Where is the line between firm discipline and child abuse? For millions of families in the United States, the answer depends on where they grew up, what their parents did, and how their culture defines good parenting. A new study published in the Journal of Child and Family Studies tackles this thorny question with an ambitious goal: building a scientific instrument that can measure how people from different backgrounds judge parenting practices that are accepted in some communities but condemned in others. The Parenting Practices and Discipline Scale, or PPDS, developed by Kevin Ortiz Diaz, Maureen C. Kenny, and Helia Baez at Florida International University, is one of the first tools designed specifically to capture culturally situated discipline practices that existing measures simply ignore.

The stakes are enormous. Child abuse is recognized as a global health issue, and in the United States it occurs at what researchers describe as epidemic proportions. Federal data show 546,159 victims of child abuse and neglect in 2023 alone, and the vast majority of these cases, roughly 89 percent, are committed by one or both parents. Children from minority groups are often the most affected, with Black or African American children among the most likely to experience abuse, second only to American Indian or Alaska Native children. Yet the definitions of what constitutes abuse vary dramatically across countries and cultures, creating a legal and ethical minefield for immigrant families who may not realize that practices considered normal at home are illegal in their new country.

The problem is particularly acute for Hispanic families, who make up 19 percent of the American population and represent the largest share of recent immigrants. Hispanic culture is far from monolithic, encompassing communities from South and Central America and the Caribbean with distinct traditions, but certain common threads run through the research. The concept of familismo, a cultural orientation and sense of obligation to family, has been linked to greater parental warmth in some studies. At the same time, physical violence as a disciplinary method remains common in many Hispanic cultures, and the chances that such behaviors are carried over by families immigrating to the United States are high. The result is a potential collision between generations of cultural practice and a legal system, anchored by the Child Abuse Prevention and Treatment Act of 1974, that defines abuse in terms of serious physical or emotional harm.

What makes the new study distinctive is its grounding in intersectionality theory, a framework rooted in feminist scholarship that examines how multiple categories of identity, such as race, ethnicity, gender, and nativity, interact to shape experience. Rather than treating demographic characteristics as independent variables, the researchers designed their analyses to test whether these dimensions jointly predict how individuals classify discipline practices as culturally acceptable or abusive. The framework has practical urgency: racial and ethnic disparities exist in both the reporting and substantiation of child maltreatment, and current child protection policies can disproportionately affect women of color who use social services and face heightened surveillance and risk of reports to child protective services.

The scale itself emerged from a careful two-stage piloting process. First, four experienced mental health clinicians specializing in child abuse, with 10 to 20 years of experience each, reviewed a draft pool of 54 items describing parenting behaviors and case vignettes. Then ten first-generation Hispanic immigrants from a range of nationalities, including Puerto Rican, Cuban, Colombian, Mexican, Dominican, and Nicaraguan participants, examined the refined items and identified which behaviors they would associate with parents in their own cultures and whether they would label those behaviors abusive or acceptable. Consensus required agreement from at least 70 percent of the pilot group. Eighteen items survived, ten identified as culturally acceptable behaviors such as yelling, using the chancla, the slipper famously wielded in many Latino households, or letting a child sleep with parents, and eight identified as abusive behaviors such as spanking, leaving a child alone in a car, or forcing a child to bathe with an adult.

The main study recruited 1,477 undergraduates at a large Hispanic-Serving Institution in Miami, ultimately yielding an analytic sample of 1,089 participants after exclusions for failed attention checks, missing data, and groups too small for meaningful analysis. The final sample was 72.2 percent Hispanic, 16.3 percent non-Hispanic Black, and 11.6 percent non-Hispanic White, nearly 70 percent U.S.-born, and predominantly female at 88.8 percent. Participants rated each of the 18 behaviors on a three-point scale: not abusive, inappropriate but not abusive, or abusive. The choice of a college population was deliberate, since emerging adults are on the cusp of parenthood, attitudes at this life stage predict later parenting intentions, and parenting practices tend to be transmitted intergenerationally.

Psychometrically, the study unfolded in two parts. Study 1 established the scale’s initial properties using exploratory factor analysis with principal axis factoring and oblique rotation on a polychoric correlation matrix, an approach suited to the ordinal, non-normally distributed responses. The analysis supported a two-factor structure: Factor 1 captured discipline and psychological aggression, including spanking, yelling, the chancla, cursing, ignoring, comparing, fighting, and religion, while Factor 2 reflected neglect and physical endangerment, such as using drugs in front of a child, refusing treatment for illness, and allowing smoking. The two factors accounted for 29.5 percent of the variance and were moderately correlated. Internal consistency for the full 18-item measure was acceptable at a Cronbach’s alpha of 0.71, though the abusive subscale showed weaker reliability, a limitation the authors acknowledge candidly.

