Artificial Intelligence Frameworks Research Articles

Artificial intelligence and machine learning at various stages and scales of process systems engineering

AbstractWe review the utility and application of artificial intelligence (AI) and machine learning (ML) at various process scales in this work, from molecules and reactions to materials to processes, plants, and supply chains; furthermore, we highlight whether the application is at the design or operational stage of the process. In particular, we focus on the distinct representational frameworks employed at the various scales and the physics (equivariance, additivity, injectivity, connectivity, hierarchy, and heterogeneity) they capture. We also review AI techniques and frameworks important in process systems, including hybrid AI modelling, human‐AI collaborations, and generative AI techniques. In hybrid AI models, we emphasize the importance of hyperparameter tuning, especially in the case of physics‐informed regularization. We highlight the importance of studying human‐AI interactions, especially in the context of automation, and distinguish the features of human‐complements‐AI systems from those of AI‐complements‐human systems. Of particular importance in the AI‐complements‐human framework are model explanations, including rule‐based explanation, explanation‐by‐example, explanation‐by‐simplification, visualization, and feature relevance. Generative AI methods are becoming increasingly relevant in process systems engineering, especially in contexts that do not belong to ‘big data’, primarily due to the lack of high quality labelled data. We highlight the use of generative AI methods including generative adversarial networks, graph neural networks, and large language models/transformers along with non‐traditional process data (images, audio, and text).

An Open Annotated Dataset and Baseline Machine Learning Model for Segmentation of Vertebrae with Metastatic Bone Lesions from CT.

Automatic analysis of pathologic vertebrae from computed tomography (CT) scans could significantly improve the diagnostic management of patients with metastatic spine disease. We provide the first publicly available annotated imaging dataset of cancerous CT spines to help develop artificial intelligence frameworks for automatic vertebrae segmentation and classification. This collection contains a dataset of 55 CT scans collected on patients with various types of primary cancers at two different institutions. In addition to raw images, data include manual segmentations and contours, vertebral level labeling, vertebral lesion-type classifications, and patient demographic details. Our automated segmentation model uses nnU-Net, a freely available open-source framework for deep learning in healthcare imaging, and is made publicly available. This data will facilitate the development and validation of models for predicting the mechanical response to loading and the resulting risk of fractures and spinal deformity.

Convergence and Innovation of Artificial Intelligence in Corporate Strategic Planning: Opportunities, Challenges and Future Research Directions

In the contemporary digital era, the swift evolution of Artificial Intelligence (AI) has not only revolutionized various sectors but has also become an indispensable asset in corporate strategic planning, decision-making processes, and operational efficiency. This paper delves into the multifaceted applications of AI across a myriad of industries, providing a comprehensive analysis of its integration into corporate strategy and its profound impact on business operations. This study adopts a comprehensive research methodology that merges qualitative case studies with quantitative data analysis to scrutinize the practical use of AI in strategic planning. It delves into the interdependent relationship between AI and conventional digital frameworks, such as Customer Relationship Management (CRM) and Enterprise Resource Planning (ERP) systems. The research aims to illuminate how the fusion of these technologies can substantially enhance corporate competencies and optimize operational workflows. The study's findings indicate that while AI has the potential to offer transformative solutions, its integration into strategic planning is not without challenges. Ethical considerations, legal ramifications, and societal implications must be carefully navigated. Additionally, the study reveals a notable void in existing literature concerning the enduring strategic ramifications of AI and its effects on corporate culture and talent management practices. This gap highlights the need for further exploration into how AI shapes the long-term strategic direction and internal dynamics within organizations.

