The rapid evolution of Large Language Models (LLM) and subsequent Agentic AI technologies requires systematic architectural guidance for building sophisticated, production-grade systems. This paper presents an approach for architecting such systems using design patterns derived from enterprise distributed systems standards, formal methods, and industry practice. We classify these patterns into three tiers: LLM Agents (task-specific automation), Agentic AI (adaptive goal-seekers), and Agentic Communities (organizational frameworks where AI agents and human participants coordinate through formal roles, protocols, and governance structures). We focus on Agentic Communities - coordination frameworks encompassing LLM Agents, Agentic AI entities, and humans - most relevant for enterprise and industrial applications. Drawing on established coordination principles from distributed systems, we ground these patterns in a formal framework that specifies collaboration agreements where AI agents and humans fill roles within governed ecosystems. This approach provides both practical guidance and formal verification capabilities, enabling expression of organizational, legal, and ethical rules through accountability mechanisms that ensure operational and verifiable governance of inter-agent communication, negotiation, and intent modeling. We validate this framework through a clinical trial matching case study. Our goal is to provide actionable guidance to practitioners while maintaining the formal rigor essential for enterprise deployment in dynamic, multi-agent ecosystems.
ObjectiveRural landscape in the Jiangnan Canal Basin is characterized by a synergistic symbiosis of water basin habitats, canal water conservancy and transportation, and rural settlement construction. Based on the perspective of the correlation and coupling of water network − canal channel − rural landscape, this research aims to explore the types of rural landscape characters and spatial coupling patterns in the Jiangnan Canal Basin, so as to ensure that rural landscape continues to play a significant role in the protection of regional historical heritage and the resilient development of urban and rural areas within the basin.MethodsThe research selects the Tiaoxi River − Canal Basin in Zhejiang Province, which is closely connected to the Jiangnan Canal, as the research object. At the river − canal basin scale, the multi-level spatial superposition technique is utilized to identify associative character zones, and a coupling coordination model is used to identify the correlation patterns of water network， canal channel and rural landscape. At the rural community scale, through spatial maximum expectation (EM) clustering, the types of rural landscape characters are classified and spatially mapped through simulation algorithms.ResultsIn terms of the spatial correlation of water network − canal channel − rural landscape within the Tiaoxi River − Canal Basin, the research identifies three major landscape character zones including the West Tiaoxi River − Ditang Canal Basin, East Tiaoxi River − Jiangnan Canal Basin, and Hangzhoutang River − Hangzhou − Jiaxing Canal Basin. In terms of the coupling pattern of water network – canal channel – rural landscape in the Tiaoxi River Canal Basin, the coupling and coordination degree between the East and West Tiaoxi River Basin and the buffer zone along the Jiangnan Canal is generally high, and the that in the west area dominated by natural water system is stronger than that in the east area dominated by canal channel. In the main river basin of the East Tiaoxi River – Jiangnan Canal, there are coupling cold spots at the gathering places of canal towns such as Jiaxing and Hangzhou, and there are significant disturbances in the coupling of the landscape system. In terms of typical rural landscape types in the Tiaoxi River – Canal Basin, six types of rural landscape paradigms are identified, including the canal hydraulic engineering, regional agricultural production, township community linkage, canal trade operation, water conservancy culture collection and water network habitat maintenance types. And significant region − basin differences are explored from the perspectives of spatial proportion, layout pattern, and coupling relationship. At the level of the spatial correlation mechanism of typical rural landscapes, the regional agricultural production and township community linkage types of landscape occupy a relatively large proportion, accounting for 24.49% and 21.89% respectively, while the water conservation culture collection and water network habitat maintenance types of landscape occupy a relatively small proportion, accounting for 12.88% and 10.60% respectively. At the level of spatial transformation from the overall scale of basin as a whole to the individual scale of rural settlements, the canal hydraulic engineering and township community linkage types of landscape are mainly distributed in the East Tiaoxi River and the Hangzhoutang River water network basin, the regional agricultural production type of landscape is mainly distributed in the West Tiaoxi River − Ditang Canal Basin, and the water network habitat maintenance type of landscape is mainly distributed in the connection area between the main veins of the East and West Tiaoxi River systems and the tributaries of the Hangzhou − Jiaxing Canal. The canal trade operation and water conservation culture collection types of landscape are mainly distributed in the East Tiaoxi River − Jiangnan Canal Basin and the Hangzhoutang River − Hangzhou − Jiaxing Canal Basin. From the perspective of coupling mechanism, the spatial coupling degrees of the regional agricultural production and township community linkage types of landscape are relatively high in the West Tiaoxi River − Ditang Canal Basin, with mean values of 0.731 and 0.775, respectively; while the mean value of the canal hydraulic engineering type is only 0.596, relatively low.ConclusionThe research delves into the underlying mechanisms that interlink the water network, canal channel and rural landscape within the human settlement system of the Jiangnan Canal Basin. The research clarifies the stratified compositional elements, the superposition of character zoning, the predominant correlation types, and the coupling mechanisms of rural landscape. The research delineates landscape character zones at the canal basin scale and identifies the dominant types of rural landscapes at the scale of rural settlements. Building on these findings, the research achieves a basin-scale spatial inversion of rural landscape types. Furthermore, the research establishes a research paradigm for examining the characters, spatial distribution and mechanisms of rural landscapes within the canal basin. The research offers resilient planning insights that integrate trans-basin zone governance, multi-type system linkage, and multi-level network construction, providing a comprehensive approach to the management and conservation of canal heritage and rural landscapes. Future research will construct a spatio-temporal dynamic dataset of canal heritage landscapes, which aims to enhance the fine-grained character extraction and type classification accuracy of machine learning algorithm models for rural landscapes. Additionally, the universal applicability of the stratified correlation perspective in the field of rural landscape research within the context of the Grand Canal Basin will also be explored.