Study 2 then deployed regression models to test how demographic identities shaped perceptions, and the results were revealing. Gender and parental status emerged as the strongest and most consistent predictors: women and parents were significantly more likely to classify culturally acceptable practices as abusive. Among Hispanic participants, those not born in the United States were more likely than U.S.-born Hispanics to view behaviors like spanking as abusive, a finding that ran counter to the researchers’ hypothesis and suggests that acculturation may shift parenting beliefs toward American child protection norms. Across racial and ethnic groups, Hispanic participants were more likely than Black participants to classify culturally acceptable practices as abusive. Notably, for the more overtly harmful behaviors in Factor 2, parental status was the only significant predictor, suggesting broad agreement that neglecting a sick child or encouraging substance use transcends cultural and gendered expectations.

The effect sizes were small, with the models explaining roughly 1 to 6 percent of the variance, a reminder that unmeasured cultural and contextual factors account for most of the variation in how people judge discipline. Still, the patterns carry real implications for prevention. The authors suggest that father-focused education programs may be valuable, since men were less likely to identify physical discipline as harmful and may therefore be at greater risk of normalizing such practices. Culturally adapted interventions have already shown promise, from brief summer parent training programs benefiting Hispanic children to the multinational Triple P Positive Parenting Program, and the PPDS could help practitioners identify which disciplinary practices are perceived as acceptable within specific communities, allowing messaging to be tailored rather than uniform.

The researchers are careful to frame their findings as preliminary. The sample was heavily female and drawn from a single Miami institution in a minority-majority city, socioeconomic status was not assessed, and social desirability bias may have influenced responses to questions about parenting. The abusive subscale needs refinement, and future work should apply confirmatory factor analysis and measurement invariance testing across cultural groups. Yet the core contribution stands: a novel measure that captures culturally situated discipline practices invisible to existing tools like the Alabama Parenting Questionnaire or the Child Abuse Potential Inventory. As parenting practices pass from generation to generation and families cross borders into legal systems with different rules, understanding exactly how culture shapes perceptions of abuse may be the first step toward protecting children without criminalizing the traditions their parents brought with them.

Subject of Research: Development and validation of a scale measuring cultural perceptions of parenting and discipline practices

Article Title: Development and Validation of the Parenting Practices and Discipline Scale in an Ethnically Diverse Sample

Article References: Ortiz Diaz, K., Kenny, M. C., & Baez, H. (2026). Development and Validation of the Parenting Practices and Discipline Scale in an Ethnically Diverse Sample. Journal of Child and Family Studies. https://doi.org/10.1007/s10826-026-03370-6

Image Credits: AI Generated

DOI: 10.1007/s10826-026-03370-6

Keywords: child abuse, parenting practices, discipline, Hispanic families, intersectionality, psychometrics, scale development, cultural perceptions, child welfare, corporal punishment, acculturation, exploratory factor analysis

Buildings are responsible for an enormous share of global electricity demand, and the way their air conditioners, lights, and heating systems are scheduled can make the difference between a grid under strain and one running smoothly. For years, engineers have tried to optimize these schedules with evolutionary algorithms and reinforcement learning, but the problem becomes brutally hard when dozens of buildings share a single electrical network and must be coordinated simultaneously. A new study published in Discover Artificial Intelligence proposes a strikingly different approach: instead of searching step by step for the best schedule, a generative artificial intelligence model learns to paint one directly out of pure noise, guided by a map of how the buildings are physically wired together.

The framework, called DMCBO, short for Diffusion-driven Multi-objective Collaborative Optimization for Building Clusters, was developed by Yalan Zheng of Zhengzhou Technical College and Taixiang Yin of North China University of Water Resources and Electric Power. Their central insight is that the joint scheduling of equipment settings and load distribution across an entire cluster of buildings can be treated as a high-dimensional data sample, much like an image. Diffusion models, the same family of generative techniques behind modern text-to-image systems, work by progressively corrupting data with Gaussian noise and then learning to reverse that corruption. The researchers reformulated multi-objective building optimization as exactly this kind of conditional denoising process: starting from random noise, the model strips away noise step by step until a complete, Pareto-optimal collaborative schedule for every piece of equipment in every building emerges.

