Abstract: This article provides general expressions for the phase velocity and the Doppler shift of the electromagnetic fields radiated from a uniformly moving Hertzian dipole measured by a uniformly moving observer. The results show that the phase velocity of the electromagnetic wave is always equal to when measured exactly in the direction pointing to the birthplace of the field. The expression for the Doppler effect is of the same form of the Newtonian type classical formula, which implies that it might be not proper to consider that the classical formula for the Doppler shift is the low speed approximation of the conventional relativistic formula.

Abstract: Industry 4.0 technologies offer substantial opportunities for sustainable business transformation, yet organisations consistently struggle to translate technological in-vestments into successful project outcomes. This study investigates the inner workings the "black box” of Industry 4.0 project implementation by examining how project management practices, team competencies, and decision-making processes interact. Using a mixed-methods case study of a leading industrial automation company, in-cluding a survey of project team members (n=50) and interviews with project managers (n=5), The identification of recursive feedback loop: competency gaps directly cause decision failures, and poor decision processes subsequently widen those competency gaps. Conversely, structured decision reviews and transparent communication trans-form routine choices into competency-building opportunities. An Integrated Imple-mentation Model (IIM) was proposed that explains these dynamics and demonstrates that sustainability outcomes, like resource efficiency, waste reduction, and circular economy practices emerge naturally when organisations manage processes, people, and decisions together. For practitioners, the core message is that every Industry 4.0 project should be treated as an opportunity to build long-term organisational learning capacity, not merely as a technology installation. This study provides both a theoretical framework for understanding implementation dynamics and actionable guidance for sustainable digital transformation.

Abstract: This study examined the mediating role of negative stress in the relationship between Psychological Capital (PsyCap) a higher-order construct comprising hope, self-efficacy, resilience, and optimism and psychological distress indicators among Ecuadorian uni-versity students. A cross-sectional survey was conducted with 1,732 students (55% women; M = 20.44, SD = 2.29), using validated self-report measures. Structural equa-tion modeling showed a good model fit (CFI = 0.947; TLI = 0.942; RMSEA = 0.055; SRMR = 0.040) Results indicated that PsyCap was negatively associated with negative stress (β = −0.261), which in turn showed strong positive effects on anxiety–depression symptoms (β = 0.782) and psychological inflexibility (β = 0.781). Direct effects of PsyCap on both outcomes were significant but comparatively small (β = −0.115 and β = −0.086, respec-tively), whereas indirect effects through stress were substantial and significant (β = −0.204), supporting a partial mediation model. The model explained 67.2% of the vari-ance in anxiety–depression and 65.2% in psychological inflexibility. These findings suggest that PsyCap operates primarily as a protective factor through its capacity to reduce negative stress, which subsequently influences downstream psy-chological outcomes. The results highlight the importance of stress-focused mecha-nisms in understanding how positive psychological resources impact mental health. From an applied perspective, the findings underscore the relevance of implementing strengths-based interventions in higher education that enhance PsyCap components while simultaneously targeting stress reduction. Such inter-ventions may contribute to decreasing psychological distress and improving students’ adaptive functioning and well-being. This study provides robust evidence from the Latin American context, advancing the understanding of transdiagnostic mechanisms linking positive resources and mental health in university populations.

Abstract: Poly(l-lactic acid) or poly(l-lactide) (PLLA) is an optically active polymer derived from renewable sources and fully biodegradable. It is known that PLLA assumes a left-handed helix in the solid state and also in solution it still keeps a certain degree of helical structure. Here we examine the Optical Rotatory Dispersion (ORD) behavior of two grades of PLLA (medium molecular weight and hexadecyl-terminated or a high molecular weight for 3D printing) in 13 different solvents and through the Moffitt-Yang equation of the ORD data. Furthermore, the ORD data of PLLA in additional 6 solvents were taken from literature and analyzed with the Moffitt-Yang approach. The results suggest that also in solution PLLA maintain the left-handed helix and the most structurizing and helicogenic solvents for PLLA are ethyl acetate, acetonitrile, and certain chlorinated solvents. The equilibrium association constant (K) and other thermodynamic parameters (ΔG°, ΔH° and ΔS°) between PLLA and polyphenylacetylene (PPA another helical polymer in the solid state and in solution) were determined in trichloromethane, dichloromethane and tetrahydrofuran. The K values found suggest a strong helix-helix interaction between the two polymers. The ORD analysis of the PLLA-PPA solutions show evidences of the extrinsic Cotton effect and confirming the chiral helicity induction between the two polymers with 1:1 complex formation.