Aesthetics of cities. City planning and beautifying, Architectural drawing and design
This paper introduces a novel disk array architecture, designated RAID-0e (Resilient Striping Array), designed to superimpose a low-overhead fault tolerance layer upon traditional RAID 0 (striping). By employing a logically and physically separate parity domain to protect a primary data domain, RAID-0e mitigates the risk of array-wide data loss from common, non-catastrophic media failures, such as isolated bad blocks, transient read errors, or sector-level corruption. The architecture is engineered to preserve the intrinsic read performance advantages of RAID 0 while significantly enhancing data availability and operational resilience. This document provides a comprehensive exposition of the architectural principles, operational workflows, performance characteristics, failure mode analysis, and security considerations of RAID-0e. It is presented as an experimental yet pragmatic solution for environments seeking a new equilibrium between I/O performance, storage cost, and data resilience, particularly where full drive failure is a secondary concern to media degradation.
Cloud architecture design is a complex process requiring both technical expertise and architectural knowledge to develop solutions from frequently ambiguous requirements. We present CloudArchitectBuddy, a system-driven cloud architecture design support application with two key mechanisms: (1) structured state management that enhances design understanding through explicit representation of requirements and architectural decisions, and (2) guided decision assistance that facilitates design progress through proactive verification and requirement refinement. Our study with 16 industry practitioners showed that while our approach achieved comparable design quality to a chat interface, participants rated our system higher for usability and appreciated its ability to help understand architectural relationships and identify missing requirements. However, participants also expressed a need for user-initiated interactions where they could freely provide design instructions and engage in detailed discussions with LLMs. These results suggest that integrating a chat interface into our structured and guided workflow approach would create a more practical solution, balancing systematic design support with conversational flexibility for comprehensive cloud architecture development.
Portugal's early seaside holiday camps emerged prior to the Estado Novo era, serving as summer destinations for underprivileged children even before the implementation of national educational programs. Educational and charitable associations, supported by private philanthropy and the working class, played a pivotal role in the development of these facilities. Newspapers and cooperatives like Voz do Operário spearheaded social initiatives to address the challenges of the late nineteenth century, targeting workers and their families. These initiatives involved hundreds of children each summer, primarily between June and October, for the care of childhood tuberculosis and the recreation of young guests. These projects found space in existing buildings or gave rise to new buildings along a mostly pristine coastline, often at a considerable distance from urban centers. A significant example is the O Século seaside children's holiday camp, funded by private and public funds, operational since 1927 in S. Pedro do Estoril, between Lisbon and the renowned Cascais. The project went through at least three construction phases, influencing city public life and benefiting from the support of Lisbon’s famous Feira Popular, in operation until a few years ago. Archival documents, historical photographs, and blueprints testify to the evolution of these facilities over time. Transformations between 1944 and 1945 highlight the importance of this camp in the collectivity, its strategies of funding, and its needs, staff, and the children it hosted. Over the years, many of these facilities disappeared due to the tourism expansion in the area, unlike the case of O Século, which still partly continues programs dedicated to childhood.
Architectural drawing and design, Aesthetics of cities. City planning and beautifying
Jakob Svennevik Notland, Jinguye Li, Mariusz Nowostawski
et al.
Cross-chain bridges are solutions that enable interoperability between heterogeneous blockchains. In contrast to the underlying blockchains, the bridges often provide inferior security guarantees and have been targets of hacks causing damage in the range of 1.5 to 2 billion USD in 2022. The current state of bridge architectures is that they are ambiguous, and there is next to no notion of how different architectures and their components are related to different vulnerabilities. Throughout this study, we have analysed 60 different bridges and 34 bridge exploits in the last three years (2021-2023). Our analyses identified 13 architectural components of the bridges. We linked the components to eight types of vulnerabilities, also called design flaws. We identified prevention measures and proposed 11 impact reduction measures based on the existing and possible countermeasures to address the imminent exploits of the design flaws. The results are meant to be used as guidelines for designing and implementing secure cross-chain bridge architectures, preventing design flaws, and mitigating the negative impacts of exploits.