Technically, the forward diffusion process is a fixed Markov chain that adds tiny amounts of Gaussian noise to real scheduling data over one hundred steps, with the noise coefficient increasing linearly from 0.0001 to 0.02 until the original data becomes indistinguishable from standard Gaussian noise. A neural network is then trained to predict the noise that was added at each step, given the noisy schedule and a conditional encoding vector describing the current situation. Once trained, the model can generate fresh solutions at inference time by sampling pure noise and running the reverse process, refining it over one hundred denoising steps into a schedule tailored to the prevailing weather, occupancy, and grid conditions. Because this generation is a gradient-based learning process rather than an iterative search, it avoids the local optima traps that plague traditional multi-objective evolutionary algorithms and the reward oscillations that destabilize deep reinforcement learning.

What makes the system genuinely cluster-aware is its use of graph neural networks. The researchers model the building cluster as a graph mapped onto a radial distribution microgrid topology, where nodes represent buildings and edges encode physical constraints such as maximum line thermal capacities and transformer limits. A two-layer graph convolutional network passes messages between neighboring buildings, fusing spatial-topological dependencies and network constraints into a single graph embedding. This embedding is concatenated with an adjustable preference weight vector to form the condition that steers the denoising trajectory. In effect, when the model generates a control strategy for one building, it simultaneously perceives the state of its neighbors and the limits of the shared electrical infrastructure, suppressing systemic coordination errors from 0.18 down to 0.06 in the team’s experiments.

The multi-objective aspect is handled through an elegant scalarization trick. Each objective, such as energy consumption, operating cost, or thermal comfort, is normalized to a dimensionless value between zero and one using Min-Max normalization based on historical operational bounds. The normalized objectives are then blended into a single scalar condition via a weight vector whose components sum to one. During training, these weights are sampled from a uniform Dirichlet distribution, exposing the model to the entire space of possible preferences so it learns the distribution of the full Pareto front. At deployment, a decision maker can slide the weight vector like an adjustment knob, instantly generating schedules that emphasize energy saving, comfort, or a balanced compromise, all within the same trained model.

The denoising backbone itself is a U-Net adapted to process one- and two-dimensional building scheduling tensors. Diffusion time steps are embedded through sinusoidal positional encodings, while the graph-derived condition is injected into every residual block via Adaptive Layer Normalization. A bottleneck layer with multi-head self-attention captures long-range dependencies across the entire daily scheduling horizon, and symmetric skip connections preserve high-resolution information during reconstruction. The authors trained the model for 300 epochs on dual NVIDIA Tesla V100 GPUs, and the loss curves dropped from 0.75 to below 0.03 with training and validation curves closely coinciding, indicating no overfitting. As the reverse diffusion step count decreased from 90 to 10 during analysis, the noise-prediction error distribution became sharply concentrated at zero, demonstrating extremely high denoising precision at the end of generation.

Validation was carried out on the open-access Building Data Genome Project 2 dataset, which supplied real hourly weather records, operational schedules, and annual electricity, cooling, and heating load profiles across diverse climate zones. Because that dataset provides metered energy data rather than physical building models, the team used United States Department of Energy Commercial Reference Buildings as templates and calibrated them in EnergyPlus, a high-fidelity thermodynamic simulation platform, to create a digital twin consistent with the real data. All results were benchmarked against a rule-based control strategy with static thermostat setpoints of 25 degrees Celsius in summer and 22 degrees in winter, and every experiment was repeated ten independent times to confirm statistical significance.

The headline numbers are impressive. In high-demand summer conditions, DMCBO achieved an average energy-saving rate of 28.6 percent and cut peak load by 38.5 percent, shaving the afternoon peak from roughly 560 kilowatts down to about 440 kilowatts and saving approximately 142.80 US dollars per day. Cumulative daily power savings exceeded 1,000 kilowatt-hours, compared with 400 kilowatt-hours for a traditional collaborative strategy and just 120 kilowatt-hours without coordination. In winter, peak shaving and valley filling reached about 39 and 42 percent respectively, far above the 22 and 25 percent of conventional methods. When the solution sets were projected into objective space, DMCBO dominated the baselines: it reached a hypervolume indicator of 0.95 within 12,000 function evaluations, versus 0.85 for NSGA-II and 0.70 for MOPSO, while keeping operating costs between 140 and 160 dollars against more than 210 dollars for deep reinforcement learning and 260 to 300 dollars for a GAN-based generative approach.

