Reduce Resource Costs Research Articles

Development of intelligent systems for monitoring and management of agricultural enterprises

Modern agriculture faces a number of challenges, among which are the growing demand for food, the impact of climate change, limited natural resources and the need to ensure the environmental sustainability of production. In this context, the introduction of new technologies, in particular intelligent systems, becomes an important tool for improving the efficiency and competitiveness of agro-industrial enterprises. Intelligent Systems Based on Use artificial intelligence, machine learning, the Internet of Things (IoT) and big data are opening up new opportunities for monitoring and managing agricultural processes. The article presents the results of a study aimed at the development of intelligent systems that provide comprehensive monitoring of the state of fields and management of production processes at agricultural enterprises. The proposed system integrates data from various sources, such as sensors, drones, satellite images and other IoT devices, which allows you to create a detailed and up-to-date view of the state of crops, soils and climatic conditions. This, in turn, allows you to make prompt and sound management decisions that increase the productivity and efficiency of agricultural production. Particular attention is paid to the development of algorithms for processing big data and machine learning methods, which are used to analyze the data obtained and generate recommendations for optimizing the use of resources. This includes managing water resources, fertilizers, crop protection products, as well as predicting yields and identifying potential threats such as plant diseases or pests. The implementation of such systems allows not only to reduce resource costs, but also to minimize the environmental impact of agriculture by reducing the use of chemicals and water. Thus, the article makes a significant contribution to the development of intelligent technologies in agriculture, offering new approaches to the management and monitoring of agricultural enterprises. These approaches can form the basis for the development of future innovations in this area, contributing to improving the resilience and efficiency of agro-industrial production in response to modern challenges.

Integrated Sensing, Communication, and Computing for Cost-effective Multimodal Federated Perception

Federated learning (FL) is a prominent paradigm of 6G edge intelligence (EI), which mitigates privacy breaches and high communication pressure caused by conventional centralized model training in the artificial intelligence of things (AIoT). The execution of multimodal federated perception (MFP) services comprises three sub-processes, including sensing-based multimodal data generation, communication-based model transmission, and computing-based model training, ultimately competitive on available underlying multi-domain physical resources such as time, frequency, and computing power. How to reasonably coordinate the multi-domain resources scheduling among sensing, communication, and computing, therefore, is vital to the MFP networks. To address the above issues, this article explores service-oriented resource management with integrated sensing, communication, and computing (ISCC). Specifically, employing the incentive mechanism of the MFP service market, the resources management problem is defined as a social welfare maximization problem, where the concept of “expanding resources” and “reducing costs” is used to enhance learning performance gain and reduce resource costs. Experimental results demonstrate the effectiveness and robustness of the proposed resource scheduling mechanisms.

Implementation of circular economy principles to promote the development of rural areas

The relevance of the study lies in identifying and explaining the impact of implementing circular economy technologies on the development of rural areas within the framework of the state strategy for regional development in Ukraine, thereby contributing to ensuring the socio-economic stability of communities and the overall security of the country. The purpose of this study was to investigate the influence of utilising circular economy opportunities on the development of rural territories. The abstract-logical method was employed to substantiate the research methodology and formulate theoretical conclusions. The analytical and comparative methods was utilised to organise researchers' scientific approaches to theoretical aspects of circular economy development, refine the methodological basis, and substantiate the applicability of circular economy principles in rural development. The research utilised a graphical method to illustrate the findings and an abstract-logical approach to draw conclusions and recommendations. The study was divided into three main stages. The literature on the contemporary definition of "circular economy" was examined, and the driving forces and barriers to systemic circular economy in rural areas were analysed. The study investigated how European Union countries apply strategies to address environmental and economic issues through the lens of circular economy. The analysis of the experience in implementing circular economy practices in EU countries revealed a clear trend towards more sustainable resource utilisation and reduced environmental impact. The implementation of the "Farm to Fork" strategy under the European Green Deal is a crucial step towards ensuring sustainable resource use and implementing environmentally friendly practices in agriculture and the food industry. An assessment of the impact of implementing circular economy on the socio-economic development of rural areas was conducted, indicating that the development of circular models can be key to optimising resource utilisation, creating new employment opportunities, improving production efficiency, and increasing the profitability of agricultural enterprises. The implementation of circular methods in agriculture can reduce resource costs and provide more stable conditions for development. The results can be beneficial for government bodies, local administrations, agricultural enterprises, and cooperatives in Ukraine striving to achieve sustainable development of rural territories through the implementation of circular approaches in the economy.