Vittorio Cortellessa, J. Andres Diaz-Pace, Daniele Di Pompeo
et al.
Several approaches have recently used automated techniques to generate architecture design alternatives by means of optimization techniques. These approaches aim at improving an initial architecture with respect to quality aspects, such as performance, reliability, or maintainability. In this context, each optimization experiment usually produces a different set of architecture alternatives that is characterized by specific settings. As a consequence, the designer is left with the task of comparing such sets to identify the settings that lead to better solution sets for the problem. To assess the quality of solution sets, multi-objective optimization commonly relies on quality indicators. Among these, the quality indicator for the maximum spread estimates the diversity of the generated alternatives, providing a measure of how much of the solution space has been explored. However, the maximum spread indicator is computed only on the objective space and does not consider architectural information (e.g., components structure, design decisions) from the architectural space. In this paper, we propose a quality indicator for the spread that assesses the diversity of alternatives by taking into account architectural features. To compute the spread, we rely on a notion of distance between alternatives according to the way they were generated during the optimization. We demonstrate how our architectural quality indicator can be applied to a dataset from the literature.
Drawing (a multiset of) coloured balls from an urn is one of the most basic models in discrete probability theory. Three modes of drawing are commonly distinguished: multinomial (draw-replace), hypergeometric (draw-delete), and Polya (draw-add). These drawing operations are represented as maps from urns to distributions over multisets of draws. The set of urns is a metric space via the Kantorovich distance. The set of distributions over draws is also a metric space, using Kantorovich-over-Kantorovich. It is shown that these three draw operations are all isometries, that is, they exactly preserve the Kantorovich distances. Further, drawing is studied in the limit, both for large urns and for large draws. First it is shown that, as the urn size increases, the Kantorovich distances go to zero between hypergeometric and multinomial draws, and also between Pólya and multinomial draws. Second, it is shown that, as the drawsize increases, the Kantorovich distance goes to zero (in probability) between an urn and (normalised) multinomial draws from the urn. These results are known, but here, they are formulated in a novel metric manner as limits of Kantorovich distances. We call these two limit results the law of large urns and the law of large draws.
Alessandra Battisti, Marco Antonini, Angela Calvano
et al.
Clean Energy for All Europeans, Green Deal and Fit for 55 at European level and the Piano Nazionale Integrato per l’Energia e il Clima have identified essential decarbonisation goals that are achievable by broadening the audience of actors involved and the ability to innovate in terms of the evolution of products/services and production processes. In this sense, decentralised energy production technologies together with Renewable Energy Communities (RECs) and bottom-up initiatives play a strategic role in the establishment of local energy systems. The objective of the paper is to illustrate these dynamisms within the territorial system, and the characteristics of local initiatives and RECs as means of energy transition and economic development.
Aesthetics of cities. City planning and beautifying, Architectural drawing and design
This study presents a case study that aims to select the ideal internal wall panel option causing less environmental impact for the Sumerbank Kayseri Textile Factory restoration process, which is now used as part of Abdullah Gul University’s main campus. Since the university has an environmental agenda, examining the environmental impacts of the materials used for the ongoing restoration process has the potential to contribute to these goals. For this purpose, the three most used interior wall panels in the Turkish building material industry, gypsum, reinforced gypsum, and cement-based panels, were selected within the scope of the case study. The life cycle assessment (LCA) method was used to compare these options, and analyses were conducted using SimaPro software. The data required for life cycle impact assessment (LCIA) were obtained based on market analyses and also from the EcoInvent Life Cycle Inventory Database. At the end of the study, damage assessment, weighting, and midpoint and endpoint data of the characterization results provided by the ReCiPe method were compared and interpreted. According to the overall results obtained for the described case conditions, reinforced gypsum panel causes the most adverse environmental impacts, followed by cement and gypsum panels, respectively.