Robustness and scalability proved equally compelling. Under injected stochastic disturbances mimicking sudden weather changes and occupancy swings, the framework confined power fluctuations to a narrow band of plus or minus 5 kilowatts, whereas the deep reinforcement learning baseline swung between minus 15 and plus 20 kilowatts. Scaling from 5 to 50 buildings, the global energy-saving rate actually rose from 28.5 to 31.2 percent, because larger clusters unlock more load-complementary opportunities that the graph encoder can safely exploit within encoded grid boundaries. Computation time grew to 9.8 seconds at the 50-building scale, but that remains comfortably within the 15-minute control horizon used by real building management systems, and the hypervolume index slipped only slightly to 0.88. Ablation experiments confirmed that both the graph neural network module and the choice of 100 diffusion steps, which yields a 50-millisecond inference latency, were essential to this balance of speed, accuracy, and stability.

The authors are candid about limitations. The framework has so far been validated only in simulation, and EnergyPlus simplifies sub-minute mechanical actuation latencies that real HVAC hardware would introduce; they propose cascading PID loops and predictive safety margins to absorb those delays in practice. Data quality, generalization to residential buildings and extreme climates, the computational cost of scaling to city-sized networks, and the black-box nature of the generative process all remain open challenges, with explainable AI and transfer learning flagged as future work. Still, the architecture is designed to slot into standard building management systems as a supervisory layer, transmitting setpoints over protocols such as BACnet or Modbus. If the simulation gains survive contact with physical buildings, the idea of generating optimal energy schedules from noise, guided by a graph of the grid itself, could become a powerful new tool in the push toward decarbonized cities.

Subject of Research: Conditional diffusion models with graph neural networks for multi-objective energy optimization in building clusters

Article Title: Diffusion model-driven multi-objective collaborative optimization for building energy management using graph neural networks

Article References: Zheng, Y., & Yin, T. (2026). Diffusion model-driven multi-objective collaborative optimization for building energy management using graph neural networks. Discover Artificial Intelligence, 6(1), Article 1383. https://doi.org/10.1007/s44163-026-02140-z

Image Credits: AI Generated

DOI: 10.1007/s44163-026-02140-z

Keywords: diffusion models, graph neural networks, building energy management, multi-objective optimization, EnergyPlus, smart buildings, Pareto optimization, deep learning, energy efficiency, HVAC control, microgrids, generative AI

Mr. Musk, whose company SpaceX owns Starlink, has attacked the top executive of Delta Air Lines for not selecting the internet provider.

September falls in the heart of the monsoon season in Thailand, where cities and towns are accustomed to the afternoon rains and thunderstorms that frequently roll through. But throughout late September and early October 2026, persistent heavy rains punctuated by intense downpours left low-lying areas across the country in Southeast Asia under sustained flooding.

Prachin Buri province was among the areas where flooding was widespread. The right image above shows land along the Bang Pakong River and nearby agricultural areas and aquaculture facilities inundated on October 2, 2026. The left image shows the same area at the same time of year in 2025. Acquired with NASA-USGS Landsat satellites, the images are false-color to better distinguish water from vegetation, developed land, and other surfaces. The bright red shapes are floating solar arrays.

Some of the flooding likely stemmed from a low-pressure system that stalled over the area for several days in late September. Stoked by moisture and energy from the warm Gulf of Thailand nearby, the system delivered especially intense rains to the region. About 100 kilometers (60 miles) to the west, Bangkok received nearly 300 millimeters (11.8 inches) of rain over a 48-hour period, according to news reports, close to the average monthly rainfall total for September of 344 millimeters (13.5 inches).

Since mid-September, flooding has occurred in 50 provinces across the country, affecting over 4 million people and causing dozens of deaths, according to the International Federation of Red Cross and Red Crescent Societies. Government agencies estimate about 270,000 hectares (670,000 acres) of farmland, including rice fields and fruit orchards, have been damaged.

NASA Earth Observatory images by Lauren Dauphin, using Landsat data from the U.S. Geological Survey. Story by Lindsey Doermann.