A method for generating the digital twin models of plants

In the process of creating digital twin models of plants, the simulation accuracy of models generated by conventional 3D reconstruction methods is insufficient. Considering the increasing number of models over time, adopting deep learning-based models not only consume a significant amount of time but also require substantial computational resources consumption. Therefore, we proposed a method for generating digital twin models of plants, aiming to ensure the simulation credibility while effectively reducing resource costs. Experimental results show that this method is capable of generating plant models that closely resemble physical entities at different growth periods. Compared with other construction methods, this approach achieves higher simulation credibility (CD=0.089, EMD=0.034, KNN=0.005, RGB=4.97) with lower computational resources consumption (ROM=7.46MB, VRAM=2456MB, Time=0h5min48s) when generating a digital twin model.

AMFL: Resource-Efficient Adaptive Metaverse-Based Federated Learning for the Human-Centric Augmented Reality Applications.

The emergence of 5G technology has enabled the development of Metaverse applications that provide users with immersive experiences through augmented reality (AR) devices, and the integration of federated learning (FL) with the Metaverse AR (MAR) systems can enable many edge intelligence services in 5G. However, the presence of nonindependent and identically distributed (Non-IID) data across all AR users' devices, coupled with limited edge communication resources, makes it challenging to achieve human-centric Metaverse-related applications such as target detection or image classification that combine virtual content with real-world. To address these challenges, we propose a novel adaptive resource-efficient Metaverse-based FL (AMFL) algorithm for AR applications that mitigates the negative effect of Non-IID data and reduces resource costs as well as improves the quality of experience (QoE). We first analyze the impact of wireless communication factors such as CPU frequency, bandwidth, and transmission power on FL training performance by a toy example in the MAR systems. Based on this analysis, furthermore, we establish a Non-IID degree, model accuracy, and resource consumption-related QoE maximization problem under given resource budgets, which is a stochastic optimization problem with strongly coupled variables, including bandwidth, CPU frequency, and transmission power. Guided by the theoretical analysis, to solve this issue, AMFL employs a deep reinforcement learning (DRL)-based method to adaptively allocate resources. Numerical results demonstrate that AMFL can significantly improve the QoE by up to 30.28 % , and reduce communication round and energy costs by up to 81.08 % and 72.20 % , respectively, even under the worst Non-IID case, compared to benchmarks.

ЭКОНОМИЧЕСКОЕ ВЛИЯНИЕ ТОЧНОГО ЗЕМЛЕДЕЛИЯ НА УСТОЙЧИВОЕ РАЗВИТИЕ СЕЛЬСКОГО ХОЗЯЙСТВА

The article evaluates the economic impact of precision farming technologies on the sustainable development of agriculture. The research is aimed at analyzing the economic efficiency of the introduction of technologies such as satellite monitoring, differential fertilization and the use of unmanned aerial vehicles. The methods used include a comparative analysis of the costs of traditional and innovative technologies, as well as an assessment of their impact on productivity and environmental sustainability. The main results of the study show that precision farming technologies contribute to reducing resource costs and increasing yields. In particular, the use of differentiated fertilization technology reduces the use of pesticides by 15%, and satellite monitoring ensures a 20% increase in yield. The peculiarity of the work is to substantiate the economic feasibility of the introduction of digital technologies and their contribution to the development of agricultural production.

Building practice of energy efficient earthquake-resistance buildings

The article discusses technologies for constructing energy-efficient, earthquake-resistant buildings that have increased rigidity and strength in their main components. These technologies utilize structures with enhanced damping and active seismic insulation systems. Practical implementations, design features, and technological aspects are examined. An analysis is conducted on the characteristics of current technologies, the conditions for production and installation, repairability, and their impact on the building's energy efficiency during operation. Technologies are systematized and categorized based on their principles of seismic protection during the analysis process. The analysis was conducted based on the following main criteria: steel consumption, suitability of repairable technologies used, possibility of implementing seismic protection during building reconstruction, ability to withstand seismic effects, feasibility of implementation, and energy efficiency. It has been determined that traditional technical solutions to increase rigidity and strength, such as introducing structures with increased damping used for buildings with precast, monolithic, and precast-monolithic frames, require a significant amount of additional resources. These solutions are characterized by accumulation of failures in individual hidden elements due to seismic loads, low maintainability, and often decrease the energy efficiency of the building. The use of seismic energy-absorbing mechanisms with high efficiency can significantly reduce resource costs for construction and increase maintainability, thus improving the overall sustainability of the structure. Based on the findings of the analysis, the following recommendations are made for conducting further studies on active protection systems in terms of design, development of construction technology, and potential application in the renovation of existing buildings.

