Related Articles
Dynamic optimizers for complex industrial systems via direct data-driven synthesis
The chemical process industry (CPI) faces significant challenges in improving sustainability and efficiency while maintaining conservative principles for managing cost, complexity, and uncertainty. This work introduces a data-driven approach to dynamic real-time optimization (D-RTO) that addresses the aforementioned concerns by directly extracting process optimization policies from historical plant data. Our method constructs a value function to evaluate trajectory quality and employs weighted regression to derive improved policies. When applied to a plant-wide industrial process control problem, the proposed optimizer demonstrates superior performance in adapting to disturbances while maintaining stability and product quality. These results challenge conventional assumptions regarding the potential of data-driven optimization in the CPI. Although limitations exist due to the black-box nature of neural networks, this study presents a promising avenue for enhancing operational efficiency in industrial settings. The proposed approach offers a practical solution for process optimization, as it leverages readily available historical data and does not require extensive modeling efforts. By demonstrating significant efficiency improvement on a realistic industrial benchmark problem, this work paves the way for the adoption of data-driven optimization techniques in real-world CPI applications.
A data-driven generative strategy to avoid reward hacking in multi-objective molecular design
Molecular design using data-driven generative models has emerged as a promising technology, impacting various fields such as drug discovery and the development of functional materials. However, this approach is often susceptible to optimization failure due to reward hacking, where prediction models fail to extrapolate, i.e., fail to accurately predict properties for designed molecules that considerably deviate from the training data. While methods for estimating prediction reliability, such as the applicability domain (AD), have been used for mitigating reward hacking, multi-objective optimization makes it challenging. The difficulty arises from the need to determine in advance whether the multiple ADs with some reliability levels overlap in chemical space, and to appropriately adjust the reliability levels for each property prediction. Herein, we propose a reliable design framework to perform multi-objective optimization using generative models while preventing reward hacking. To demonstrate the effectiveness of the proposed framework, we designed candidates for anticancer drugs as a typical example of multi-objective optimization. We successfully designed molecules with high predicted values and reliabilities, including an approved drug. In addition, the reliability levels can be automatically adjusted according to the property prioritization specified by the user without any detailed settings.
Macroevolution along developmental lines of least resistance in fly wings
Evolutionary change requires genetic variation, and a reigning paradigm in biology is that rates of microevolution can be predicted from estimates of available genetic variation within populations. However, the accuracy of such predictions should decay on longer evolutionary timescales, as the influence of genetic constraints diminishes. Here we show that intrinsic developmental variability and standing genetic variation in wing shape in two distantly related flies, Drosophila melanogaster and Sepsis punctum, are aligned and predict deep divergence in the dipteran phylogeny, spanning >900 taxa and 185 million years. This alignment cannot be easily explained by constraint hypotheses unless most of the quantified standing genetic variation is associated with deleterious side effects and is effectively unusable for evolution. However, phenotyping of 71 genetic lines of S. punctum revealed no covariation between wing shape and fitness, lending no support to this hypothesis. We also find little evidence for genetic constraints on the pace of wing shape evolution along the dipteran phylogeny. Instead, correlational selection related to allometric scaling, simultaneously shaping developmental variability and deep divergence in fly wings, emerges as a potential explanation for the observed alignment. This suggests that pervasive natural selection has the potential to shape developmental architectures of some morphological characters such that their intrinsic variability predicts their long-term evolution.
Influenza A virus rapidly adapts particle shape to environmental pressures
Enveloped viruses such as influenza A virus (IAV) often produce a mixture of virion shapes, ranging from 100 nm spheres to micron-long filaments. Spherical virions use fewer resources, while filamentous virions resist cell-entry pressures such as antibodies. While shape changes are believed to require genetic adaptation, the mechanisms of how viral mutations alter shape remain unclear. Here we find that IAV dynamically adjusts its shape distribution in response to environmental pressures. We developed a quantitative flow virometry assay to measure the shape of viral particles under various infection conditions (such as multiplicity, replication inhibition and antibody treatment) while using different combinations of IAV strains and cell lines. We show that IAV rapidly tunes its shape distribution towards spheres under optimal conditions but favours filaments under attenuation. Our work demonstrates that this phenotypic flexibility allows IAV to rapidly respond to environmental pressures in a way that provides dynamic adaptation potential in changing surroundings.
Bicomponent nano- and microfiber aerogels for effective management of junctional hemorrhage
Managing junctional hemorrhage is challenging due to ineffective existing techniques, with the groin being the most common site, accounting for approximately 19.2% of potentially survivable field deaths. Here, we report a bicomponent nano- and microfiber aerogel (NMA) for injection into deep, narrow junctional wounds to effectively halt bleeding. The aerogel comprises intertwined poly(lactic acid) nanofibers and poly(ε-caprolactone) microfibers, with mechanical properties tunable through crosslinking. Optimized aerogels demonstrate improved resilience, toughness, and elasticity, enabling rapid re-expansion upon blood contact. They demonstrate superior blood absorption and clotting efficacy compared to commercial products (i.e., QuikClot® Combat Gauze and XStat®). Most importantly, in a lethal swine junctional wound model (Yorkshire swine, both male and female, n = 5), aerogel treatment achieved immediate hemostasis, a 100% survival rate, no rebleeding, hemodynamic stability, and stable coagulation, hematologic, and arterial blood gas testing.
Responses