Abstract: High consumption of ultra-processed foods (UPFs) contributes to the growing burden of non-communicable disease, yet many consumers struggle to recognise and interpret levels of processing. Digital tools using artificial intelligence (AI) offer potential to support nutrition literacy and UPF awareness. This study explored user perceptions, usability and cultural relevance of a Human Interference Scoring System (HISS)-based mobile application designed to classify foods and support reflection on food quality and dietary choices. A qualitative study was conducted in New Zealand, where participants used the HISS app for three days followed by semi-structured interviews. Thirty-one participants were recruited via social media and word of mouth, including adolescents (n=13), tertiary students (n=9), and Māori and Pacific health coaches (n=9). Transcripts were analysed using inductive thematic analysis. Three evaluative categories were identified: positive user experiences (intuitive interface, perceived AI accuracy, enhanced nutrition literacy, visual feedback, inclusivity of cultural foods); challenges (technical issues, database gaps, limited depth for advanced users); and suggested improvements (expanded food database, enhanced logging, culturally tailored education, optional advanced features). Participants reported increased awareness of UPF intake and reflection on food choices. The HISS app was perceived as usable, acceptable and relevant across diverse user groups, particularly for those with lower nutrition literacy. Addressing technical limitations and expanding functionality may enhance engagement and applicability. AI-enabled, culturally responsive food classification tools such as HISS show promise as scalable health promotion approaches to support UPF awareness and dietary reflection in community and clinical settings.

Abstract: Background: Retained surgical items (RSIs) remain preventable never events associated with failures in counting processes, communication, documentation, and perioperative safety culture. Although patient safety in the operating room has been widely studied in healthcare teams, evidence remains limited regarding how these competencies are developed during undergraduate surgical instrumentation training, particularly in Latin American settings. Objective. To assess knowledge, safe practices, and perceived safety culture related to RSI prevention among surgical instrumentation students in clinical training at a university in Barranquilla, Colombia. Methods. A quantitative cross-sectional study was conducted among students in advanced semesters of a Surgical Instrumentation program. Data were collected through a structured, self-administered questionnaire that included sociodemographic characteristics, knowledge of RSI prevention, safe perioperative practices, perceived safety culture, and need for further training. The instrument showed good internal consistency (Cronbach’s alpha = 0.88) in the 48 completed questionnaires analyzed. Descriptive statistics were calculated for all variables. Because of sample size and ordinal outcomes, Mann-Whitney U, Kruskal-Wallis H, and Spearman’s rho were used. Results. A total of 48 complete questionnaires were analyzed. Mean age was 21.10 years (SD 2.25), and 75.0% of participants were women. Global knowledge scores were high (mean 8.60/9; SD 0.92), with ceiling effects across several items. Safe practices (mean 4.58/5; SD 0.63) and safety culture (mean 4.67/5; SD 0.49) were also high overall. However, lower-performing items were institutional count documentation, verification during staff handovers, and requesting a formal count pause. Knowledge scores were significantly higher among students reporting prior training in count/recording procedures (p=0.018). The strongest association was a moderate positive correlation between safety culture and safe practices (rho=0.520; p<0.001). Conclusions. Surgical instrumentation students showed strong theoretical knowledge and favorable self-reported preventive practices regarding RSI prevention. Nonetheless, important gaps persisted in handover verification, institutional documentation, and count-pause activation. The positive association between safety culture and safe practices suggests that undergraduate perioperative education should integrate technical counting skills with structured communication, supervision, teamwork, and speaking-up behaviors.

Abstract: In this work, a novel entropy-weighted fuzzy c-means variation is proposed. This varia-tion introduces a semantic level of partitioning of features into groups. This approach en-ables the provision of optimal semantic meaning to the clusters, thereby capturing the in-trinsic structure of the features, which are naturally grouped into homogeneous semantic sets. Additionally, it is computationally more efficient than other cluster-specific weighted fuzzy clustering algorithms, due to the independence of the weights from the clusters. The efficacy of the method was assessed by evaluating census data from 16 Italian cities, with the objective of partitioning urban settlements based on characteristics of residential buildings, including construction technique, period, number of floors, and state of con-servation. The findings suggest that the proposed algorithm effectively captures the se-mantic meaning of clusters. A comparative analysis of the two algorithms reveals that the new algorithm offers similar outcomes to the traditional EWFCM algorithm while signifi-cantly enhancing computational speed.

