Generation Strategy Research Articles

The application of virtual synchronous generator technology in inertial control of new energy vehicle power generation

Introduction: With the rapid development of human society and economy, the power generation technology of various new energy vehicles has begun to receive widespread attention.Methods: Due to the lack of inertia and frequency stability in the new energy vehicle power generation system, this paper proposes a power generation control method that combines linear active disturbance rejection control technology and virtual synchronous generator technology. This method first introduces the control strategy and inertial response of the virtual synchronous generator. Then, it uses linear active disturbance rejection control technology to improve the virtual synchronous generator technology to deal with the uncertainty and external interference in the system.Results: The results showed that when the virtual inertia coefficient was 0, and the new energy vehicles would hardly intervene in the regulation of the grid voltage. When the virtual inertia coefficient was 5, the decline rate of the DC bus voltage of new energy vehicles had slowed down. When the virtual inertia coefficient increased, the power output of new energy vehicles can be increased to the grid. When the load suddenly increased, and the corresponding DC bus voltage decreased more slowly. In the VSG output power comparison, under the research method, the frequency fluctuation only increased by 0.09 Hz and returned to the rated frequency of 50 Hz. Additionally, the dynamic process of the system output power was the shortest, lasting only 0.05 s.Discussion: The above results show that the research method has significant superiority and effectiveness in improving the inertial response and overall stability of the new energy vehicle power system.

Accelerating deep neural network learning using data stream methodology

This paper introduces a novel machine learning approach, called BBATDD (Boosting-Based Algorithm Trained with Drift Detector), that departs from traditional epoch-based methods by adopting a data stream generation strategy from the original dataset. The proposed method is versatile, applicable to a broad range of deep structures and other machine learning architectures. The key to its effectiveness is the conversion of original data into a streaming sequence, which is then fed into the neural network structure. To monitor the learning process and prevent overfitting, a drift detector, commonly used in stream data mining, is employed. Additionally, two methods, ELB and ENT, are introduced to prioritize problematic data elements during the learning process, enhancing overall model performance. Extensive simulations on benchmark datasets (MNIST and CIFAR) validate that BBATDD achieves higher accuracy with significantly fewer batches, without sacrificing generalization capability. For instance, when applied to the MNIST dataset and processing 10 000 batches while comparing loss function values, our approach demonstrates a remarkable 45.6% improvement.

An Underwater Biomimetic Robot that can Swim, Bipedal Walk and Grasp

In developing and exploring extreme and harsh underwater environments, underwater robots can effectively replace humans to complete tasks. To meet the requirements of underwater flexible motion and comprehensive subsea operation, a novel octopus-inspired robot with eight soft limbs was designed and developed. This robot possesses the capabilities of underwater bipedal walking, multi-arm swimming, and grasping objects. To closely interact with the underwater seabed environment and minimize disturbance, the robot employs a cable-driven flexible arm for its walking in underwater floor through a bipedal walking mode. The multi-arm swimming offers a means of three-dimensional spatial movement, allowing the robot to swiftly explore and navigate over large areas, thereby enhancing its flexibility. Furthermore, the robot’s walking arm enables it to grasp and transport objects underwater, thereby enhancing its practicality in underwater environments. A simplified motion models and gait generation strategies were proposed for two modes of robot locomotion: swimming and walking, inspired by the movement characteristics of octopus-inspired multi-arm swimming and bipedal walking. Through experimental verification, the robot’s average speed of underwater bipedal walking reaches 7.26 cm/s, while the horizontal movement speed for multi-arm swimming is 8.6 cm/s.

Unveiling the Reaction Mechanism of Nitrate Reduction to Ammonia Over Cobalt-Based Electrocatalysts.

Electrocatalytic reduction of nitrate to ammonia (NRA) has emerged as an alternative strategy for sewage treatment and ammonia generation. Despite excellent performances having been achieved over cobalt-based electrocatalysts, the reaction mechanism as well as veritable active species across a wide potential range are still full of controversy. Here, we adopt CoP, Co, and Co3O4 as model materials to solve these issues. CoP evolves into a core@shell structured CoP@Co before NRA. For CoP@Co and Co catalysts, a three-step relay mechanism is carried out over superficial dynamical Coδ+ active species under low overpotential, while a continuous hydrogenation mechanism from nitrate to ammonia is unveiled over superficial Co species under high overpotential. In comparison, Co3O4 species are stable and steadily catalyze nitrate hydrogenation to ammonia across a wide potential range. As a result, CoP@Co and Co exhibit much higher NRA activity than Co3O4 especially under a low overpotential. Moreover, the NRA performance of CoP@Co is higher than Co although they experience the same reaction mechanism. A series of characterizations clarify the reason for performance enhancement highlighting that CoP core donates abundant electrons to superficial active species, leading to the generation of more active hydrogen for the reduction of nitrogen-containing intermediates.

