone banking, and more. Additionally, it can greatly enhance the robustness of Automatic Speech Recognition (ASR) systems. Non-native accents in speech signals are characterized by distinct pronunciations, prosody, and voice characteristics of the speaker. However, automatically identifying foreign accents poses significant challenges, particularly in the context of multi-class modeling. Multi-classification models face difficulties in achieving high performance and dealing with computational challenges when confronted with multi-dimensional and unbalanced datasets, such as those with more than two accents. Furthermore, the choice of features remains a bottleneck problem for Foreign Accent Identification (FAID), further hindering performance in these tasks. Consequently, the accuracy of current systems is typically low. To address these challenges, this paper proposes a framework based on the Multi-Kernel Extreme Learning Machine (MKELM) model for the multi-classification of FAID. The MKELM model utilizes a novel weighted scheme to classify various non-native English accents, including Arabic, Chinese, Korean, French, and Spanish. The model first combines Mel-frequency cepstral coefficients (MFCCs) and prosodic features as input, trains pairwise binary classifiers independently, and subsequently employs a weighting scheme to distinguish between classes and identify accents. Through experiments, the proposed model achieves an accuracy rate of 84.72% using a paired weighting scheme. In contrast, the accuracy rate drops to 66.5% when employing the traditional non-weighted multi-classification scheme. A comparison with other models demonstrates the significant advantages of the proposed model in FAID multi-class classification, showcasing improved accuracy, reduced computational complexity (requiring fewer computations, faster learning rates, and shorter training time), and enhanced stability compared to state-of-the-art classification methods.

Encoder-decoder convolutional neural network for simple CT segmentation of COVID-19 infected lungs.

This work presents the application of an Encoder-Decoder convolutional neural network (ED-CNN) model to automatically segment COVID-19 computerised tomography (CT) data. By doing so we are producing an alternative model to current literature, which is easy to follow and reproduce, making it more accessible for real-world applications as little training would be required to use this. Our simple approach achieves results comparable to those of previously published studies, which use more complex deep-learning networks. We demonstrate a high-quality automated segmentation prediction of thoracic CT scans that correctly delineates the infected regions of the lungs. This segmentation automation can be used as a tool to speed up the contouring process, either to check manual contouring in place of a peer checking, when not possible or to give a rapid indication of infection to be referred for further treatment, thus saving time and resources. In contrast, manual contouring is a time-consuming process in which a professional would contour each patient one by one to be later checked by another professional. The proposed model uses approximately 49 k parameters while others average over 1,000 times more parameters. As our approach relies on a very compact model, shorter training times are observed, which make it possible to easily retrain the model using other data and potentially afford "personalised medicine" workflows. The model achieves similarity scores of Specificity (Sp) = 0.996 ± 0.001, Accuracy (Acc) = 0.994 ± 0.002 and Mean absolute error (MAE) = 0.0075 ± 0.0005.

Predicting viscosity for steelmaking slag: A stacking regression approach

The viscosity of steel slag plays a pivotal role in steelmaking, impacting various aspects of the steelmaking operation and product quality. This paper introduces a novel approach for predicting the viscosity of turndown slag from a basic oxygen furnace using a stacking regression model. The stacking model in this approach incorporates four base models: multiple linear regression (LR), random forest (RF), k-nearest neighbour (KNN) and AdaBoost (ADB), while random forest is selected as a meta-model. The selection of meta-models is done through a systematic framework provided for identifying suitable meta-models in a stacking model. Similarly, the framework also demonstrated the workflow for constructing and evaluating a series of stacking models with different configurations. On the evaluation of 20 stacking models (including two and three-layers), the optimal configuration is found to be LR + KNN with R2 value of 0.993 and the shortest training time. Additionally, a correlation analysis of slag viscosity with composition and temperature is done, which will help the operators of steel melting shops in optimize the steelmaking process based on the required slag viscosity. Ultimately, this study emphasizes the efficacy of optimal-configured stacking regression models in industrial applications, achieving both maximum accuracy and minimal computational time.