An empirical study of AI techniques in mobile applications

The integration of artificial intelligence (AI) into mobile applications has significantly transformed various domains, enhancing user experiences and providing personalized services through advanced machine learning (ML) and deep learning (DL) technologies. AI-driven mobile apps typically refer to applications that leverage ML/DL technologies to perform key tasks such as image recognition and natural language processing. Despite existing research exploring how mobile apps exploit AI techniques, they have the following main limitations: (1) Most existing studies focus on DL-based apps, with limited research on ML-based apps. (2) Existing research typically focuses on investigating the apps and the technologies utilized in the apps, lacking user-level analysis. (3) The number of apps studied is limited, with only 1,000 to 2,000 ML/DL apps identified after filtering. To fill the gap, in this paper, we conducted the most extensive empirical study on AI applications, exploring on-device ML apps, on-device DL apps, and AI service-supported (cloud-based) apps. Our study encompasses 56,682 real-world AI applications, focusing on three crucial perspectives: (1) Application analysis, where we analyze the popularity of AI apps and investigate the update states of AI apps; (2) Framework and model analysis, where we analyze AI framework usage and AI model protection; (3) User analysis, where we examine user privacy protection and user review attitudes. Our study has strong implications for AI app developers, users, and AI R&D. On one hand, our findings highlight the growing trend of AI integration in mobile applications, demonstrating the widespread adoption of various AI frameworks and models. On the other hand, our findings emphasize the need for robust model protection to enhance app security. Additionally, our study highlights the importance of user privacy and presents user attitudes towards the AI technologies utilized in current AI apps. We provide our AI app dataset (currently the most extensive AI app dataset) as an open-source resource for future research on AI technologies utilized in mobile applications.

Green knowledge management: a bibliometric analysis, research trends and future directions

Purpose Green knowledge management (GKM) has become a critical factor for modern organizations, offering a multifaceted approach to enhancing business operations while promoting environmental sustainability. The purpose of this paper is to review the literature and explore how GKM helps companies reduce their ecological footprint by digitalizing processes and minimizing physical storage needs, thus aligning with growing expectations for corporate environmental responsibility. The necessity for conducting a GKM review is underscored by the increasing recognition of sustainability, environmental performance and as a critical component of organizational corporate success in the contemporary business landscape. Design/methodology/approach This study conducted a bibliometric analysis of 994 papers related to GKM from 2005 until August 2024. Findings Over the 20 years of research on GKM, this bibliometric analysis drew the interlinked relationships between main theme of KM in the context of sustainable practices, processes, social and related theories. This study identifies green innovation, digital transformation, green and lean supply chains, leadership and technology tools. Future research can explore themes such as green and lean processes, barriers and enablers, role of leadership and ambidexterity in green supply chain integration using different theoretical lenses. Research limitations/implications Although the findings of this study are important to GKM research, there are a number of limitations to this investigation. The research database used exclusively was the Web of Science and Scopus, which contained the indexed papers. Future investigations may draw from various databases, such as EBSCO and Google Scholar. Additionally, the use of topic modelling and other artificial intelligence frameworks could enhance research analysis. Subsequently, research papers may expand their scope to encompass conference papers, books and book chapters, in addition to articles. Practical implications This review gives a holistic understanding of GKM practices across various sectors which not only enhances organizational performance but also contributes to broader environmental goals. This study offers an understanding on GKM on holistic approach that combines total quality management (TQM), absorptive capacity and collaborative frameworks to foster innovation and sustainability. Originality/value It is evident from the review that research in GKM is evolving gaining momentum. This review provides a comprehensive understanding of the GKM literature, from its earliest forms of basic concepts aligned with the sustainability and progressing more into the concepts of GKM adoption for green innovation by highlighting multiple theories, factors and gaps in the studies.

Exploring the Artificial Intelligence and Its Impact in Pharmaceutical Sciences: Insights Toward the Horizons Where Technology Meets Tradition.