Sparse training is one of the promising techniques to reduce the computational cost of DNNs while retaining high accuracy. In particular, N:M fine-grained structured sparsity, where only N out of consecutive M elements can be nonzero, has attracted attention due to its hardware-friendly pattern and capability of achieving a high sparse ratio. However, the potential to accelerate N:M sparse DNN training has not been fully exploited, and there is a lack of efficient hardware supporting N:M sparse training. To tackle these challenges, this paper presents a computation-efficient training scheme for N:M sparse DNNs using algorithm, architecture, and dataflow co-design. At the algorithm level, a bidirectional weight pruning method, dubbed BDWP, is proposed to leverage the N:M sparsity of weights during both forward and backward passes of DNN training, which can significantly reduce the computational cost while maintaining model accuracy. At the architecture level, a sparse accelerator for DNN training, namely SAT, is developed to neatly support both the regular dense operations and the computation-efficient N:M sparse operations. At the dataflow level, multiple optimization methods ranging from interleave mapping, pre-generation of N:M sparse weights, and offline scheduling, are proposed to boost the computational efficiency of SAT. Finally, the effectiveness of our training scheme is evaluated on a Xilinx VCU1525 FPGA card using various DNN models and datasets. Experimental results show the SAT accelerator with the BDWP sparse training method under 2:8 sparse ratio achieves an average speedup of 1.75x over that with the dense training, accompanied by a negligible accuracy loss of 0.56% on average. Furthermore, our proposed training scheme significantly improves the training throughput by 2.97~25.22x and the energy efficiency by 1.36~3.58x over prior FPGA-based accelerators.
Deep learning models have proven to be successful in a wide range of machine learning tasks. Yet, they are often highly sensitive to perturbations on the input data which can lead to incorrect decisions with high confidence, hampering their deployment for practical use-cases. Thus, finding architectures that are (more) robust against perturbations has received much attention in recent years. Just like the search for well-performing architectures in terms of clean accuracy, this usually involves a tedious trial-and-error process with one additional challenge: the evaluation of a network's robustness is significantly more expensive than its evaluation for clean accuracy. Thus, the aim of this paper is to facilitate better streamlined research on architectural design choices with respect to their impact on robustness as well as, for example, the evaluation of surrogate measures for robustness. We therefore borrow one of the most commonly considered search spaces for neural architecture search for image classification, NAS-Bench-201, which contains a manageable size of 6466 non-isomorphic network designs. We evaluate all these networks on a range of common adversarial attacks and corruption types and introduce a database on neural architecture design and robustness evaluations. We further present three exemplary use cases of this dataset, in which we (i) benchmark robustness measurements based on Jacobian and Hessian matrices for their robustness predictability, (ii) perform neural architecture search on robust accuracies, and (iii) provide an initial analysis of how architectural design choices affect robustness. We find that carefully crafting the topology of a network can have substantial impact on its robustness, where networks with the same parameter count range in mean adversarial robust accuracy from 20%-41%. Code and data is available at http://robustness.vision/.
The first ACM/IEEE TinyML Design Contest (TDC) held at the 41st International Conference on Computer-Aided Design (ICCAD) in 2022 is a challenging, multi-month, research and development competition. TDC'22 focuses on real-world medical problems that require the innovation and implementation of artificial intelligence/machine learning (AI/ML) algorithms on implantable devices. The challenge problem of TDC'22 is to develop a novel AI/ML-based real-time detection algorithm for life-threatening ventricular arrhythmia over low-power microcontrollers utilized in Implantable Cardioverter-Defibrillators (ICDs). The dataset contains more than 38,000 5-second intracardiac electrograms (IEGMs) segments over 8 different types of rhythm from 90 subjects. The dedicated hardware platform is NUCLEO-L432KC manufactured by STMicroelectronics. TDC'22, which is open to multi-person teams world-wide, attracted more than 150 teams from over 50 organizations. This paper first presents the medical problem, dataset, and evaluation procedure in detail. It further demonstrates and discusses the designs developed by the leading teams as well as representative results. This paper concludes with the direction of improvement for the future TinyML design for health monitoring applications.
Our research aims to unify existing works' diverging opinions on how architectural components affect the adversarial robustness of CNNs. To accomplish our goal, we synthesize a suite of three generalizable robust architectural design principles: (a) optimal range for depth and width configurations, (b) preferring convolutional over patchify stem stage, and (c) robust residual block design through adopting squeeze and excitation blocks and non-parametric smooth activation functions. Through extensive experiments across a wide spectrum of dataset scales, adversarial training methods, model parameters, and network design spaces, our principles consistently and markedly improve AutoAttack accuracy: 1-3 percentage points (pp) on CIFAR-10 and CIFAR-100, and 4-9 pp on ImageNet. The code is publicly available at https://github.com/poloclub/robust-principles.
Software engineers share their architectural knowledge (AK) in different places on the Web. Recent studies show that architectural blogs contain the most relevant AK, which can help software engineers to make design steps. Nevertheless, we know little about blogs, and specifically architectural blogs, where software engineers share their AK. In this paper, we conduct an exploratory study on architectural blogs to explore their types, topics, and their AK. Moreover, we determine the relevance of architectural blogs to make design steps. Our results support researchers and practitioners to find and re-use AK from blogs.