Modern algorithm of linear and quadratic programming in optimization and problems of deformation of structures of variable structure in contact conditions

Various mathematical problems, in which the goal of finding the extremum of a functional is set, belong to the problems of mathematical programming and optimization problems. Practically directed problems of finding the optimal solution are extremely numerous in economics, management, technology, and others. They are related to increasing production efficiency, reducing resource costs, improving design solutions and technological processes, reducing mass, dimensions, etc. Among them, an important role is given to the methods of limiting the maximum stresses caused by external loads. Solving such problems begins with mathematical formalization. Constructive, economic or technological indicators are chosen as variation parameters. The search for the best solution is reduced to the selection of a set of parameters that provide a stationary value of the objective function. Extreme problems of practical orientation contain equality-inequality constraints in mathematical models. In improving the technical characteristics of machines, a significant role belongs to engineering and technical workers, who find optimal options at the design stage. At the same time, an essential element of the design process is the modeling of the determining processes in structures, taking into account the main influencing factors and behavior scenarios. Optimization is an important area of applied mathematics that provides effective tools for such modeling. Universal Algorithm is proposed in work [3] – a numerous scheme for solving quadratic programming (QP) problems for calculating the optimal point of a wide range of applied problems. At the same time, the linear programming (LP) problem is considered as a partial case of the (QP) problem. That is, in the universal algorithm for setting 2 optimization problems, they are formalized in a single and convenient form of symmetric matrix dependence, which makes it possible to build a single effective algorithm based on matrix algebra operations. In particular, it allows you to consider the practical tasks of calculating VAT in constructions of a variable structure consisting of separate parts connected by one-way connections. The main goal of this work is to analyze the behavior of the algorithm when increasing the number of constraints of the inequality type, to refine the computational scheme, and to formulate conclusions. Two model problems are considered as examples of the algorithm. This is a classic "transport" problem of LP and the behavior of a model of a bridge structure with one-way connections in cables under variations of wind loads. The number of ropes has been increased to 20, and the limits of one-way connections to 40.

Utilisation of livestock by-products for resource-saving biogas production in industrial pork production

Modern energy systems are complex multi-component systems that use mechanical, thermal, and electrical energy. As fossil fuel reserves decline, interest in renewable energy sources is growing everywhere, which is driving research into biogas production technologies. The purpose of this study was to evaluate technological approaches to reduce resource costs to produce renewable energy sources from organic waste (manure) from pig farms. This study employed the bibliographic method of research, laboratory method, biochemical methods (determination of the chemical composition of manure and wastewater), statistical, mathematical (calculation of economic efficiency), multicriteria analysis, and analytical method. Based on the conducted research, a resource-saving technology for the preparation of wastewater for use in the production of renewable energy in industrial pork production was developed, which ensures a reduction in capital and operating costs for biogas generation. It was found that the settling of the initial wastewater with a moisture content of 96.94±0.18% produces a sediment with a moisture content of 91.23±0.25% and a liquid fraction with a moisture content of 98.86±0.42%. At the same time, the organic matter in the sediment extracted from the original wastewater is about 87%, and in the liquid fraction – about 98% of the dry matter. The technological process includes centrifugal and gravity separation of the initial wastewater into fractions (sediment and liquid fraction), gravitational thickening of the sediment and its dosing for methane digestion to produce biogas. It was shown that the application of the proposed approaches can reduce capital investments by about 30%. Based on the multicriteria analysis, it was found that the proposed biogas production technology has a significant advantage in terms of the complex efficiency indicator of each of the considered options N(Ck) compared to the idealised one. For this technology, the objective function is the smallest according to the criteria considered and amounts to 0.1672, while the objective function of the baseline technology is 1.9 times worse. The findings of the study on the use of livestock by-products for resource-saving biogas production are planned to be used at pork production complexes and farms

Sonnet: A control-theoretic approach for resource allocation in cluster management