Building Nationalism through Digital-Based Political Education

This research discusses implementing digital-based political education and nationalism-building strategies for the younger generation in Indonesia's nation-building context. This research uses a descriptive qualitative method with a literature review approach. The study highlight the importance of political education as a vital instrument in strengthening political awareness and digital literacy in society. Challenges such as low political awareness and the spread of hoaxes affect the effectiveness of digital-based political education. The proposed strategies, such as curating relevant content, active participation in social media, online training, collaboration with figures and influencers, and civic projects, can help overcome these challenges. Periodic evaluations are needed to assess the effectiveness of strategies and tactics. By combining creative approaches in digital-based political education and focusing on building nationalistic character in the younger generation, this research shows great potential to increase political participation and strengthen national identity among the younger generation.

Using unsupervised capsule neural network reveal spatial representations in the human brain.

Humans can extract high-level spatial features from visual signals, but spatial representations in the brain are complex and remain unclear. The unsupervised capsule neural network (U-CapsNet) is sensitive to the spatial location and relationship of the object, contains a special recurrent mechanism and uses a self-supervised generation strategy to represent images, which is similar to the computational principle in the human brain. Therefore, we hypothesized that U-CapsNet can help us understand how the human brain processes spatial information. First, brain activities were studied using functional magnetic resonance imaging during spatial working memory in which participants had to remember the locations of circles for a short time. Then, U-CapsNet served as a computational model of the brain to perform tasks that are identical to those performed by humans. Finally, the representational models were used to compare the U-CapsNet with the brain. The results showed that some human-defined spatial features naturally emerged in the latent space of U-CapsNet. Moreover, representations in U-CapsNet captured the response structure of two types of brain regions during different activity patterns, as well as important factors associated with human behavior. Together, our study not only provides a computationally feasible framework for modeling how the human brain encodes spatial features but also provides insights into the representational format and goals of the human brain.

Recognition of cyanobacteria promoters via Siamese network-based contrastive learning under novel non-promoter generation.

It is a vital step to recognize cyanobacteria promoters on a genome-wide scale. Computational methods are promising to assist in difficult biological identification. When building recognition models, these methods rely on non-promoter generation to cope with the lack of real non-promoters. Nevertheless, the factitious significant difference between promoters and non-promoters causes over-optimistic prediction. Moreover, designed for E. coli or B. subtilis, existing methods cannot uncover novel, distinct motifs among cyanobacterial promoters. To address these issues, this work first proposes a novel non-promoter generation strategy called phantom sampling, which can eliminate the factitious difference between promoters and generated non-promoters. Furthermore, it elaborates a novel promoter prediction model based on the Siamese network (SiamProm), which can amplify the hidden difference between promoters and non-promoters through a joint characterization of global associations, upstream and downstream contexts, and neighboring associations w.r.t. k-mer tokens. The comparison with state-of-the-art methods demonstrates the superiority of our phantom sampling and SiamProm. Both comprehensive ablation studies and feature space illustrations also validate the effectiveness of the Siamese network and its components. More importantly, SiamProm, upon our phantom sampling, finds a novel cyanobacterial promoter motif ('GCGATCGC'), which is palindrome-patterned, content-conserved, but position-shifted.