The technological revolutions in computers and the advancement of high-throughput screening technologies have driven the application of artificial intelligence (AI) for faster discovery of drug molecules with more efficiency, and cost-friendly finding of hit or lead molecules. The ability of software and network frameworks to interpret molecular structures' representations and establish relationships/correlations has enabled various research teams to develop numerous AI platforms for identifying new lead molecules or discovering new targets for already established drug molecules. The prediction of biological activity, ADME properties, and toxicity parameters in early stages have reduced the chances of failure and associated costs in later clinical stages, which was observed at a high rate in the tedious, expensive, and laborious drug discovery process. This review focuses on the different AI and machine learning (ML) techniques with their applications mainly focused on the pharmaceutical industry. The applications of AI frameworks in the identification of molecular target, hit identification/hit-to-lead optimization, analyzing drug-receptor interactions, drug repurposing, polypharmacology, synthetic accessibility, clinical trial design, and pharmaceutical developments are discussed in detail. We have also compiled the details of various startups in AI in this field. This review will provide a comprehensive analysis and outline various state-of-the-art AI/ML techniques to the readers with their framework applications. This review also highlights the challenges in this field, which need to be addressed for further success in pharmaceutical applications.

Unveiling Uncertainty: Exploring the Potential of Neutrosophic Statistics for Artificial Intelligence

Artificial intelligence (AI) has revolutionized various fields, but its effectiveness can hampered by limitations in handling uncertainty inherent in real-world data. Traditional statistical methods often rely on binary logic (true/false) and may not adequately capture the nuances of uncertainty present in real-world data. This paper explores neutrosophic statistics (NS) as a novel approach to address uncertainty in AI applications. NS, introduced by Smarandache (1998), generalizes classical set theory by incorporating three truth-values: truth (T), indeterminacy (I), and falsity (F). This allows for a more nuanced representation of uncertainty compared to traditional binary methods. We discuss the limitations of classical statistics and how NS can overcome them. The paper explores the potential of NS for AI in areas like machine learning, data fusion, and Explainable AI (XAI). We propose research questions and methodology to investigate the effectiveness of NS integration with AI frameworks. The expected outcome is to contribute to the development of more robust and reliable AI systems by demonstrating the effectiveness of NS in handling uncertainty. Finally, we discuss future research directions to further explore the potential of NS for AI.

1.5 million materials narratives generated by chatbots

The advent of artificial intelligence (AI) has enabled a comprehensive exploration of materials for various applications. However, AI models often prioritize frequently encountered material examples in the scientific literature, limiting the selection of suitable candidates based on inherent physical and chemical attributes. To address this imbalance, we generated a dataset consisting of 1,453,493 natural language-material narratives from OQMD, Materials Project, JARVIS, and AFLOW2 databases based on ab initio calculation results that are more evenly distributed across the periodic table. The generated text narratives were then scored by both human experts and GPT-4, based on three rubrics: technical accuracy, language and structure, and relevance and depth of content, showing similar scores but with human-scored depth of content being the most lagging. The integration of multimodal data sources and large language models holds immense potential for AI frameworks to aid the exploration and discovery of solid-state materials for specific applications of interest.

Systematic Review of Real-Time Analytics and Artificial Intelligence Frameworks for Financial Fraud Detection

Systematic Review of Real-Time Analytics and Artificial Intelligence Frameworks for Financial Fraud Detection

ATOMMIC: An Advanced Toolbox for Multitask Medical Imaging Consistency to facilitate Artificial Intelligence applications from acquisition to analysis in Magnetic Resonance Imaging