References & Resources

• ABC News (2026, September 28) Residents brave floodwater for supplies after Bangkok inundated. Accessed October 7, 2026.
• The Conversation (2026, October 5) As the planet warms, cities like Bangkok could be increasingly inundated by deadly floods. Accessed October 7, 2026.
• The Nation (2026, October 7) Thailand floods hit 1.7m rai including over 26,000 rai of orchards. Accessed October 7, 2026.
• ReliefWeb (2026, October 6) THA: Flood – 09-2026 – Heavy rainfall caused flooding #3. Accessed October 7, 2026.

Surfshark One réunit VPN, antivirus et outils de protection dans une même application. Le pack est actuellement proposé à 2,79 €/mois, avec 85 % de réduction et jusqu’à 3 mois offerts.

Avec son format 14 pouces, ses 16 Go de RAM et son affichage au format 16:10, l’Acer Aspire Go 14 réunit plusieurs éléments recherchés sur un PC portable destiné au travail et aux usages quotidiens légers. Cette configuration assurant de belles performances est actuellement proposée à prix...

Conçu pour les usages courants et offrant une complément calibré, le CHUWI CoreBook i3 propose un PC portable offrant un écran de 14 pouces et un SSD de 256 Go. Ce modèle très abordable est en ce moment au meilleur prix sur AliExpress où il s'affiche à moins de 240 € grâce à une remise de 10 %....

La glacière électrique portable VEVOR 53 L met les provisions au frais et les utilisateurs d’accord. Avec plus de 1 000 commentaires positifs, sa grande capacité, son froid jusqu’à -20 °C et son format nomade ont de solides arguments pour prendre la route.

Bose baisse le volume sur le prix, mais certainement pas sur le son. L’enceinte sans fil portable Bluetooth Bose SoundLink Plus tombe à 179,99 € au lieu de 199,95 € chez Amazon, soit 10 % de remise. Avec son autonomie généreuse, sa certification IP67 et ses basses musclées, elle a de sérieux...

Face à un SMS frauduleux, le réflexe immédiat consiste à ne jamais cliquer sur le lien, à transférer le message au 33700, puis à bloquer l’expéditeur avant de supprimer la conversation. Cette arnaque, appelée smishing, exploite l’urgence et l’usurpation d’identité pour voler des données...

Ecovacs muscle son milieu de gamme avec le Deebot T90S Pro Omni. Lavage par rouleau, aspiration de 40 000 Pa et intelligence artificielle : ce nouveau robot aspirateur compte bien faire une croix sur les corvées de ménage.

L’iScooter iX3 sort les watts et allège la facture chez Cdiscount. Cette trottinette électrique pliable de 800 W combine autonomie, confort et équipement complet pour donner un sérieux coup d’accélérateur aux trajets du quotidien.

Italie : Au coeur d’une ville vivante, entre rires de passants et cliquetis de verres, elle se dresse — immobile, immense, indifférente au tumulte. L’Arène de Vérone, cicatrice sublime d’un empire révolu, attend. Sous ses arches érodées, les siècles se taisent. Un silence dense, presque sacré,...

À 11,99 €/mois, le forfait RED by SFR 100 Go 5G offre une enveloppe confortable, les appels et SMS/MMS illimités, sans engagement, avec 30 Go utilisables en Europe.

Avis à tous les jardiniers qui souhaitent entretenir leurs arbres et arbustes sans se fatiguer ni prendre de risques inutilement ! Cet élagueur sans fil avec perche télescopique fait l'objet d'une baisse de prix particulièrement agressive

Avast One Premium ne se limite pas à bloquer les virus : la suite ajoute un VPN, des outils de confidentialité et plusieurs protections pour sécuriser les usages quotidiens.

Associant une dalle AMOLED de presque 7 pouces, une conséquente batterie de 7 700 mAh et une configuration équilibrée à 6 Go de RAM, le Xiaomi Redmi Note 17 est un vrai exemple de ce qu'offre un bon rapport qualité-prix. En ce moment, il est possible de l'acquérir pour moins de 255 € en...

Le Cotiar S25 est un aspirateur-balai sans fil performant conçu pour passer rapidement des sols aux zones plus difficiles d’accès. Il bénéficie actuellement d’une remise sur AliExpress, où il s'affiche à moins de 95 € sans compte le coupon de réduction de 4 € disponible pour tous et toutes.

VEVOR sort les dents avec cette tronçonneuse thermique de 62 cm³ taillée pour les travaux musclés. Puissance, gros diamètre de coupe et promo bien affûtée sont au programme.