An imbalance-aware BiLSTM for control chart patterns early detection

Digital twins-based predictive models find their roots in smart manufacturing. However, their potential applications to control chart pattern recognition (CCPR) algorithms, which lie at the heart of advanced fault detection systems, remain underexplored. A key challenge in CCPR models arises from the intrinsic imbalance between classes, which can compromise the model’s performance if left untreated. Further, existing CCPR models are often trained over simulated control chart data in which abnormal patterns are generated separately from abnormal signals; the resulting classifiers, however, perform poorly in the early detection of abnormalities in real-time production environments. To address these challenges, we develop a cost-sensitive, bi-directional long short-term memory neural network for data sequences with mixed normal and abnormal signals. We further introduce a novel adaptive weighting strategy for data generation by enforcing the rates of abnormal signals within mini-batch distributions. Our model benefits from a new bi-objective early stopping technique, which optimally balances loss minimization and G-mean maximization for model training. Finally, we introduce a novel rolling window-based metric for evaluating CCPR classifier stability. We conduct a comprehensive experimental study of our model using both simulated data and two real-world datasets collected from biomanufacturing and wafer industries. The results of our study consistently demonstrate the superiority of our proposed stopping technique over traditional methods. Our experiments further show the effectiveness of our proposed model in maintaining the classifier stability and specifying optimal process monitoring window length within the datasets.

KGTS: Contrastive Trajectory Similarity Learning over Prompt Knowledge Graph Embedding

Trajectory similarity computation serves as a fundamental functionality of various spatial information applications. Although existing deep learning similarity computation methods offer better efficiency and accuracy than non-learning solutions, they are still immature in trajectory embedding and suffer from poor generality and heavy preprocessing for training. Targeting these limitations, we propose a novel framework named KGTS based on knowledge graph grid embedding, prompt trajectory embedding, and unsupervised contrastive learning for improved trajectory similarity computation. Specifically, we first embed map grids with a GRot embedding method to vigorously grasp the neighbouring relations of grids. Then, a prompt trajectory embedding network incorporates the resulting grid embedding and extracts trajectory structure and point order information. It is trained by unsupervised contrastive learning, which not only alleviates the heavy preprocessing burden but also provides exceptional generality with creatively designed strategies for positive sample generation. The prompt trajectory embedding adopts a customized prompt paradigm to mitigate the gap between the grid embedding and the trajectory embedding. Extensive experiments on two real-world trajectory datasets demonstrate the superior performance of KGTS over state-of-the-art methods.

Hot or Cold? Adaptive Temperature Sampling for Code Generation with Large Language Models

Recently, Large Language Models (LLMs) have shown impressive abilities in code generation. However, existing LLMs' decoding strategies are designed for Natural Language (NL) generation, overlooking the differences between NL and programming languages (PL). Due to this oversight, a better decoding strategy for code generation remains an open question. In this paper, we conduct the first systematic study to explore a decoding strategy specialized in code generation. With an analysis of loss distributions of code tokens, we find that code tokens can be divided into two categories: challenging tokens that are difficult to predict and confident tokens that can be easily inferred. Among them, the challenging tokens mainly appear at the beginning of a code block. Inspired by the above findings, we propose a simple yet effective method: Adaptive Temperature (AdapT) sampling, which dynamically adjusts the temperature coefficient when decoding different tokens. We apply a larger temperature when sampling for challenging tokens, allowing LLMs to explore diverse choices. We employ a smaller temperature for confident tokens avoiding the influence of tail randomness noises. We apply AdapT sampling to LLMs with different sizes and conduct evaluations on two popular datasets. Results show that AdapT sampling significantly outperforms state-of-the-art decoding strategy.

CycleVTON: A Cycle Mapping Framework for Parser-Free Virtual Try-On

Image-based virtual try-on aims to transfer a target clothing onto a specific person. A significant challenge is arbitrarily matched clothing and person lack corresponding ground truth to supervised learning. A recent pioneering work leveraged an improved cycleGAN to enable one network to generate the desired image for another network during training. However, there is no difference in the result distribution before and after the clothing changes. Therefore, using two different networks is unnecessary and may even increase the difficulty of convergence. Furthermore, the introduced human parsing used to provide body structure information in the input also have a negative impact on the try-on result. How to employ a single network for supervised learning while eliminating human parsing? To tackle these issues, we present a Cycle mapping Virtual Try-On Network (CycleVTON), which can produce photo-realistic try-on results by using a cycle mapping framework without the parser. In particular, we introduce a flow constraint loss to achieve supervised learning of arbitrarily matched clothing and person as inputs to the deformer, thus naturally mimicking the interaction between clothing and the human body. Additionally, we design a skin generation strategy that can adapt to the shape of the target clothing by dynamically adjusting the skin region, i.e., by first removing and then filling skin areas. Extensive experiments conducted on challenging benchmarks demonstrate that our proposed method exhibits superior performance compared to state-of-the-art methods.