Background and Objectives:Artificial intelligence (AI) is revolutionizing Magnetic Resonance Imaging (MRI) along the acquisition and processing chain. Advanced AI frameworks have been applied in various successive tasks, such as image reconstruction, quantitative parameter map estimation, and image segmentation. However, existing frameworks are often designed to perform tasks independently of each other or are focused on specific models or single datasets, limiting generalization. This work introduces the Advanced Toolbox for Multitask Medical Imaging Consistency (ATOMMIC), a novel open-source toolbox that streamlines AI applications for accelerated MRI reconstruction and analysis. ATOMMIC implements several tasks using deep learning (DL) models and enables MultiTask Learning (MTL) to perform related tasks in an integrated manner, targeting generalization in the MRI domain. Methods:We conducted a comprehensive literature review and analyzed 12,479 GitHub repositories to assess the current landscape of AI frameworks for MRI. Subsequently, we demonstrate how ATOMMIC standardizes workflows and improves data interoperability, enabling effective benchmarking of various DL models across MRI tasks and datasets. To showcase ATOMMIC’s capabilities, we evaluated twenty-five DL models on eight publicly available datasets, focusing on accelerated MRI reconstruction, segmentation, quantitative parameter map estimation, and joint accelerated MRI reconstruction and segmentation using MTL. Results:ATOMMIC’s high-performance training and testing capabilities, utilizing multiple GPUs and mixed precision support, enable efficient benchmarking of multiple models across various tasks. The framework’s modular architecture implements each task through a collection of data loaders, models, loss functions, evaluation metrics, and pre-processing transformations, facilitating seamless integration of new tasks, datasets, and models. Our findings demonstrate that ATOMMIC supports MTL for multiple MRI tasks with harmonized complex-valued and real-valued data support while maintaining active development and documentation. Task-specific evaluations demonstrate that physics-based models outperform other approaches in reconstructing highly accelerated acquisitions. These high-quality reconstruction models also show superior accuracy in estimating quantitative parameter maps. Furthermore, when combining high-performing reconstruction models with robust segmentation networks through MTL, performance is improved in both tasks. Conclusions:ATOMMIC advances MRI reconstruction and analysis by leveraging MTL and ensuring consistency across tasks, models, and datasets. This comprehensive framework serves as a versatile platform for researchers to use existing AI methods and develop new approaches in medical imaging.

Systematic exploration and in-depth analysis of ChatGPT architectures progression

The fast evolution of artificial intelligence frameworks has resulted in the creation of increasingly sophisticated large language models (LLM), ChatGPT being the most famous one. This study paper dives into this LLM with a case study of ChatGPT’s architecture and provides a thorough comparative analysis of its numerous versions, tracking its history from its conception to its most recent incarnations. This research intends to give a full knowledge of the model’s history by investigating the underlying mechanisms and enhancements provided in each edition. The comparative analysis covers key aspects such as model size, training data, fine-tuning techniques, and performance metrics. Furthermore, this study evaluates the limits of ChatGPT in its many incarnations. These limitations include common sense reasoning difficulties, biased replies, verbosity, sensitivity to input wording, and others. Each constraint is investigated for potential remedies and workarounds. This research article also provides a complete analysis of the ChatGPT architecture and its progress through multiple iterations. It gives vital insights for academics, developers, and users wanting to harness the promise of ChatGPT while managing its restrictions by exploring both the model’s strengths and limitations. The distinctiveness of this paper rests in its comprehensive assessment of ChatGPT’s architectural development and its practical strategy for resolving the myriad difficulties in producing cohesive and contextually relevant replies.

Exploring AI's Role in Supporting Diversity and Inclusion Initiatives in Multicultural Marketplaces

This paper explores the relationship between diversity, inclusion, and artificial intelligence (AI) in multicultural workplaces and markets. Growing recognition of the value of diversity and inclusion (D&I) initiatives is changing organizational practices globally. Despite progress, challenges such as unconscious bias, resistance to change, and difficulty in measuring progress still exist.This study is based on a comprehensive literature review that adheres to academic ethical standards. These findings highlight the potential of AI to improve recruitment processes, personalize marketing strategies, and develop inclusive technologies. However, challenges such as bias in AI algorithms and limited data diversity need to be addressed.The paper emphasizes the need for ethical AI frameworks, diverse representation in AI development, and transparency in the use of AI to promote justice. AI has the potential to foster a more equitable society by encouraging diversity and inclusion. Further research is needed to explore the long-term impact of AI on diversity and inclusion and to develop ethical AI frameworks tailored to different industries and cultural contexts.The study also explores how AI can support D&I initiatives by providing data-driven insights, automating processes, and offering personalized interventions to reduce bias and increase inclusivity. By reviewing recent research and case studies, this paper offers practical recommendations for leveraging AI in D&I efforts, aiming to build a more inclusive and equitable workplace culture.

