Adaptive AI Research Articles

ENHANCING ADAPTIVE LEARNING AND DECISION-MAKING SYSTEMS USING SWARM INTELLIGENCE AND DEEP LEARNING FOR ADVANCED AI APPLICATIONS

The rapid development of autonomous vehicles (AVs) demands robust and adaptive AI systems capable of handling complex real-world environments. Traditional optimization and learning algorithms often struggle with dynamic and uncertain conditions, leading to suboptimal decision-making. Swarm intelligence, particularly Hawk Fire Optimization (HFO), offers a promising solution by simulating cooperative behaviors seen in nature, like hawks in hunting, to optimize decision-making processes. Coupled with advanced deep learning techniques like Federated Dropout Learning (FDL), this hybrid approach can enhance the adaptability, scalability, and efficiency of AI systems. This paper addresses the challenge of improving decisionmaking and learning in autonomous vehicles by integrating HFO with FDL. HFO optimizes parameters in real-time, allowing AVs to adapt rapidly to changing environments. Federated Dropout Learning, a variant of federated learning, further improves system resilience by sharing learning across distributed nodes while minimizing communication overhead and enhancing privacy. By combining these methods, the proposed system ensures robust performance in unpredictable scenarios. Experimental results show that the hybrid model outperforms traditional methods in terms of decision accuracy, response time, and energy efficiency. Specifically, the system achieved a 12% improvement in decision accuracy, reduced processing time by 18%, and cut energy consumption by 22%, compared to standard algorithms. These findings suggest that the combination of HFO and FDL can significantly improve the performance of autonomous vehicles, providing safer and more efficient AI-driven navigation.

NOVEL ADAPTIVE AI-POWER INSULIN NAVIGATOR FOR DIABETES USING MACHINE LEARNING

Adaptive AI-powered insulin delivery systems represent a significant advancement in diabetes management, offering enhanced precision and safety for patients with diabetes. This paper examines the development and implementation of AI-driven insulin delivery systems that dynamically adjust insulin dosages based on real-time glucose monitoring and predictive analytics. The system integrates continuous glucose sensors with adaptive AI algorithms to analyze blood glucose trends, activity levels, and dietary inputs, enabling the automatic adjustment of insulin delivery to meet individual patient needs. The study evaluates the performance of these systems in optimizing glycemic control, reducing the incidence of hypoglycemia and hyperglycemia, and improving overall patient outcomes. Key features include the ability to learn from historical data to refine dosing algorithms and the incorporation of safety mechanisms to prevent adverse events. Additionally, the paper explores user experiences and feedback, highlighting the system’s potential to improve quality of life and provide more personalized diabetes management. The integration of adaptive AI in insulin delivery systems promises to revolutionize diabetes care by offering a more responsive, accurate, and user-centric approach to insulin therapy.

A Hierarchical Model for Complex Adaptive System: From Adaptive Agent to AI Society

As complex adaptive system involves human and social factors (e.g., changing demands, competition and collaboration among agents), accurately modeling the complex features of adaptive agents and AI society is crucial for the effective analysis and governance of complex adaptive systems. However, existing modeling methods struggle to accurately represent these complex features, there is a gap between existing technologies and complex features modeling. In this context, this paper proposes a hierarchical model based on the computational experiments method, which consists of four layers (i.e., L1, L2, L3 and L4) modeling the autonomous, evolutionary, interactive, and emergent features respectively from adaptive agent to AI society. Additionally, taking intelligent transportation system as an example, a computational experiments system is constructed to demonstrate the effectiveness of the proposed model. This model builds a bridge between complex feature modeling and various technologies, thereby offering theoretical support for further research in complex adaptive systems.

When digital spaces matter: The influence of uniqueness and place attachment on self‐identity expression with brands using generative AI on the metaverse

AbstractGenerative artificial intelligence technologies are transforming brand‐customer interactions by offering highly personalized and dynamic virtual experiences. This underscores the importance of studying their significant impact on customer experiences in the Metaverse. Accordingly, the study aims to explore how users' attachment to the Metaverse relates to the underlying psychological drivers of customer engagement with brands using generative AI in the digital realm. A mixed‐method approach was adopted, beginning with an exploratory study that involved in‐depth interviews with 24 participants to gain an initial understanding of consumers' psychological and behavioral reactions to brands in the Metaverse. An empirical study was then conducted, surveying 407 UK‐based users of virtual worlds who interacted with brands using generative AI on these platforms. The goal was to understand how the need for uniqueness influences self‐identity expression and to examine how these factors are affected by attachment to virtual places. The research highlights the nuanced relationship between the pursuit of uniqueness and the integration of brand narratives into one's identity particularly when supported by highly personalized, adaptive AI technologies.