Abstract: The proliferation of Unmanned Aerial Vehicles (UAVs) in various applications has created a pressing need for robust and efficient communication systems. Fifth generation (5G) networks, with their high bandwidth and low latency, are poised to support the massive connectivity requirements of UAVs. However, the high mobility of drones presents significant challenges for handover management, leading to frequent service interruptions and degraded performance. This paper proposes a novel, first-of-its-kind framework that integrates multi-UAV trajectory prediction with proactive handover optimization in 5G networks. Our approach utilizes a Long Short-Term Memory (LSTM)-based Recurrent Neural Network (RNN) to predict the future flight path of each UAV. The predicted trajectories are then fed into a Deep Reinforcement Learning (DRL) agent, which makes optimal handover decisions to ensure seamless connectivity and high Quality of Service (QoS). Unlike existing solutions that primarily rely on simulated data, our framework is validated using a real-world drone trajectory dataset. The experimental results demonstrate that our proposed method significantly outperforms traditional and existing machine learning-based handover schemes in terms of handover success rate, average Signal-to-Interference-plus-Noise Ratio (SINR), and handover delay. The proposed framework paves the way for more reliable and efficient drone operations in 5G and beyond networks.

Abstract: Settlement discrepancies in multi-party electricity trading are difficult to localize because final outcomes are produced by multi-stage pipelines that combine heterogeneous data, rule versions, parameters, and execution contexts across organizational boundaries. In such settings, numerical reconciliation is not enough: investigators must identify where divergence entered the pipeline and support that judgment with evidence that can be checked independently. We formulate discrepancy localization as an auditable inference problem and introduce RootTrace, a trusted path-traceability method for this setting. Settlement processing is represented as an evidence graph over versioned artifacts and explicit events; RootTrace uses backward tracing and version-difference tracing to derive a suspect set of stages and/or artifact versions. The same pipeline exports a verifiable evidence bundle that preserves the trace used in localization. To support accountability under an explicit threat model, RootTrace includes trusted recording and verification procedures for tamper-evident capture and minimal-disclosure bundle export. On a semi-synthetic benchmark with ten independent replications, RootTrace achieves mean Top-1/Top-5/MRR of 0.47/0.83/0.60, compared with Top-1 0.17 and 0.04 for representative rule-based and stage-order baselines; exported bundles verify cleanly with full detection of the configured tamper classes (mean verification latency below 5ms on our grid); and an eight-hour unattended stress run completes over 1.3×10⁶ iterations without runtime failure.

Abstract: Urban eco-efficiency (UEE) is fundamental to achieving China's dual-carbon goals. However, literature has overlooked green space carbon sequestration, and linear models fail to capture complex nonlinear relationships. This study integrates green space carbon sinks into the evaluation framework, employing the global super-efficiency EBM model to measure the UEE of 108 cities in the Yangtze River Economic Belt (YREB) from 2012 to 2023. It combines XGBoost-SHAP with Geographically and Temporally Weighted Regression (GTWR) to examine UEE's spatiotemporal dynamics and driving mechanisms. The findings reveal that: (1) UEE in the YREB increased from 1.0760 in 2012 to 1.0990 in 2023, while spatial polarization became more pronounced. (2) Core driving factors exhibited significant nonlinear threshold and interactive effects. Specifically, fiscal decentralization's environmental dividend is contingent on active government intervention to circumvent localized "race to the bottom" behaviors. Furthermore, population density transitions from yielding scale dividends to inducing "crowding effects" beyond optimal capacities—a degradation advanced financial systems appear unable to mitigate. (3) A spatiotemporal misalignment was observed: fiscal decentralization unleashed green institutional dividends downstream (coefficients up to 0.0682), but caused a race to the bottom in middle and upper reaches (extending to -0.6548); excessive population agglomeration in megacities induced a crowding effect eroding early pollution control dividends. This study supports abandoning one-size-fits-all approaches and developing precise, spatiotemporally differentiated low-carbon policies.