Energy retrofitting of hospital buildings considering climate change: An approach integrating automated machine learning with NSGA-III for multi-objective optimization

In the context of sustainable development, energy-saving renovation of existing buildings can not only effectively reduce energy consumption but also improvement the residential comfort and office environments. This study proposes an optimization method for building energy-saving renovation based on automated machine learning and the NSGA-III algorithm, aiming at efficiently and accurately identifying the best energy-saving renovation schemes. The results showed that automated machine learning could provide a predictive model for multi-objective optimization analysis in a relatively short training time, with an average prediction accuracy of over 95%. And the Stacked Ensemble algorithm and Gradient Boosting Machine algorithm perform the best among all the algorithms considered in this study. Additionally, climate change significantly affects the energy-saving benefits and thermal comfort performance of renovation measures, thereby altering the choice of optimal schemes. The final renovation schemes obtained through the NSGA-III and entropy weight TOPSIS method can achieve an energy saving rate of 22.06% to 26.48% under typical climate scenarios, while also improving thermal comfort time by 31.87% to 43.73%. In future climate scenarios, an energy saving rate of 32.73% to 33.02% and a thermal comfort time of 40.20% to 42.29% can be achieved. Combining these two technologies can quickly and effectively find the best balance among energy consumption, whole life cycle cost, and thermal comfort time. The method proposed in this paper offers new perspectives and tools for researchers in the field of building energy conservation, demonstrating the application potential of automated machine learning and the NSGA-III algorithm in the energy-saving renovation of hospital buildings.

Prediction of Vacuum Ultraviolet/Ultraviolet Gas-Phase Absorption Spectra Using Molecular Feature Representations and Machine Learning.

Ultraviolet (UV) absorption spectroscopy is a widely used tool for quantitative and qualitative analyses of chemical compounds. In the gas phase, vacuum UV (VUV) and UV absorption spectra are specific and diagnostic for many small molecules. An accurate prediction of VUV/UV absorption spectra can aid the characterization of new or unknown molecules in areas such as fuels, forensics, and pharmaceutical research. An alternative to quantum chemical spectral prediction is the use of artificial intelligence. Here, different molecular feature representation techniques were used and developed to encode chemical structures for testing three machine learning models to predict gas-phase VUV/UV absorption spectra. Structure data files (.sdf) and VUV/UV absorption spectra for 1397 volatile and semivolatile chemical compounds were used to train and test the models. New molecular features (termed ABOCH) were introduced to better capture pi-bonding, aromaticity, and halogenation. The incorporation of these new features benefited spectral prediction and demonstrated superior performance compared to computationally intensive molecular-based deep learning methods. Of the machine learning methods, the use of a Random Forest regressor returned the best accuracy score with the shortest training time. The developed machine learning prediction model also outperformed spectral predictions based on the time-dependent density functional theory.

Deep neural network method to predict the dynamical system response under random excitation of combined Gaussian and Poisson white noises

Nonlinear dynamical systems excited by combined Gaussian and Poisson white noises are widely found in the scientific field. However, the introduction of Poisson white noise results in the associated forward Kolmogorov equation becoming an integro-differential equation (IDE), which makes it difficult to solve the response of these dynamical systems. At present, traditional numerical IDE solvers are limited due to both the mesh generation and high computational costs. In this context, this paper integrates the multi-element Gauss–Legendre (GL) quadrature into the physics-informed neural networks (PINNs) algorithm to predict the response of systems excited by combined Gaussian and Poisson white noises. A residual-based adaptive distribution (RAD) sampling method, which replaces Latin hypercube sampling (LHS) method, is implemented to adjust the residual points to improve the accuracy and training efficiency of the proposed algorithm. The results demonstrate that the integration of the PINNs algorithm with RAD sampling method dramatically enhances the tolerance of noisy data and expedites the convergence of the PINNs algorithm compared with the LHS sampling method. It has the ability to achieve better performance in a shorter training time and reduces the demand for data during the training process. Additionally, the Monte Carlo (MC) simulation is carried out to demonstrate the effectiveness of the numerical results in this paper.