Deep Incomplete Multi-View Learning Network with Insufficient Label Information

Due to the efficiency of integrating semantic consensus and complementary information across different views, multi-view classification methods have attracted much attention in recent years. However, multi-view data often suffers from both the miss of view features and insufficient label information, which significantly decrease the performance of traditional multi-view classification methods in practice. Learning for such simultaneous lack of feature and label is crucial but rarely studied. To tackle these problems, we propose a novel Deep Incomplete Multi-view Learning Network (DIMvLN) by incorporating graph networks and semi-supervised learning in this paper. Specifically, DIMvLN firstly designs the deep graph networks to effectively recover missing data with assigning pseudo-labels of large amounts of unlabeled instances and refine the incomplete feature information. Meanwhile, to enhance the label information, a novel pseudo-label generation strategy with the similarity constraints of unlabeled instances is proposed to exploit additional supervisory information and guide the completion module to preserve more semantic information of absent multi-view data. Besides, we design view-specific representation extractors with the autoencoder structure and contrastive loss to learn high-level semantic representations for each view, promote cross-view consistencies and augment the separability between different categories. Finally, extensive experimental results demonstrate the effectiveness of our DIMvLN, attaining noteworthy performance improvements compared to state-of-the-art competitors on several public benchmark datasets. Code will be available at GitHub.

AE-NeRF: Audio Enhanced Neural Radiance Field for Few Shot Talking Head Synthesis

Audio-driven talking head synthesis is a promising topic with wide applications in digital human, film making and virtual reality. Recent NeRF-based approaches have shown superiority in quality and fidelity compared to previous studies. However, when it comes to few-shot talking head generation, a practical scenario where only few seconds of talking video is available for one identity, two limitations emerge: 1) they either have no base model, which serves as a facial prior for fast convergence, or ignore the importance of audio when building the prior; 2) most of them overlook the degree of correlation between different face regions and audio, e.g., mouth is audio related, while ear is audio independent. In this paper, we present Audio Enhanced Neural Radiance Field (AE-NeRF) to tackle the above issues, which can generate realistic portraits of a new speaker with few-shot dataset. Specifically, we introduce an Audio Aware Aggregation module into the feature fusion stage of the reference scheme, where the weight is determined by the similarity of audio between reference and target image. Then, an Audio-Aligned Face Generation strategy is proposed to model the audio related and audio independent regions respectively, with a dual-NeRF framework. Extensive experiments have shown AE-NeRF surpasses the state-of-the-art on image fidelity, audio-lip synchronization, and generalization ability, even in limited training set or training iterations.

Generative Multi-Modal Knowledge Retrieval with Large Language Models

Knowledge retrieval with multi-modal queries plays a crucial role in supporting knowledge-intensive multi-modal applications. However, existing methods face challenges in terms of their effectiveness and training efficiency, especially when it comes to training and integrating multiple retrievers to handle multi-modal queries. In this paper, we propose an innovative end-to-end generative framework for multi-modal knowledge retrieval. Our framework takes advantage of the fact that large language models (LLMs) can effectively serve as virtual knowledge bases, even when trained with limited data. We retrieve knowledge via a two-step process: 1) generating knowledge clues related to the queries, and 2) obtaining the relevant document by searching databases using the knowledge clue. In particular, we first introduce an object-aware prefix-tuning technique to guide multi-grained visual learning. Then, we align multi-grained visual features into the textual feature space of the LLM, employing the LLM to capture cross-modal interactions. Subsequently, we construct instruction data with a unified format for model training. Finally, we propose the knowledge-guided generation strategy to impose prior constraints in the decoding steps, thereby promoting the generation of distinctive knowledge clues. Through experiments conducted on three benchmarks, we demonstrate significant improvements ranging from 3.0% to 14.6% across all evaluation metrics when compared to strong baselines.