Abstract: Cellular stress signaling conveys vital messages to the cell’s machinery to respond dynamically to internal and external environmental cues. One prevailing hypothesis for such signaling is to bring about crucial downstream functional changes in the cell’s ability to withstand intra- and extra-cellular stress, as exemplified by the NRF2-ARE pathway. Reactive electrophilic metabolites (REMs) are often generated from membrane lipids or respiratory metabolites in inflammatory and stress signaling contexts. REMs harbour an innate ability to irreversibly bind protein-cysteines, and form DNA and RNA adducts. Our work has led us to propose a new hypothesis regarding the role of locale-specific REM build-up in stress signalling; whereby they can label both the majority stable protein residents in a given locale, as well as minority guests transiently existing within a subcellular compartment, at a kinetically significant rate to trigger a “gain of function” signalling cascade. Much like the NRF2 signalling pathway, such downstream signalling messages could assist the cell’s ability to survive against microenvironment-specific stress and adapt on-demand. As REMs accumulate in several disease-specific cells, especially age-related disorders such as cancer and neurodegeneration, understanding the functional signal propagation mechanisms shaped by specific REMs engaging with specific biomolecular targets will prove vital for future therapeutic interventions with enhanced precision and context.

Abstract: The fall of droplets of an aqueous NaCl solution in a vertical channel, filled with heated dry air, is studied. Water from the droplets evaporates quickly, and crystals of a solid salt crust form on their surface. At a later stage of the process, the remaining solution is removed from the droplet using a jet of water vapor that passes through the pores of the polycrystalline crust. It was first observed that some of the drying droplets suddenly shifted to one side under the influence of the reactive force generated by the vapor jet. The resulting salt particles are weakly porous and consist of many crystals. It has been proven that these particles don’t have a central cavity. The use of seawater and the role of salt particles in protecting against thermal radiation from fires are briefly discussed. Calculations based on Mie theory have shown that the contribution of light scattering by hollow sea salt particles formed above the ocean surface during relatively slow evaporation of seawater droplets can be significant in the ocean's heat balance.

Abstract: Physics makes two questionable assumptions: (1) Distant galaxies are accelerating relative to Earth. (2) Entangled objects are spatially separated from each other. Why questionable? Acceleration relative to Earth has never been observed in a single galaxy. Observers perceive entangled objects as spatially separated, yet 3D space is relative. We show that physical realities are projections of a mathematical background reality: 4D Euclidean space (ES). In Euclidean relativity (ER), all objects move through ES at the speed C. There is no time coordinate in ES. All action is due to a monotonically increasing, absolute, external evolution parameter θ. An observer experiences two projections of ES as space and time. The axis of his current 4D motion is his proper time τ. Three orthogonal axes form his 3D space x₁, x₂, x₃. His physical reality is his spacetime x₁(ϑ), x₂(ϑ), x₃(ϑ), τ(ϑ), where τ is a natural time coordinate and θ converts to absolute parameter time ϑ. Without gravity, his spacetime is Minkowski-like. As in general relativity (GR), gravity in ER is the curvature of spacetime. Since coordinates in GR are merely labels, the Einstein field equations also hold in systems that use τ as the time coordinate. ER predicts time’s arrow, relativistic effects, galactic motion, the Hubble tension, and entanglement. Remarkably, ER manages without cosmic inflation, expanding space, dark energy, and non-locality. ER tells us: (1) Distant galaxies maintain their recession speeds. (2) From their perspective, entangled objects have never been spatially separated, yet their proper time flows in opposite 4D directions.

Abstract: Let k be a positive integer. A polynomial A∈F_2[x] is called k-unitary perfect if the sum of the k-th powers of its distinct unitary divisors equals A^k. In this paper, we focus on the case k=2^n and prove that every 2^n-unitary perfect polynomial over F_2 is necessarily even. Moreover, we obtain a complete classification of all even 2^n-unitary perfect polynomials having at most three distinct irreducible factors. In particular, we characterize all such polynomials of the formA=x^a (x+1)^b P^h, where P is a Mersenne prime over F_2 and a, b, and h are positive integers. As a consequence, several explicit infinite families of k-unitary perfect polynomials over F_2 are obtained.

Abstract: Abstract—In-vehicle intrusion detection systems (IDSs) are increasingly proposed to protect automotive networks, yet most prior work emphasizes detection accuracy while overlooking system-level constraints that determine real-world deployability. This paper addresses the mismatch between IDS design assumptions and the computational, architectural, and real-time limitations of production automotive electronic control units (ECUs). This issue is particularly critical in safety-critical automotive systems, where security mechanisms must operate within strict timing and resource bounds without interfering with control functions. The objective of this work is to provide a deploymentaware feasibility analysis of in-vehicle IDS techniques across heterogeneous automotive computing platforms. We introduce a baseline-driven methodology that defines two representative ECU tiers: microcontroller-based safety ECUs operating under AUTOSAR Classic and higher-performance domain or zonal controllers based on AUTOSAR Adaptive and POSIX-compliant operating systems. IDS approaches are evaluated against nonnegotiable constraints including processing capacity, memory availability, worst-case execution time, operating system compatibility, and in-vehicle network technology. The results show that microcontroller-based ECUs support only lightweight, messagelevel IDS mechanisms with strictly bounded execution behavior, while machine learning–based IDSs require controller-class platforms and remain constrained by determinism and interference requirements. This work demonstrates that feasibility, rather than accuracy alone, must be treated as a first-class criterion in automotive IDS design.