A Review on DISDAIN: An Auto Content Generation VR Game

This document discusses the manner in which DISDAIN has been conceptualized, developed, implemented, and evaluated, looking into its major mechanics, algorithm- based content creation as well as user interaction and all impact it has made on virtual reality gaming. It explores DISDAIN’s design and implementation with much detail to highlight how procedural content generation can improve replayability and immersion in virtual reality games. The results from user testing and gameplay analysis indicate that the game has been successful in providing various experiences that are interesting while at the same time ensuring high levels of player contentment. Besides, this study analyzes technical performance and adaptive AI of DISDAIN so as to project what may be expected from dynamic content generation in VR gaming in future. In general, these results indicate that DISDAIN is a starting point for new developments in VR gaming henceforth while at the same time setting new standards for this kind of game category

Adaptive AI

Adaptive AI is a smart new technology that changes its re-actions to adapt to a new situation and continues patterns. As it collects information from previous experiences and adapt over time. It can solve occurring problems as well as dealing with changing conditions in businesses.

Adaptive and Flexible Model-Based AI for Deep Receivers in Dynamic Channels

Adaptive and Flexible Model-Based AI for Deep Receivers in Dynamic Channels

Towards interculturally adaptive conversational AI

Abstract Among the many ways that AI technologies are becoming embedded in our social worlds is the proliferation of Conversational User Interfaces, such as voice assistants (e.g. Apple Siri and Amazon Alexa), chatbots and voice-based conversational agents. Such conversational AI technologies are designed to draw upon the designers’ understanding of interactional practices employed in human–human conversation, and therefore have implications for intercultural communication (ICC). In this paper, we highlight some of the current shortcomings of conversational AI, and how these relate to ICC. We also draw on findings from Conversation Analysis to discuss how pragmatic norms vary across linguacultural groups (see Risager, Karen. 2019. Linguaculture. In Carol A. Chapelle (ed.). Encyclopedia of applied linguistics. Wiley-Blackwell for a discussion of the term ‘linguaculture’), noting that this poses further challenges for designers of conversational AI systems. We argue that the solution is to work towards what we call interculturally adaptive conversational AI. Finally, we propose a framework for how this can be conceptualised and researched, and argue that researchers with expertise in language and ICC are uniquely placed to contribute to this endeavour.

Advancing Warehouse Management Systems: Optimizing Loading-Unloading, Conditioning, Packing and Marking Processes with Adaptive AI Technology

Advancing Warehouse Management Systems: Optimizing Loading-Unloading, Conditioning, Packing and Marking Processes with Adaptive AI Technology

Survey of Advanced Nonlinear Control Strategies for UAVs: Integration of Sensors and Hybrid Techniques.

This survey paper explores advanced nonlinear control strategies for Unmanned Aerial Vehicles (UAVs), including systems such as the Twin Rotor MIMO system (TRMS) and quadrotors. UAVs, with their high nonlinearity and significant coupling effects, serve as crucial benchmarks for testing control algorithms. Integration of sophisticated sensors enhances UAV versatility, making traditional linear control techniques less effective. Advanced nonlinear strategies, including sensor-based adaptive controls and AI, are increasingly essential. Recent years have seen the development of diverse sliding surface-based, sensor-driven, and hybrid control strategies for UAVs, offering superior performance over linear methods. This paper reviews the significance of these strategies, emphasizing their role in addressing UAV complexities and outlining future research directions.

Adaptive AI in precision agriculture: A review: Investigating the use of self-learning algorithms in optimizing farm operations based on real-time data

Adaptive AI in precision agriculture: A review: Investigating the use of self-learning algorithms in optimizing farm operations based on real-time data

AI model for predicting adult cochlear implant candidacy using routine behavioral audiometry

ObjectiveTo describe an AI model to facilitate adult cochlear implant candidacy prediction based on basic demographical data and standard behavioral audiometry. MethodsA machine-learning approach using retrospective demographic and audiometric data to predict candidacy CNC word scores and AzBio sentence in quiet scores was performed at a tertiary academic center. Data for the model were derived from adults completing cochlear implant candidacy testing between January 2011 and March 2023. Comparison of the prediction model to other published prediction tools and benchmarks was performed. ResultsThe final dataset included 770 adults, encompassing 1045 AzBio entries, and 1373 CNC entries. Isophoneme scores and word recognition scores exhibited strongest importance to both the CNC and AzBio prediction models, followed by standard pure tone average and low-frequency pure tone average. The mean absolute difference between the predicted and actual score was 15 percentage points for AzBio sentences in quiet and 13 percentage points for CNC word scores, approximating anticipated test-retest constraints inherent to the variables incorporated into the model. Our final combined model achieved an accuracy of 87 % (sensitivity: 90 %; precision: 80 %). ConclusionWe present an adaptive AI model that predicts adult cochlear implant candidacy based on routine behavioral audiometric and basic demographical data. Implementation efforts include a public-facing online prediction tool and accompanying smartphone program, an embedded notification flag in the electronic medical record to alert providers of potential candidates, and a program to retrospectively engage past patients who may be eligible for cochlear implantation based on audiogram results.