WORLD GREENHOUSE GAS EMISSION CLASSIFICATION USING SUPPORT VECTOR MACHINE (SVM) METHOD

The phenomenon of Heatwaves has struck several countries across the globe due to climate change. This climate change has led to an increase in greenhouse gas emissions surpassing the limits set by the IPCC Fourth Assessment Report GWPs. This study utilizes the Support Vector Machine (SVM) classification method to identify and categorize greenhouse gas emission data from 1990 to 2020 using four kernels function such as linear, polynomial, radial basis function (RBF), and sigmoid. The SVM method demonstrates excellent performance in constructing classification models with a polynomial kernel function. This is evidenced by high values of training accuracy, testing accuracy, and F1-score, accompanied by short training and testing analysis times. Successively, these values are 97.39%, 97.69%, 96.82%, 0.59 seconds, and 0.22 seconds.

Energy-Efficient Edge and Cloud Image Classification with Multi-Reservoir Echo State Network and Data Processing Units.

In an era dominated by Internet of Things (IoT) devices, software-as-a-service (SaaS) platforms, and rapid advances in cloud and edge computing, the demand for efficient and lightweight models suitable for resource-constrained devices such as data processing units (DPUs) has surged. Traditional deep learning models, such as convolutional neural networks (CNNs), pose significant computational and memory challenges, limiting their use in resource-constrained environments. Echo State Networks (ESNs), based on reservoir computing principles, offer a promising alternative with reduced computational complexity and shorter training times. This study explores the applicability of ESN-based architectures in image classification and weather forecasting tasks, using benchmarks such as the MNIST, FashionMnist, and CloudCast datasets. Through comprehensive evaluations, the Multi-Reservoir ESN (MRESN) architecture emerges as a standout performer, demonstrating its potential for deployment on DPUs or home stations. In exploiting the dynamic adaptability of MRESN to changing input signals, such as weather forecasts, continuous on-device training becomes feasible, eliminating the need for static pre-trained models. Our results highlight the importance of lightweight models such as MRESN in cloud and edge computing applications where efficiency and sustainability are paramount. This study contributes to the advancement of efficient computing practices by providing novel insights into the performance and versatility of MRESN architectures. By facilitating the adoption of lightweight models in resource-constrained environments, our research provides a viable alternative for improved efficiency and scalability in modern computing paradigms.

Towards Efficient Neural Decoder for Dexterous Finger Force Predictions.

Dexterous control of robot hands requires a robust neural-machine interface capable of accurately decoding multiple finger movements. Existing studies primarily focus on single-finger movement or rely heavily on multi-finger data for decoder training, which requires large datasets and high computation demand. In this study, we investigated the feasibility of using limited single-finger surface electromyogram (sEMG) data to train a neural decoder capable of predicting the forces of unseen multi-finger combinations. We developed a deep forest-based neural decoder to concurrently predict the extension and flexion forces of three fingers (index, middle, and ring-pinky). We trained the model using varying amounts of high-density EMG data in a limited condition (i.e., single-finger data). We showed that the deep forest decoder could achieve consistently commendable performance with 7.0% of force prediction errors and R2 value of 0.874, significantly surpassing the conventional EMG amplitude method and convolutional neural network approach. However, the deep forest decoder accuracy degraded when a smaller amount of data was used for training and when the testing data became noisy. The deep forest decoder shows accurate performance in multi-finger force prediction tasks. The efficiency aspect of the deep forest lies in the short training time and small volume of training data, which are two critical factors in current neural decoding applications. This study offers insights into efficient and accurate neural decoder training for advanced robotic hand control, which has the potential for real-life applications during human-machine interactions.