A causality-inspired generalized model for automated pancreatic cancer diagnosis

Pancreatic cancer (PC) is a severely malignant cancer variant with high mortality. Since PC has no obvious symptoms, most PC patients are belatedly diagnosed at advanced disease stages. Recently, artificial intelligence (AI) approaches have demonstrated promising prospects for early diagnosis of pancreatic cancer. However, certain non-causal factors (such as intensity and texture appearance variations, also called confounders) tend to induce spurious correlation with PC diagnosis. This undermines the generalization performance and the clinical applicability of the AI-based PC diagnosis approaches. Therefore, we propose a causal intervention based automated method for pancreatic cancer diagnosis with contrast-enhanced computerized tomography (CT) images, where a confounding effects reduction scheme is developed for alleviating spurious correlations to achieve unbiased learning, thereby improving the generalization performance. Specifically, a continuous image generation strategy was developed to simulate wide variations of intensity differences caused by imaging heterogeneities, where Monte Carlo sampling is added to further enhance the continuity of simulated images. Then, to enhance the pancreatic texture variability, a texture diversification method was introduced in conjunction with gradient-based data augmentation. Finally, a causal intervention strategy was proposed to alleviate the adverse confounding effects by decoupling the causal and non-causal factors and combining them randomly. Extensive experiments showed remarkable diagnosis performance on a cross-validation dataset. Also, promising generalization performance with an average accuracy of 0.87 was attained on three independent test sets of a total of 782 subjects. Therefore, the proposed method shows high clinical feasibility and applicability for pancreatic cancer diagnosis.

Abstract 3097: CETSA HT: Employing rapid high-throughput target engagement technology in primary screening of P53

Abstract The confirmation of target engagement for candidate drugs is a fundamental requisite in drug discovery. A significant portion of projects fail to reach the clinic due to lack of efficacy or failure to show lead candidates interacting with the intended target in a more complex environment. The utility of using the CETSA® (CEllular Thermal Shift Assay) technology to measure target engagement in intact cells is increasingly exploited in high throughput screening, where examples demonstrate low false positive rates in screens of up to 500K compounds. The p53 tumor suppressor protein, a crucial guardian against cancer, functions as a transcription factor that mediates cancer suppression by regulating genes involved in a wide range of cellular processes such as DNA repair, cell cycle and apoptosis. Unfortunately, frequent mutations in the TP53 gene hinder p53's tumor suppressor role in human cancers, thus designating it as a promising target for anticancer therapy. However, attempts to restore p53 function for therapeutic purposes have not yet advanced to clinical approval, highlighting the complexities in leveraging p53 for cancer treatment. Its undruggable nature, attributed to the absence of accessible deep pockets, lack of enzymatic activity, and its intracellular location, further complicates targeted interventions with both low and high molecular weight drugs. Here, we demonstrate the analysis of a high-throughput plate based CETSA® HT screen on the challenging target p53, performed using AlphaLISA® SureFire® Ultra™ technology and our newly developed high-throughput screening platform. Hits identified from this screen are privileged from the onset by binding to the target in a physiologically relevant cellular matrix and in a binding-mode agnostic way. This is in contrast to traditional screening methods targeting activity or in vitro interaction where cellular activity is addressed in later stages. This assay cascade highlights a rapid target-to-hit lead generation strategy placing the emphasis on finding chemical matter that interacts with endogenous target in living cells first, followed by analysis of hit material activity. The ‘shortcut’ to chemical matter active in living cells represents significant potential cost savings in comparison to assay development time on recombinantly produced proteins and/or modified cell lines for traditional HTS screening approaches that still rely on target engagement assessment downstream. Citation Format: Merve Kacal, Tomas Friman, Laurence Arnold, Victoria Brehmer, Daniel Martinez Molina, Stina Lundgren. CETSA HT: Employing rapid high-throughput target engagement technology in primary screening of P53 [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3097.

Spiky metal-organic framework nanosystem for enhanced cuproptosis-mediated cancer immunotherapy

Immunotherapy, particularly αPD-1 therapy, has been hailed as a significant breakthrough in cancer treatment. However, the limited response rates and unknown mechanisms hinder the clinical application of αPD-1 blockade boosting with hybrid nanoparticles. Herein, we introduce a strategy to enhance cancer immunotherapy by leveraging cuproptosis, initiated by the spiky copper-based metal-organic framework (S@Cu-MOF) loaded with polyphyllin I (PPI) and synergized with αPD-1 therapy. Interestingly, the unique spiky structure of S@Cu-MOF not only significantly enhances tumor cell uptake but also leads to pronounced DNA damage. The application of S@Cu-MOF/PPI with αPD-1 therapy led to obvious tumor regression and prolonged survival due to cuproptosis, which involved substantial mitochondrial damage. This process amplifies the cGAS/STING immune pathway, enhancing the immune response against cancer. Additionally, this strategy promoted immunogenic cell death (ICD), recruitment of mature dendritic cells (DCs), T cell-mediated immunity, and effective memory CD8+ T cells, resulting in significant abscopal tumor suppression and lung metastasis reduction. Our study highlights the potential of S@Cu-MOF/PPI as a sensitizer for αPD-1 with the mechanism of cuproptosis and the cGAS/STING pathway, offering a promising strategy for the next generation of immunotherapy.

