Payer un abonnement tous les mois pour utiliser un VPN n’attire pas forcément tout le monde. Beaucoup d’utilisateurs cherchent aujourd’hui une solution plus souple. Une offre gratuite, un plan sans engagement ou une version limitée peuvent suffire pour protéger ses données sur internet, regarder...
L’Asteroid Day rassemble chaque année chercheurs, passionnés d’astronomie et curieux autour des enjeux liés aux objets célestes. Mais l’événement dépasse désormais largement les conférences physiques : une grande partie des échanges se déroule en ligne, entre retransmissions, forums spécialisés...
O Copilot Cowork, que passa a estar disponível a nível internacional, é um agente de IA que promete tratar do trabalho real e ajudar na produtividade. Tarefas que levavam dias a concluir passam agora a ser finalizadas numa manhã, afirma a Microsoft.
In recent years, the evolution of industrial sensing technologies has been pivotal to the advancement of intelligent manufacturing systems. Among various approaches, artificial neural networks (ANNs) have garnered immense attention due to their remarkable capability to model complex data relationships. However, applying these networks effectively in real-world industrial environments remains a formidable challenge, primarily due to the noisy, nonlinear, and often limited datasets encountered in such settings. Recognizing these constraints, a team of researchers led by Professor Lanxiang Sun at the Shenyang Institute of Automation, Chinese Academy of Sciences, has forged a transformative approach. Their groundbreaking solution, known as the Partial Least Squares-assisted Optimization Network (PLSaoNET), promises to redefine the application of neural networks in industrial sensing by embedding statistical interpretability into the heart of deep learning models.
Traditional neural networks typically commence training through random initialization of weights—a process that, while standard, lacks any guiding framework grounded in data characteristics. This randomness poses inherent limitations, especially when the amount of training data is limited or when the data exhibits high noise levels and nonlinear relationships, both common in industrial scenarios. Such conditions often result in poor generalization and unstable predictive outcomes, thereby hindering practical deployment. The innovation from Professor Sun’s team circumvents these issues through a statistical method known as Partial Least Squares (PLS), a classic tool renowned for its interpretability and robustness in linear regression tasks.
PLS works by extracting latent variables that capture the fundamental relationships between independent variables (input features) and dependent variables (output targets), offering a transparent statistical framework. While PLS excels in linear modeling, it lacks the capability to model complex nonlinear dynamics. The research team ingeniously leveraged the strengths of PLS by integrating its solutions as the initial conditions within a neural network architecture. Specifically, the number of PLS latent variables dictates the number of neurons in the network’s hidden layer, and the PLS-derived weight matrix initializes the neural connections. This initialization endows the neural network with a statistically meaningful starting point, effectively eliminating the guesswork inherent in random weight assignments.
Such a well-founded initialization transforms the learning process from an unfocused blind search into a systematic refinement of the PLS solution, integrating nonlinear modeling capabilities while preserving interpretability. This fusion of statistical rigor and deep learning adaptability addresses one of the fundamental bottlenecks in industrial data modeling, enabling the network to capture both the global trends identified by PLS and the nuanced nonlinear interactions vital for accurate predictions.
Moreover, industrial datasets frequently suffer from uneven label distributions, posing significant challenges during the training of neural networks using stochastic gradient descent. Fluctuations in loss values caused by biased mini-batch samples can lead to instability and degrade the overall learning process. To counter this, the researchers implemented a stratified sampling-based training strategy. In this approach, the entire dataset is divided into strata based on label value ranges, ensuring that each mini-batch during training samples from each stratum proportionally. This methodology smooths gradient updates, fortifying the network’s robustness against skewed data distributions and enhancing training stability substantially.
Professor Sun has emphasized that the primary motivation behind PLSaoNET is not to pursue architectural complexity but to advocate for models that produce stable, reliable outputs even when faced with small, noisy datasets that typify industrial environments. By embedding PLS’s statistical solutions into neural network training, the model’s learning trajectory gains a trustworthy, transparent foundation from which to build more complex representations.
To validate the efficacy of PLSaoNET, the research team conducted comprehensive experiments in two realistic industrial applications: the online monitoring of iron ore concentrate slurry grade using laser-induced breakdown spectroscopy (LIBS), and diesel fuel quality assessment with near-infrared (NIR) spectroscopy. Both cases involve complex spectral data with significant noise and nonlinear characteristics—ideal testbeds to showcase the robustness of PLSaoNET. Results consistently demonstrated superior prediction accuracy and generalization capabilities over traditional PLS regression and conventionally initialized backpropagation neural networks.