AI-Driven Deep Learning Techniques in Protein Structure Prediction.

Protein structure prediction is important for understanding their function and behavior. This review study presents a comprehensive review of the computational models used in predicting protein structure. It covers the progression from established protein modeling to state-of-the-art artificial intelligence (AI) frameworks. The paper will start with a brief introduction to protein structures, protein modeling, and AI. The section on established protein modeling will discuss homology modeling, ab initio modeling, and threading. The next section is deep learning-based models. It introduces some state-of-the-art AI models, such as AlphaFold (AlphaFold, AlphaFold2, AlphaFold3), RoseTTAFold, ProteinBERT, etc. This section also discusses how AI techniques have been integrated into established frameworks like Swiss-Model, Rosetta, and I-TASSER. The model performance is compared using the rankings of CASP14 (Critical Assessment of Structure Prediction) and CASP15. CASP16 is ongoing, and its results are not included in this review. Continuous Automated Model EvaluatiOn (CAMEO) complements the biennial CASP experiment. Template modeling score (TM-score), global distance test total score (GDT_TS), and Local Distance Difference Test (lDDT) score are discussed too. This paper then acknowledges the ongoing difficulties in predicting protein structure and emphasizes the necessity of additional searches like dynamic protein behavior, conformational changes, and protein-protein interactions. In the application section, this paper introduces some applications in various fields like drug design, industry, education, and novel protein development. In summary, this paper provides a comprehensive overview of the latest advancements in established protein modeling and deep learning-based models for protein structure predictions. It emphasizes the significant advancements achieved by AI and identifies potential areas for further investigation.

Artificial neural network framework for the selection of deep eutectic solvents promoted enhanced oil recovery by interfacial tension reduction mechanism

Estimation of interfacial tension (IFT) and % reduction in IFT of Brazil and China regions’ heavy crude oil is employed by Artificial Intelligence (AI) led frameworks. The uniqueness of proposed approach lies in using AI frameworks that are constructed using Artificial Neural Network (ANN) and meta-heuristic algorithms i.e. Genetic Algorithm (GA) and Grey Wolf Optimization (GWO) forms three AI frameworks namely ANN, ANN-GA, and ANN-GWO is employed for IFT estimation of heavy oil and deep eutectic solvent-brine interface. The hyper-parameters of the frameworks are trained, validated, and tested over 18228 IFT data points obtained from the Omani heavy crude oil. The intermolecular interactions between oil, brine, and DES are obtained by generating surface segments in COnductor like Screening MOdel for Real Solvent (COSMO-RS) framework. The composition of brine, mole ratio of oil, DES, and temperature are other input parameters for AI model. The IFT estimation leads to the selection of 3 hydrophilic DESs capable of greater than 75 % IFT reduction and 14 hydrophilic DESs capable of 50–75 % IFT reduction. ANN-GWO scores best AI framework among 3 whereas ANN-GA fails to estimate IFT reduction greater than 50 %. Analysis of IFT with respect to mole fraction reveals the most suitable range of DES addition for IFT reduction.

Maritime Fuel Price Prediction of European Ports using Least Square Boosting and Facebook Prophet: Additional Insights from Explainable Artificial Intelligence

Prediction of bunker fuel spot prices at a port and understanding the dependence on key determinants is an arduous and challenging activity. The present work strives to analyze the temporal spectrum of daily spot prices of Very Low Sulphur fuel Oil (VLSFO), a critical bunker fuel, in five European Ports, Amsterdam, Antwerp, Gothenburg, Hamburg, and Rotterdam. The lack of prior research in the allied domain has motivated to undertake the modeling of VLSFO spot prices through the lens of applied predictive analytics. The Least Square Boosting (LSBoost) and Facebook Prophet algorithms are used to draw forecasts in multivariate framework leveraging constructs related to the same fuel prices at different ports, different fuel prices at the same ports, economic indicator, etc. The dynamics have been explicitly examined during the Russia-Ukraine military conflict. Additionally, Explainable Artificial Intelligence (XAI) frameworks have been used to demystify the influence of the chosen explanatory variables at a granular scale. The overall findings espouse the effectiveness of the predictive framework in accurately estimating spot prices of VLSFO in any of the selected ports, and the same heavily depends on VLSFO prices at different ports.