Abstract: Guardrail systems for large language models (LLMs) are designed under a foundational but rarely examined assumption: that safety is a property of individual input–output exchanges. This assumption is adequate for single-turn deployments but fails structurally in multi-turn conversational systems, where risk does not reside in any single message but emerges from the accumulated trajectory of a session. We formalize this failure mode as Conversational Risk Accumulation (CRA), a class of adversarial and incidental threat patterns in which individually policy-compliant turns collectively produce outcomes that violate safety intent. We propose a stateful guardrail architecture, the CRA Framework, comprising three novel constructs: (1) a Semantic Drift Monitor that tracks divergence from declared session intent; (2) an Information Accumulation Graph (IAG) that models cross-turn entity and attribute disclosure as a growing knowledge structure; and (3) a Compliance Gradient Detector that identifies progressive erosion of refusal behavior across turns. These three signals are fused into a session-level CRA Score, which triggers guardrail intervention at the conversation layer rather than the message layer. We formalize the threat taxonomy, define the mathematical properties of the CRA Score, and derive theoretical bounds on detection latency. The framework is domain-agnostic and architecturally composable with existing single-turn guardrail systems. We discuss instantiation across the enterprise RAG deployments, agentic pipelines, and educational AI systems, and identify open problems in stateful safety that the framework surfaces.

Abstract: Background/Objectives: In laparoscopic and robot-assisted surgery, bleeding may rapidly impair operative-field readability and procedural safety. In the broader Robin Heart teleoperation framework, interpretation of such events is relevant not only for scene understanding, but also as a potential prerequisite for future safety-oriented supervisory functions under communication-degraded conditions. The aim of this study was to assess whether a deep learning model for blood segmentation could provide outputs suitable for preliminary image-level temporal characterization of visible blood-region behavior in laparoscopic video. Methods: The model was first trained on a simulated bleeding dataset prepared under controlled conditions and then fine-tuned on annotated frames from robot-assisted laparoscopic hysterectomy video. Additional limited adaptation and held-out evaluation were performed on annotated bleeding-related episodes derived from the public GynSurg dataset. Segmentation performance was assessed using the Dice coefficient and Intersection over Union (IoU). Temporal analysis was performed on representative internal and external sequences using mask-derived descriptors and auxiliary optical-flow-based motion descriptors computed after camera-motion compen-sation within the detected blood ROI. Results: The model achieved Dice/IoU values of 0.94/0.89 on the simulated validation set, 0.907/0.830 on the internal operative validation set, and 0.764/0.626 on the annotated external GynSurg subset. The combined descriptor set differentiated more dynamic and unstable progression profiles from more spatially coherent ones across both datasets. Peak dA/dt reflected abrupt visible blood-area ex-pansion, temporal IoU described mask stability over time, and optical-flow-based de-scriptors provided additional information on local motion activity. A peak-only descrip-tion was insufficient to fully characterize the observed progression patterns. Conclusions: The results support the feasibility of combining deep-learning-based blood segmentation with temporal and optical-flow-based descriptors for exploratory image-level character-ization of visible blood-region behavior in laparoscopic video. Within the Robin Heart development pathway, such descriptors may in the future serve as candidate components of image-analysis support modules for safety-oriented teleoperative scenarios. At this stage, they should be interpreted as exploratory image-derived indicators rather than clinically validated markers of bleeding severity.