Cluster users expect to minimize the resource costs while ensuring target performance for different applications. It is particularly difficult to reach such a goal, because the applications are diverse with dynamic load changes, and interference exists between them. In addition, the performance of the applications depends on heterogeneous resources with different costs. However, existing works either use simplistic and generalized heuristics that disregard resource-specific characteristics or need suspending service to get expert knowledge to optimize the resource cost for a brand-new application or runtime, which fails to optimize the resource allocation finely.In this paper, we propose Sonnet, a control-theoretic approach to perform efficient resource allocation. Sonnet can efficiently optimize the cost of resources while satisfying the SLO by quickly establishing new application performance models through only online profiling and without affecting service. Experiments on Docker Swarm using various open-source benchmarks demonstrate that Sonnet can decrease the SLO violation rate by 91% while reducing resource costs up to 47% compared with the state-of-the-arts.

АЛГОРИТМ МНОГОКРИТЕРИАЛЬНОГО АНАЛИЗА ТЕ КСТОВОЙ ИНФОРМАЦИИ

The article presents the author's algorithm for multi-criteria analysis of textual information, the main purpose of which is to increase the efficiency of emergency services, including the Ministry of Emergency Situations of Russia. This algorithm is based on a comprehensive assessment of information coming from social networks using a range of different types of heterogeneous data studies.
 The developed algorithm will speed up data collection, optimize their analysis, develop preventive scenarios for possible incidents, as well as reduce resource costs, optimize management proposals and justify decisions made.

AHPA: Adaptive Horizontal Pod Autoscaling Systems on Alibaba Cloud Container Service for Kubernetes

The existing resource allocation policy for application instances in Kubernetes cannot dynamically adjust according to the requirement of business, which would cause an enormous waste of resources during fluctuations. Moreover, the emergence of new cloud services puts higher resource management requirements. This paper discusses horizontal POD resources management in Alibaba Cloud Container Services with a newly deployed AI algorithm framework named AHPA - the adaptive horizontal pod auto-scaling system. Based on a robust decomposition forecasting algorithm and performance training model, AHPA offers an optimal pod number adjustment plan that could reduce POD resources and maintain business stability. Since being deployed in April 2021, this system has expanded to multiple customer scenarios, including logistics, social networks, AI audio and video, e-commerce, etc. Compared with the previous algorithms, AHPA solves the elastic lag problem, increasing CPU usage by 10% and reducing resource cost by more than 20%. In addition, AHPA can automatically perform flexible planning according to the predicted business volume without manual intervention, significantly saving operation and maintenance costs.

РЕЗУЛЬТАТИ ДОСЛІДЖЕННЯ ВПЛИВУ ЯКІСНИХ ПОКАЗНИКІВ ВИКОРИСТАННЯ РУХОМОГО СКЛАДУ НА СОБІВАРТІСТЬ ВАНТАЖНИХ ПЕРЕВЕЗЕНЬ

The article presents the results of a study of the impact on the cost of freight transportation of qualitative indicators of the use of rolling stock. The cost of transportation is a complex economic indicator. This indicator evaluates the costs of all types of resources for transportation. Reducing the cost of transportation indicates a reduction in resource costs. The cost of freight transportation is influenced by the type of traction, labor productivity, material consumption and energy intensity of transportation. The change in the qualitative indicators of the use of rolling stock has a significant impact on the economic indicators of the operational activities of railway transport. The cost of freight transportation can be reduced with the efficient use of rolling stock. This is possible by increasing the dynamic load of the loaded car, the mass of the train, the productivity of the freight car and the locomotive, sectional speed, and the average daily mileage of the freight car and the locomotive. Two areas of research are distinguished on the basis of determining the cost of transportation. The first direction is to identify the relationship between costs and traffic volumes. The second direction is to identify the dependence of costs on individual meters. The impact of qualitative indicators of the use of rolling stock on the cost of freight transportation is determined by the method of unit cost rates. This method establishes the dependence of the cost of freight transportation on the qualitative indicator of the use of rolling stock. As a result of the study, qualitative indicators of the use of rolling stock have been identified, which have a significant impact on the cost of freight transportation. The results of the study can be used for further development of technical and operational measures to improve the efficiency of the railway transport of Ukraine.