A particularly revealing insight emerged from the visualization and analysis of hidden layer weights in the LIBS slurry model. While the PLS model alone lacked precision, the dominant latent variables it extracted aligned clearly with known spectral emission lines from key mineral elements like silicon, magnesium, and iron, reflecting true physical relevance. In contrast, networks initialized randomly exhibited neuron weights that failed to selectively identify meaningful spectral features, resulting in noisy, uninterpretable representations. The PLSaoNET model, however, preserved the interpretability inherent to PLS by focusing attention on these dominant spectral features while simultaneously allowing fine-tuned retraining that accentuated secondary information related to elements like calcium and aluminum. This balance effectively suppresses overfitting and elevates model reliability.
Not content with theoretical validation alone, the team successfully deployed PLSaoNET within a live LIBS slurry analyzer at a mineral processing plant for real-time iron ore concentrate grade monitoring. This real-world application underscores PLSaoNET’s practical value in lowering the barriers to widespread adoption of artificial neural networks in industrial contexts. By anchoring the initial network parameters in statistically meaningful solutions and employing thoughtful training strategies, PLSaoNET represents a crucial step toward integrating cutting-edge AI models into complex, noise-prone industrial environments where data scarcity and variability have historically impeded progress.
Looking beyond its immediate applications, PLSaoNET delivers a broader paradigm shift, suggesting that harmonizing classical statistical techniques with modern deep learning can unlock new frontiers in intelligent industrial sensing. As manufacturing systems increasingly embrace digital transformation, methods like PLSaoNET that offer interpretability, stability, and adaptability will become indispensable tools for building resilient, data-driven infrastructures.
In conclusion, the work by Professor Lanxiang Sun and colleagues embodies a fusion of statistical principled approaches and adaptive machine learning techniques, producing an ANN model that addresses longstanding challenges in industrial data analysis. With its innovative PLS-based initialization and stratified sampling training, PLSaoNET not only enhances predictive performance but also paves the way for more interpretable, reliable AI systems in industrial manufacturing. As industries worldwide continue to seek intelligent sensing solutions that function robustly under real-world constraints, PLSaoNET stands out as a promising, practical breakthrough.
Subject of Research: Advanced artificial neural network design for improved industrial sensing under constraints of small, noisy, and nonlinear datasets.
Article Title: PLSaoNET: A Generalized ANN Model Under PLS Statistical Constraints for Industrial Sensing
L’intelligence artificielle, le calcul haute performance et la recherche scientifique reposent de plus en plus sur des infrastructures cloud accessibles à distance. Cette évolution facilite le travail collaboratif, mais elle augmente aussi les besoins en matière de sécurité. À l’occasion de...
In an era where education continually evolves through collaboration, the partnership between schools and universities represents a crucial frontier for innovation and professional growth. Yet, despite the increasingly common alliances forged between these educational institutions, evaluating their efficacy and sustaining them over time remains a complex challenge. Addressing this gap, a pioneering doctoral thesis from Tallinn University introduces a robust framework designed to monitor and enhance school–university partnerships systematically.
Edna Milena Sarmiento Marquez, PhD in Educational Sciences, has contributed a transformational approach to understanding how educational collaborations can thrive. She emphasizes that strong partnerships catalyze professional development for educators, enrich the practical application of research within classrooms, and ultimately, elevate student learning experiences. More than just administrative or academic joint ventures, these partnerships can ignite educational innovation, blending theoretical insights with lived classroom realities to the benefit of all stakeholders.
However, despite such promise, many partnerships falter due to a lack of clear indicators for progress, and the difficulty in sustaining engagement over extended periods. This represents a critical obstacle, as partnerships often involve diverse participants—teachers, researchers, administrators—each with differing expectations and needs. Without structured methods to assess their collaboration, these alliances risk stagnation or dissolution.
To navigate this intricate landscape, Sarmiento Marquez developed the SUP.ME (School–University Partnerships) framework. This innovative tool acts as both compass and map, guiding partners through the lifecycle of their joint ventures. SUP.ME moves beyond traditional evaluation metrics—such as goal achievement alone—and integrates qualitative measures like mutual trust, effective communication, and the partners’ capacity to learn collaboratively and adapt to evolving conditions. It transforms assessment into an ongoing, reflective learning process where each participant shares responsibility for monitoring progress.