AI optimization framework using digital layouts of array structures: A case study for fuel cells

It is challenging for common artificial intelligence (AI) frameworks to optimize layouts of array structures, because AI is unable to recognize and adjust layouts. To address this challenge, a novel AI optimization framework has been proposed, where layouts of array structures are represented in digital codes to enable AI in layout optimizations. The optimizations of baffle layout and modular channel layout in proton exchange membrane fuel cells are used to validate the methodology. For the multi-objective optimization of baffle layout, although the channel with optimal baffle layout (OC) exhibits a little higher pressure drop than the conventional channel, it increases the net power density and oxygen uniformity by 2.4% and 1.5%, respectively. Consistently, for the multi-objective optimization of channel layout, the optimal flow field with an optimal channel layout (OFF-OC) shows a small increase of pressure drop that can be compensated by the much-enhanced output. So, the net power density of OFF-OC has been elevated by 7.0% in comparison with that of the conventional parallel flow field. Meanwhile, the oxygen uniformity of OFF-OC has also been improved by 4.0%. The simulation results have been verified experimentally. These examples have demonstrated the effectiveness of the AI optimization framework using digital layouts.

Artificial Intelligence for Sustainable Energy Transition: Optimising Renewable Energy Integration and Management

As climate change and long-term energy security drive the global energy sector towards renewable resources, powerful tools are required to optimise integration and management. A novel framework is proposed for effectively utilising Artificial Intelligence (AI) in the renewable energy landscape. AI algorithms can analyse weather patterns, historical generation data, and environmental factors to predict renewable energy output. Energy dispatch is optimised, grid integration is improved, and energy storage requirements are reduced. A system powered by artificial intelligence also significantly reduces downtime, optimises maintenance schedules, and minimises operational costs in wind turbines, solar panels, and other renewable infrastructure. AI can also optimise energy flows, reduce grid instability, and ensure efficient resource utilisation within the smart grid by dynamically managing renewable sources, energy storage systems, and demand profiles. Furthermore, AI-driven spatial analysis and resource mapping can identify optimal locations for renewable installations, considering factors like wind speed, solar irradiance, and environmental constraints. This paper presents two AI frameworks, one for solar energy and one for wind energy, to demonstrate possible applications. They both utilise comprehensive data acquisition, including real-time sensor data and external factors like weather forecasts and historical generation patterns. AI algorithms use these combined data to perform critical tasks such as predictive maintenance, minimising downtime, and maximising efficiency. Power output forecasting enables real-time adjustments based on weather, and optimal site selection maximises energy production. AI is used for proactive issue identification, accurate power output forecasting based on wind conditions, grid demand, storage capacity, dynamic load optimisation for maximum energy efficiency, and wind farm site selection. Integrating these tailored AI frameworks with solar and wind energy can achieve significant benefits such as increased efficiency, reduced operational costs, and seamless grid integration. In addition to analysing the challenges and opportunities associated with this AI integration, the paper explores infrastructure development, ethical considerations, and data acquisition. A second benefit of the research methodology is that it highlights how these tailored AI frameworks can optimise the integration of solar and wind renewable energy sources, providing valuable insights for researchers, practitioners, and policymakers who wish to use AI to create a more sustainable and efficient energy system. Keyword: Artificial Intelligence, renewable energy, climate change.