Resource Allocation With Workload-Time Windows for Cloud-Based Software Services: A Deep Reinforcement Learning Approach

As the workloads and service requests in cloud computing environments change constantly, cloud-based software services need to adaptively allocate resources for ensuring the Quality-of-Service (QoS) while reducing resource costs. However, it is very challenging to achieve adaptive resource allocation for cloud-based software services with complex and variable system states. Most of the existing methods only consider the current condition of workloads, and thus cannot well adapt to real-world cloud environments subject to fluctuating workloads. To address this challenge, we propose a novel Deep Reinforcement learning based resource Allocation method with workload-time Windows (DRAW) for cloud-based software services that considers both the current and future workloads in the resource allocation process. Specifically, an original Deep Q-Network (DQN) based prediction model of management operations is trained based on workload-time windows, which can be used to predict appropriate management operations under different system states. Next, a new feedback-control mechanism is designed to construct the objective resource allocation plan under the current system state through iterative execution of management operations. Extensive simulation results demonstrate that the prediction accuracy of management operations generated by the proposed DRAW method can reach 90.69%. Moreover, the DRAW can achieve the optimal/near-optimal performance and outperform other classic methods by 3~13% under different scenarios.

Application of resource-saving information technologies during preventive medical examinations in Orsha central district hospital

The use of resource-saving information technologies during preventive medical examinations leads to an increase in qualitative and quantitative indicators and a reduction in resource costs. The results of the introduction and use of resource-saving information technologies are described in this article. The purpose of the work: to assess the organizational and economic efficiency of the use of resource-saving information technologies during preventive medical examinations in the Orsha Central District Hospital. Results. The issues of national informatization of health care and preventive medicine, prerequisites, experience in the creation and use of resource-saving information technologies in a medical organization were studied; the material, technical, personnel costs and planned and economic indicators of the activities of a medical organization during preventive medical examinations were studied; the cost of time resources was studied during preventive medical examinations by medical personnel using resource-saving information technologies; the satisfaction of patients with preventive medical examinations using resource-saving information technologies was studied; the effectiveness of the use of resource-saving information technology during preventive medical examinations was evaluated. Conclusion. The results obtained indicate that all participants in the procedure for conducting preliminary and periodic medical examinations (medical organization, medical personnel, patients) are in dire need of saving resources, namely resource saving of time and personnel.

HetSev: Exploiting Heterogeneity-Aware Autoscaling and Resource-Efficient Scheduling for Cost-Effective Machine-Learning Model Serving

To accelerate the inference of machine-learning (ML) model serving, clusters of machines require the use of expensive hardware accelerators (e.g., GPUs) to reduce execution time. Advanced inference serving systems are needed to satisfy latency service-level objectives (SLOs) in a cost-effective manner. Novel autoscaling mechanisms that greedily minimize the number of service instances while ensuring SLO compliance are helpful. However, we find that it is not adequate to guarantee cost effectiveness across heterogeneous GPU hardware, and this does not maximize resource utilization. In this paper, we propose HetSev to address these challenges by incorporating heterogeneity-aware autoscaling and resource-efficient scheduling to achieve cost effectiveness. We develop an autoscaling mechanism which accounts for SLO compliance and GPU heterogeneity, thus provisioning the appropriate type and number of instances to guarantee cost effectiveness. We leverage multi-tenant inference to improve GPU resource utilization, while alleviating inter-tenant interference by avoiding the co-location of identical ML instances on the same GPU during placement decisions. HetSev is integrated into Kubernetes and deployed onto a heterogeneous GPU cluster. We evaluated the performance of HetSev using several representative ML models. Compared with default Kubernetes, HetSev reduces resource cost by up to 2.15× while meeting SLO requirements.

The distributed programming and data processing methods for computational task optimization in cloud environments

Cloud computing environments have become an essential component for executing computational tasks of various scales and complexities in the modern information-driven world. The use of distributed programming and efficient data processing in cloud environments is a topical and strategically important issue. The analysis of recent research and publications in the field of cloud computing indicates a continuous development and expansion of capabilities in these environments. Special emphasis is placed on container-based distributed programming and orchestration, particularly using Kubernetes. These technologies enable scalability and automation of the computational processes in cloud environments. The aim of this article is to investigate methods for optimizing computational tasks in cloud environments. The research encompasses the analysis of distributed programming, real-time data processing, and the integration of state-of-the-art technologies to achieve optimal efficiency. One of the key aspects of the research is the study of distributed programming. This approach allows the distribution of computational tasks among various cloud nodes, leading to improved performance and fault tolerance. Real-time data processing plays a pivotal role in addressing tasks where time is a critical factor. The integration of cutting-edge technologies, such as artificial intelligence and machine learning, aids in achieving optimal results. The conclusions of this study underscore the importance of combining distributed programming and efficient data processing to optimize computational tasks in cloud environments. This approach enables increased productivity and reduced resource costs. Future research in this direction envisions further development of methods and tools to achieve the best possible outcomes in cloud computing. Exploring new opportunities in cloud computing environments will ensure a continuous enhancement of the quality of computational services and contribute to the advancement of the modern information society.