Abstract: Manual compliance auditing in cloud environments consumes up to 40% of IT security budgets annually, yet existing approaches verify control presence rather than effectiveness, leaving institutions vulnerable to adversarial evasion. This paper presents an AI-augmented hybrid ML–LLM compliance auditing system evaluated on a national cybersecurity standards framework (143 controls, 200,000 training events). The system combines multi-label XGBoost classification with LLM-based semantic log analysis, grounded in a formal effectiveness model. Key findings: XGBoost achieves 99.88% F1 after 5% domain fine-tuning but collapses to 7.98% zero-shot, a 92-point generalization gap bridged by the hybrid LLM path; adversarial validation exposes effectiveness deficits invisible to checkbox auditing (SI-3: 20%detection rate; SI-10: 32% XSS bypass); GPT-4o-mini achieves 93.5% zero-shot accuracy across four log types (n=200), while Llama-3.2-3B on CPU-only hardware achieves 84.0%, validating on-premise deployment viability. A vocabulary-coverage gating router achieves 94.5% accuracy at $0.15/10K logs. The system runs at 2.0 CPU cores, $50/month, producing audit reports in 0.77s, demonstrating that effectiveness-based compliance auditing is accessible without enterprise-grade infrastructure.

Abstract: Background and Aim: Liver disease in patients receiving chronic hemodialysis (HD) is frequently underrecognized and may contribute to adverse outcomes. Its evaluation is complicated by dialysis-related changes in volume status that can influence noninvasive measurements. We aimed to assess the prevalence and determinants of Metabolic Dysfunction–Associated Steatotic Liver Disease (MASLD) and liver fibrosis using transient elastography (TE), with particular attention to metabolic, volume-related, and dialysis-specific factors, and to examine the effect of a single HD session on Controlled Attenuation Parameter (CAP) and liver stiffness measurement (LSM). Methods: In this prospective cross-sectional study, adult patients receiving maintenance HD for ≥3 months were enrolled. TE (FibroScan®) was performed immediately before (pre-HD) and after (post-HD) a midweek HD session. Steatosis was defined as CAP >257 dB/m and significant fibrosis as LSM ≥8 kPa. Volume status was assessed using interdialytic weight gain (IDWG) and ultrafiltration (UF) volume. Associations with clinical, biochemical, and dialysis-related variables were analyzed. Results: Forty patients were included (mean age, 63.5 ± 17.3 years; 70% male; 30% diabetic). MASLD prevalence was 25% pre-HD and 22.5% post-HD. The prevalence of significant fibrosis decreased from 17.5% pre-HD to 10% post-HD, while mean LSM did not change significantly, suggesting that volume removal may influence fibrosis classification in a subset of patients. CAP correlated positively with body mass index, IDWG, UF volume, and triglyceride levels, and inversely with serum albumin. Post-HD LSM showed a significant inverse association with serum albumin and positive associations with alanine aminotransferase, total cholesterol, and low-density lipoprotein cholesterol, whereas pre-HD LSM was mainly associated with age and bilirubin. The FIB-4 index did not correlate with LSM. Conclusions: In maintenance HD patients, hepatic steatosis is closely associated with metabolic burden and fluid overload, whereas fibrosis assessment is substantially influenced by nutritional status. Dialysis-related volume changes may modify LSM-based fibrosis estimates, supporting post-dialysis TE for more reliable assessment. TE provides clinically informative evaluation and appears superior to surrogate indices such as FIB-4 in this population.

Abstract: Retrofitting existing residential buildings is a critical strategy for achieving urban decarbonization while addressing public health disparities, particularly in communities disproportionately affected by environmental and socioeconomic stressors. This study presents a scalable urban building energy modeling framework that integrates physics-based simulations with machine learning to evaluate and prioritize health-driven retrofit strategies across residential building stocks. Synthetic datasets were generated through parametric simulations of representative building archetypes and retrofit scenarios, capturing variations in envelope performance, HVAC systems, infiltration rates, and ventilation strategies. Machine learning models were trained as surrogate predictors of building energy performance, enabling rapid evaluation of retrofit impacts. A range of algorithms—including decision trees, random decision forests, gradient boosting machines, support vector machines, k-nearest neighbors, and artificial neural networks—were evaluated. An artificial neural network implemented as a multilayer perceptron was selected for further analysis due to its strong predictive performance (R² = 0.94) and ability to capture complex nonlinear relationships among retrofit variables. The final model used the Port optimization algorithm for stable convergence and improved generalization. The framework is applied to Seattle’s Duwamish Valley, a community experiencing disproportionate environmental and health burdens. The results highlight retrofit priorities—particularly infiltration reduction, HVAC upgrades, and improved envelope performance—that deliver co-benefits for energy efficiency, indoor environmental quality, and occupant health. The results demonstrate that machine learning–enhanced physics-based UBEM can significantly accelerate retrofit evaluation while preserving the interpretability of simulation-based approaches. The proposed framework provides a scalable approach for identifying health-informed retrofit pathways that support equitable urban decarbonization.