Prediction of Epidermal Growth Factor Receptor Mutation Subtypes in Non–Small Cell Lung Cancer From Hematoxylin and Eosin–Stained Slides Using Deep Learning

Accurate assessment of epidermal growth factor receptor (EGFR) mutation status and subtype is critical for the treatment of non–small cell lung cancer patients. Conventional molecular testing methods for detecting EGFR mutations have limitations. In this study, an artificial intelligence–powered deep learning framework was developed for the weakly supervised prediction of EGFR mutations in non–small cell lung cancer from hematoxylin and eosin–stained histopathology whole-slide images. The study cohort was partitioned into training and validation subsets. Foreground regions containing tumor tissue were extracted from whole-slide images. A convolutional neural network employing a contrastive learning paradigm was implemented to extract patch-level morphologic features. These features were aggregated using a vision transformer-based model to predict EGFR mutation status and classify patient cases. The established prediction model was validated on unseen data sets. In internal validation with a cohort from the University of Science and Technology of China (n = 172), the model achieved patient-level areas under the receiver-operating characteristic curve (AUCs) of 0.927 and 0.907, sensitivities of 81.6% and 83.3%, and specificities of 93.0% and 92.3%, for surgical resection and biopsy specimens, respectively, in EGFR mutation subtype prediction. External validation with cohorts from the Second Affiliated Hospital of Anhui Medical University and the First Affiliated Hospital of Wannan Medical College (n = 193) yielded patient-level AUCs of 0.849 and 0.867, sensitivities of 79.2% and 80.7%, and specificities of 91.7% and 90.7% for surgical and biopsy specimens, respectively. Further validation with The Cancer Genome Atlas data set (n = 81) showed an AUC of 0.861, a sensitivity of 84.6%, and a specificity of 90.5%. Deep learning solutions demonstrate potential advantages for automated, noninvasive, fast, cost-effective, and accurate inference of EGFR alterations from histomorphology. Integration of such artificial intelligence frameworks into routine digital pathology workflows could augment existing molecular testing pipelines.

Initial orbit determination via artificial intelligence for too-short arcs

In the framework of Space Traffic Management (STM) the growing number of space debris and the consequent increased risk of collisions have become a topic of great interest. Moreover, the difficulty of detecting and characterizing a Resident Space Object (RSO) in terms of orbital parameters makes this research field very challenging. In most cases the available observations have a very short duration and an accurate estimation of the orbit is very difficult. This paper introduces the Initial Orbit Determination via Artificial intelligence (IODA) algorithm, a reliable and accurate orbit estimator combining metaheuristic and artificial intelligence frameworks to perform precise orbit determination from Very Short Arcs (VSAs). Starting from an initial orbit estimate derived through an Artificial Bee Colony (ABC) algorithm, a Physics Informed Extreme Learning Machine is employed to increase the accuracy of the predicted orbit. Results on simulated data have shown the robustness of the algorithm to random parameters initialization and a percentage of success of 89% of retrieving the reference orbit with an error under 1 km and 0.01 km/s on the initial target’s state.

AIfES: A Next-Generation Edge AI Framework.

Edge Artificial Intelligence (AI) relies on the integration of Machine Learning (ML) into even the smallest embedded devices, thus enabling local intelligence in real-world applications, e.g. for image or speech processing. Traditional Edge AI frameworks lack important aspects required to keep up with recent and upcoming ML innovations. These aspects include low flexibility concerning the target hardware and limited support for custom hardware accelerator integration. Artificial Intelligence for Embedded Systems Framework (AIfES) has the goal to overcome these challenges faced by traditional edge AI frameworks. In this paper, we give a detailed overview of the architecture of AIfES and the applied design principles. Finally, we compare AIfES with TensorFlow Lite for Microcontrollers (TFLM) on an ARM Cortex-M4-based System-on-Chip (SoC) using fully connected neural networks (FCNNs) and convolutional neural networks (CNNs). AIfES outperforms TFLM in both execution time and memory consumption for the FCNNs. Additionally, using AIfES reduces memory consumption by up to 54% when using CNNs. Furthermore, we show the performance of AIfES during the training of FCNN as well as CNN and demonstrate the feasibility of training a CNN on a resource-constrained device with a memory usage of slightly more than 100 kB of RAM.