FEATURE STRUCTURE OF A COTTON LEAF AS A MECHANISM OF PROTECTION AGAINST INSECT PESTS

According to the International Cotton Advisory Committee, Uzbekistan accounts for about 4% of all cotton produced in the world and 10%of world exports. At the same time, cotton makes up 45% of all Uzbek exports. But besides the purely commercial meaning, "white gold"has always had a huge cultural significance for Uzbekistan. Cotton is a strategic agricultural crop, as after processing it is used in thechemical, light industry and many other industries. Currently, legal documents are being adopted at the national level to solve the problemof crop rotation in the cotton industry and the necessary mechanisms are being developed for their implementation. In addition, in the systemof insect control in cotton fields, the issue of resource conservation is very relevant, since by reducing harmful insects it is necessary toachieve a reduction in resource costs in terms of increasing (or not reducing) the quantity and quality of the cotton crop. Protection of plantsfrom pests by various methods is widely used in world agricultural production. At the same time, according to the FAO, the losses ofagricultural products in the world still remain quite high and reach 30% of the volume of the world crop. To solve the problem of crop safety,extensive and systematic scientific and practical work is being carried out to improve and develop methods and means of plant protection inagriculture to prevent harm caused to plants by pests, diseases and weeds. In this regard, the purpose of this work was to establish thecorrelating structural features of the leaves of wild representatives of the genus Gossypium L with economically valuable traits. 19representatives were studied, among which were 7 parental species and 12 hybrids of their first generation. As a result, signs were establishedthat correlate with drought resistance and tolerance to insect pests.

Reconfigurable Intelligent Surface-Aided Cooperative NOMA with p-CSI Fading Channel toward 6G-Based IoT System.

Addressing the challenges of internet-based 5G technology, namely increasing density through micro-cell systems, frequency spectrum, and reducing resource costs, is needed to meet the use of IoT-based 6G technology with the goal of high-speed, high-capacity, and low-latency communication. In this research, we considered the coverage performance and ergodic capacity of the Reconfigurable Intelligent Surface (RIS)-aided cooperative nonorthogonal multiple-access network (NOMA) of an IoT system. This enables the upgrading of 5G- toward 6G-technology-based IoT systems. We developed a closest-form formula of near and far user coverage probabilities as a function of perfect channel statistical information (p-CSI) using only a single-input single-output (SISO) system with a finite number of RIS elements under the Nakagami-m fading channel. We also define ergodic capacity as a simple upper limit by simplifying the use of symbolic functions and it could be used for a sustained period. The simulation findings suggest that RIS-assisted NOMA has a reduced risk of outage than standard NOMA. All of the derived closed-form formulas agree with Monte Carlo simulations, indicating that the distant user’s coverage probability outperforms the nearby user. The bigger the number of RIS parts, however, the greater the chance of coverage. They also disclose the scaling law of the number of phase shifts at the RIS-aided NOMA based on the asymptotic analysis and the upper bound on channel capacity. In both arbitrary and optimum phase shifts, the distant user’s ergodic capacity outperforms the near user.

Resource-efficient high-dimensional subspace teleportation with a quantum autoencoder.

Quantum autoencoders serve as efficient means for quantum data compression. Here, we propose and demonstrate their use to reduce resource costs for quantum teleportation of subspaces in high-dimensional systems. We use a quantum autoencoder in a compress-teleport-decompress manner and report the first demonstration with qutrits using an integrated photonic platform for future scalability. The key strategy is to compress the dimensionality of input states by erasing redundant information and recover the initial states after chip-to-chip teleportation. Unsupervised machine learning is applied to train the on-chip autoencoder, enabling the compression and teleportation of any state from a high-dimensional subspace. Unknown states are decompressed at a high fidelity (~0.971), obtaining a total teleportation fidelity of ~0.894. Subspace encodings hold great potential as they support enhanced noise robustness and increased coherence. Laying the groundwork for machine learning techniques in quantum systems, our scheme opens previously unidentified paths toward high-dimensional quantum computing and networking.