An Adaptive Explainable AI Framework for Securing Consumer Electronics-Based IoT Applications in Fog-Cloud Infrastructure

An Adaptive Explainable AI Framework for Securing Consumer Electronics-Based IoT Applications in Fog-Cloud Infrastructure

EEG Brainwave Controlled Robotic Arm for Neurorehabilitation Training

This study presents an EEG-based control method using AI for a robotic arm neurorehabilitation system. The system employs advanced EEG technology to capture and interpret brainwaves, utilizing AI algorithms for analysis and translation. This enables users to control the robotic arm via neural signals during neurorehabilitation training. Real-time feedback and adaptive AI ensure a personalized, interactive experience, and adjust the program based on user progress. This approach aims to enhance neurorehabilitation outcomes by promoting neuroplasticity and motor skill recovery in individuals with neurological disorders or injuries. The study’s results highlight AI’s positive impact on the system, suggesting EEG-based AI control could play a pivotal role in neurorehabilitation.

Exploring interactive video learning: Techniques, applications, and pedagogical insights

Multimedia, interaction, and customized content make interactive video learning a popular and effective teaching tool. This pedagogical technique uses movies, animations, and graphics to create engaging and attractive instructional materials. Interactive aspects such as quizzes, simulations, and decision-making encourage active learning and critical thinking. Personalization is critical because it allows the educational process to be tailored to individual needs and preferences. This is done through adaptive learning systems and AI algorithms. This review examines interactive video-based learning methods, applications, and educational effects. This study examines how interactive video learning is impacting education in many sectors and locations. This technique allows students to view interactive videos at their convenience, accommodating different learning styles and schedules. The study also examines how interactive video-based learning affects knowledge, motivation, and participation in formal education and corporate training. This review analyzes existing studies and trends to assess the usefulness and future of interactive video learning in modern education. Interactivity, personalization, and multimedia help provide effective, learner-centered instruction. This study shows that interactive video learning can improve pedagogy and educational outcomes.

Initial Clinical Experience Using Daily Artificial Intelligence-Assisted Adaptive Radiotherapy on Cone-Beam CT for Cancer of the Head and Neck

Adaptive radiation therapy (ART) has emerged as a promising method to modify treatment plans in response to structural and spatial changes that occur during a treatment course. In particular, patients with locally advanced Head and Neck (HN) Cancers often experience changes in target volumes and organs-at risk requiring replanning throughout treatment. The implementation of daily ART using daily cone-beam Cone-Beam CT (CBCT) imaging is a solution that may improve the therapeutic ratio by accounting for intra-fractional changes throughout the patient's treatment course. We evaluated the initial clinical experience of daily ART for patients with locally advanced HN cancer using a commercially available, online adaptive platform with AI-assisted workflows on daily CBCT. For treatment, AI-based auto-contour and structure deformation of OARs and Targets were reviewed and edited by the treating physician. Two plans were generated including, a CT sim-based plan with deformed structures (scheduled) and a re-optimized plan (adaptive) of which both plans evaluated and the best one approved and delivered. Clinical outcomes and dosimetric parameters were reviewed. A total of 23 patients with locally advanced HN cancers (7 Nasopharynx, 6 Oropharynx, 2 oral cavity, 8 larynx) were treated with daily ART from December 2020 until November 2022. 655 adaptive and scheduled radiotherapy plans were generated of which 95.7% of the adaptive plans were chosen. Adaptive plans demonstrated superior V95 PTV70, PTV60, PTV56 mean coverage by 1.02%, 8.14%, and 7.2%, respectively. At a median follow up of 10.1 months, the local control was 91.3%. The 1-year actuarial PFS was 86.9% and 1 year actuarial OS was 82.6%. Two patients required prophylactic g-tube. 30% and 21% of patients had acute grade 3 dysphagia and odynophagia, respectively; 13% with chronic grade 2 xerostomia. There were no grade 4 or grade 5 toxicities. These findings provide evidence for the feasibility and potential clinical benefit of incorporating daily adaptive AI radiotherapy on CBCT in the treatment of locally advanced HN cancer. Further prospective studies are needed to confirm these findings and determine the optimal clinical applications of this approach.