The conceptual underpinning of SUP.ME is the recognition that partnerships function as dynamic ecosystems. They evolve through a continual exchange of knowledge and experiences among partners who bring complementary perspectives. This iterative process supports collective decision-making that is informed by real-time insights rather than distant or static assessments. By embedding evaluation within partnership activities, SUP.ME ensures that feedback loops remain immediate and actionable.
From a methodological standpoint, the framework demands active engagement with observational data, gathered through interactions, meetings, and cooperative initiatives within the partnership. Utilizing this data, partners collaboratively analyze outcomes and challenges, fostering an adaptive environment. Such an approach aligns with contemporary educational theories emphasizing reflective practice and experiential learning, which acknowledge that collaboration itself can be a transformative pedagogical tool.
Sarmiento Marquez’s research highlights that successful partnerships often manifest flexibility—adjusting goals, roles, and strategies as situations evolve. This adaptability is underpinned by the strength of interpersonal relationships and the development of trust. Partnerships that sustain themselves are not merely about completing tasks but about cultivating a shared vision and mutual respect that motivates continued engagement despite inevitable challenges.
This research adds a critical new dimension to partnership studies by situating evaluation as an integral, continuous component rather than a periodic, external audit. In doing so, it disrupts traditional views on educational collaboration and suggests that long-term sustainability emerges through mutual growth and learning. The framework’s practical application promises to provide schools and universities with a much-needed mechanism to quantify and qualify their joint efforts meaningfully.
Equally important is how SUP.ME emphasizes inclusivity and co-ownership of the evaluation process. It eschews hierarchical models where evaluation is imposed by one institution over the other. Instead, it encourages a democratic evaluation culture, promoting transparency and shared accountability. This shift resonates with broader educational trends aiming to democratize knowledge and empower practitioners at all levels.
The broader implications of this work extend beyond school–university partnerships. The principles of collaborative, embedded evaluation may influence how other cross-sectoral educational initiatives are designed and sustained, including community engagement projects and interdisciplinary research programs. By modeling how partnerships can tactically and effectively self-monitor, SUP.ME offers a blueprint for multifaceted educational transformation.
Reflecting on the metaphor provided by Sarmiento Marquez, managing a partnership without structured evaluation is akin to undertaking a trek through challenging terrain without navigational aids. The SUP.ME framework serves as that critical navigation system, enabling partners to map their journey, recognize landmarks of success, and recalibrate their route in response to unforeseen obstacles.
In essence, this doctoral research contributes not just a theoretical model but a functional toolkit aimed at bridging the often elusive gap between partnership aspirations and tangible outcomes. It empowers educational institutions to harness the full potential of collaboration, nurturing an environment where teachers, researchers, and students collectively thrive through shared knowledge, reflective practice, and adaptive strategy.
The significance of this work lies in its potential to reshape educational collaboration worldwide. As schools and universities confront unprecedented challenges—from digitization and shifting pedagogical paradigms to socio-economic disparities—the capacity to build and maintain vibrant partnerships becomes increasingly critical. SUP.ME offers a beacon of guidance for those navigating this demanding yet rewarding path of joint educational advancement.
By integrating observational data, ongoing reflective dialogue, and adaptive learning cycles, Sarmiento Marquez’s framework transcends traditional boundaries of partnership evaluation. It positions collaboration not merely as a strategic imperative but as a living, evolving practice anchored in trust, communication, and mutual growth. This perspective is not only innovative but essential for fostering the sustained success of educational partnerships in the 21st century.
Ultimately, the SUP.ME framework invites education professionals to embrace partnerships as reciprocal journeys of learning and innovation. It challenges institutions to reconceptualize evaluation from a retrospective checkpoint to a real-time compass guiding continuous improvement. As this approach gains traction, it promises to enhance educational ecosystems, ensuring that collaborations realize their full transformative potential.
Subject of Research: People Article Title: Available from DOI link Web References: http://dx.doi.org/10.60518/etera/148 Image Credits: Edna Milena Sarmiento Marquez and Tallinn University Keywords: Educational Partnerships, School–University Collaboration, SUP.ME Framework, Collaborative Learning, Evaluation Framework, Professional Development, Educational Innovation, Reflective Practice
Political correctness has been blamed for the failure of social workers to stop a gay couple murdering their adopted baby, with the child's grandmother and a social worker suggesting officials' judgement was clouded.
Elon Musk’s trillionaire milestone raises a bigger AI question: Will the future belong to the best models or the companies that own the infrastructure behind them?
Login