“Four-in-One” Nanozyme for Amplified Catalytic-Photothermal Therapy

The cancer therapeutic efficacy of the peroxidase (POD)-mimicking nanozyme-based monotherapy is significantly hindered due to insufficient intratumoral hydrogen peroxide (H2O2) and glutathione (GSH) consumption effect on reactive oxygen species (ROS). In this study, we present the development of poly(o-phenylenediamine)@gold nanoparticles (AuNPs) (PoPD@Au) nanocomposites for multifunctional catalytic-photothermal therapy. These nanocomposites exhibit triple distinct nanozymatic activities, i.e., POD-like activity that catalyzes H2O2 to ROS, glucose oxidase (GOx)-like activity that supplements endogenous H2O2, and GSH depleting activity that decreases the ROS consumption efficiency. This open source and reduce expenditure strategy for ROS generation allows for the amplification of tumor oxidative stress, thereby enhancing anti-tumor efficiency. Additionally, the PoPD@Au nanocomposites demonstrate outstanding photothermal conversion efficiency, contributing to the synergistic effect between PoPD and AuNPs. Moreover, we reveal the improved photothermal performance of PoPD@Au triggered by the tumor microenvironment pH, which provides additional benefits for targeted catalytic-photothermal therapy. This “four-in-one” design of PoPD@Au enables efficient anti-tumor effects both in vitro and in vivo, making it a universal strategy for engineering catalytic-photothermal therapeutic nanoagents.

Electrocardiogram identification based on data generative network and non-fiducial data processing

Nowadays, the use of biological signals as a criterion for identity recognition has gained increasing attention from various organizations and companies. Therefore, it has become crucial to have a biometric identity recognition method that is fast and accurate. In this paper, we propose a linear electrocardiogram (ECG) data preprocessing algorithm based on Kalman filters for rapid noise data filtering (wavelet transform filtering algorithm). Additionally, we introduce a generative network model called Data Generation Strategy Network (DRCN) based on generative networks. The DRCN is employed to augment training samples for convolutional classification networks, ultimately improving the classification performance of the model. Through the final experiments, our method successfully reduced the average misidentification rate of ECG-based identity recognition to 2.5%, and achieved an average recognition rate of 98.7% for each category, significantly surpassing previous achievements. In the future, this method is expected to be widely applied in the field of ECG-based identity recognition.

Three novel machine learning-based adaptive controllers for a photovoltaic shunt active power filter performance enhancement

This study develops three new machine learning-based algorithms using the SVR prediction approach. The overall objective is to enhance the performance of the PV shunt active power filter (PV-SAPF) in order to successfully fulfil its multi-functionality in terms of PV power generation along with power quality upgrades. The first technique, named "SVR-INC MPPT", incorporates an ML-based duty cycle prediction function, which is intended to speed up the MPP finding process. The algorithm then switches to the INC approach, guaranteeing an accurate steady-state response. The second and third techniques are the ML-based Adaline for DC voltage regulation, and the ML-based Adaline SRF strategy for reference currents generation. The adopted approaches are based on the prediction of two weights performing the actions of the proportional and integral of DC voltage controller, and nine weights for the fundamental currant extraction. The advantage lies in overcoming the Adaline-based algorithm's limitations, requiring an on-line large number of iterations. Three simulation scenarios along with six controllers based on classical and intelligent techniques, are adopted to prove the effectiveness of the proposed ML-controllers. The results display (i) an accurate and immediate ML-based harmonics identification (minimizing the extraction error by up to 99% compared to the Adaline approach). (ii) a response time reduction range from 65% to 100%, of the ML-based DC voltage output controller compared to the PI-based one. (iii) a THD range from 2.41% to 4.45%. (iv) 20 ms and 0% of PV power response time and overshoot, respectively (v) DC voltage overshoot range from 0.04% to 1.1%. Consequently, these three ML-based controllers offer the best options for a powerful PV-SAPF system in terms of performance tracking and harmonic attenuation. Moreover, the obtained metrics comply with IEEE-519 standard.