A Study on the Influence of Sensors in Frequency and Time Domains on Context Recognition

Adaptive AI for context and activity recognition remains a relatively unexplored field due to difficulty in collecting sufficient information to develop supervised models. Additionally, building a dataset for human context activities “in the wild” demands time and human resources, which explains the lack of public datasets available. Some of the available datasets for activity recognition were collected using wearable sensors, since they are less invasive than images and precisely capture a user’s movements in time series. However, frequency series contain more information about sensors’ signals. In this paper, we investigate the use of feature engineering to improve the performance of a Deep Learning model. Thus, we propose using Fast Fourier Transform algorithms to extract features from frequency series instead of time series. We evaluated our approach on the ExtraSensory and WISDM datasets. The results show that using Fast Fourier Transform algorithms to extract features performed better than using statistics measures to extract features from temporal series. Additionally, we examined the impact of individual sensors on identifying specific labels and proved that incorporating more sensors enhances the model’s effectiveness. On the ExtraSensory dataset, the use of frequency features outperformed that of time-domain features by 8.9 p.p., 0.2 p.p., 39.5 p.p., and 0.4 p.p. in Standing, Sitting, Lying Down, and Walking activities, respectively, and on the WISDM dataset, the model performance improved by 1.7 p.p., just by using feature engineering.

Методика построения защищенных систем искусственного интеллекта для проведения электроретинографии в офтальмологии

Relevance. It is known that the method of electroretinography (hereinafter – ERG) in ophthalmology, which works on the key registration of changes in the bioelectric potential of the retina, uses the potential of exposure to light passing through the optical media of the eye. A similar method is conditionally applicable to the transmission of a light pulse over a fiber-optic cable, during which the correct transmission of information is carried out. If there is a violation or change in the electrical potential, there is every reason to believe that a person has any diseases. Purpose. To develop a technology for creating intellectual information security systems (IISS) is of a complex nature in which a quasi-biological paradigm is put in the first place, where the form of programming information processes, machine learning systems (MLS) and the construction of neural systems and ending with an AI architecture with built – in mechanisms for ensuring information security. Material and methods. In this article, a methodology for using a new artificial intelligence system (hereinafter referred to as AI) in a protected version for working with corrective devices for electroretinography in ophthalmology is compiled. Results. A set of methodological and scientific and technical solutions for the artificial intelligence system has been developed in order to ensure its «viability» and resistance to computer attacks aimed at violating the integrity. Conclusion. The article raises a key issue – the need to develop software architectural solutions for AI and adaptive neuro-fuzzy AI, which have «neurons» built into the software and hardware information protection systems with a universal set of commands for EG devices. Keywords: artificial intelligence, multi-agent system, electroretinography, neural network, integrity control, information security

Logistics 4.0 in intermodal freight transport

While Industry 4.0 is the use of cyber-physical systems in manufacturing and supply chains, Logistics 4.0 is the use of similar systems within logistics (robotics and material handling) such as adaptive programming and AI. The EU's environmental policy - more precisely - the pressing need to reduce CO2 emissions from freight requires technological advancement. A possible solution to the problem is to increase the proportion of rail freight instead of road freight. This is not a new idea, but recent decades brought modest solutions. By examining current rail-road intermodal freight transport and container handling procedures, we established a new freight transport model and propose an efficient automated container handling procedure that will bring us closer to the environmental objectives. Efficient horizontal container handling proposed in this paper can facilitate growth in continental intermodal freight transport. The technological and process solutions presented in this paper increase competitiveness in terms of reducing costs and transportation times.

The Placebo Effect of Artificial Intelligence in Human–Computer Interaction

In medicine, patients can obtain real benefits from a sham treatment. These benefits are known as the placebo effect. We report two experiments (Experiment I: N = 369; Experiment II: N = 100) demonstrating a placebo effect in adaptive interfaces. Participants were asked to solve word puzzles while being supported by no system or an adaptive AI interface. All participants experienced the same word puzzle difficulty and had no support from an AI throughout the experiments. Our results showed that the belief of receiving adaptive AI support increases expectations regarding the participant’s own task performance, sustained after interaction. These expectations were positively correlated to performance, as indicated by the number of solved word puzzles. We integrate our findings into technological acceptance theories and discuss implications for the future assessment of AI-based user interfaces and novel technologies. We argue that system descriptions can elicit placebo effects through user expectations biasing the results of user-centered studies.