搜索结果：RSC pharmaceutics

Spiking Neural Networks (SNNs) have received widespread attention due to their unique neuronal dynamics and low-power nature. Previous research empirically shows that SNNs with Poisson coding are more robust than Artificial Neural Networks (ANNs) on small-scale datasets. However, it is still unclear in theory how the adversarial robustness of SNNs is derived, and whether SNNs can still maintain its adversarial robustness advantage on large-scale dataset tasks. This work theoretically demonstrates that SNN's inherent adversarial robustness stems from its Poisson coding. We reveal the conceptual equivalence of Poisson coding and randomized smoothing in defense strategies, and analyze in depth the trade-off between accuracy and adversarial robustness in SNNs via the proposed Randomized Smoothing Coding (RSC) method. Experiments demonstrate that the proposed RSC-SNNs show remarkable adversarial robustness, surpassing ANNs and achieving state-of-the-art robustness results on large-scale dataset ImageNet. Our open-source implementation code is available at this https URL: https://github.com/KemingWu/RSC-SNN.

In recent years, there is an increasing interests in reconstruction based generative models for image One-Class Novelty Detection, most of which only focus on image-level information. While in this paper, we further exploit the latent space of Variational Auto-encoder (VAE), a typical reconstruction based model, and we innovatively divide it into three regions: Normal/Anomalous/Unknown-semantic-region. Based on this hypothesis, we propose a new VAE architecture, Recoding Semantic Consistency Based VAE (RSC-VAE), combining VAE with recoding mechanism and constraining the semantic consistency of two encodings. We come up with three training modes of RSC-VAE: 1. One-Class Training Mode, alleviating False Positive problem of normal samples; 2. Distributionally-Shifted Training Mode, alleviating False Negative problem of anomalous samples; 3. Extremely-Imbalanced Training Mode, introducing a small number of anomalous samples for training to enhance the second mode. The experimental results on multiple datasets demonstrate that our mechanism achieves state-of-the-art performance in various baselines including VAE.

Existing visual grounding benchmarks primarily evaluate alignment between image regions and literal referring expressions, where models can often succeed by matching a prominent named category. We explore a complementary and more challenging setting of scenario-based visual grounding, where the target must be inferred from roles, intentions, and relational context rather than explicit naming. We introduce Referring Scenario Comprehension (RSC), a benchmark designed for this setting. The queries in this benchmark are paragraph-length texts that describe object roles, user goals, and contextual cues, including deliberate references to distractor objects that often require deep understanding to resolve. Each instance is annotated with interpretable difficulty tags for uniqueness, clutter, size, overlap, and position which expose distinct failure modes and support fine-grained analysis. RSC contains approximately 31k training examples, 4k in-domain test examples, and a 3k out-of-distribution split with unseen object categories. We further propose ScenGround, a curriculum reasoning method serving as a reference point for this setting, combining supervised warm-starting with difficulty-a

RSC is an essential chromatin remodeling factor that is required for the control of several processes including transcription, repair and replication. The ability of RSC to relocate centrally positioned mononucleosomes at the end of nucleosomal DNA is firmly established, but the data on RSC action on oligo-nucleosomal templates remains still scarce. By using Atomic Force Microscopy (AFM) imaging, we have quantitatively studied the RSC- induced mobilization of positioned di- and trinucleosomes as well as the directionality of mobilization on mononucleosomal template labeled at one end with streptavidin. AFM imaging showed only a limited set of distinct configurational states for the remodeling products. No stepwise or preferred directionality of the nucleosome motion was observed. Analysis of the corresponding reaction pathways allows deciphering the mechanistic features of RSC-induced nucleosome relocation. The final outcome of RSC remodeling of oligosome templates is the packing of the nucleosomes at the edge of the template, providing large stretches of DNA depleted of nucleosomes. This feature of RSC may be used by the cell to overcome the barrier imposed by the presence of nucl

Herd immunity is shaped not only by the infection capacity of a spreading epidemic or the contact structure of the hosting population, but also by how and under what circumstances individuals acquire immunity. Immunization strategies may interact with ongoing non-pharmaceutical interventions, which commonly aim to reduce social contact numbers. We demonstrate that these interactions can induce unexpectedly strong and counterintuitive effects on herd immunity. We explore these phenomena on spatially embedded contact networks and uncover a reversal in the relative effectiveness of disease- versus vaccine-induced immunization schemes, highlighting the average number of contacts as a critical determinant of emerging herd immunity. In sparse geometric networks with limited degree heterogeneity, uniform vaccination proves most effective; however, as average contact numbers increase, naturally acquired immunity ultimately becomes the better strategy. We show that this phenomenon may emerge not only in synthetic networks but also in real-world mixing networks, observed during non-pharmaceutical intervention periods across multiple states of the United States.

Pharmaceutical supply chains (PSCs) struggle with inventory management (IM) due to unpredictable demand patterns and variable lead times associated with restocking. This complexity is further compounded by the finite shelf lives of pharmaceutical products, which necessitate a delicate balance between adequate stock and minimal waste. These intertwined factors create a complex optimization problem that requires sophisticated inventory strategies to ensure both product availability and PSC efficiency. This study aims to develop an optimal inventory replenishment policy for pharmaceutical products that can handle the stochasticity arising from uncertain demand and variable PSC conditions. The objective is to maximize the profitability of the PSC while maintaining a high patient service level. We formulate the problem as a Markov decision process and propose a deep reinforcement learning (DRL) approach, specifically, a hybrid asynchronous advantage actor critic distributed proximal policy optimization (A3C DPPO)algorithm. The A3C DPPO algorithm is tailored to handle the continuous action space inherent in IM. The numerical results demonstrate that the proposed algorithm adaptively upda

Research and Development is the largest budget position in the pharmaceutical industry, with clinical trials being a critical, yet costly and time-consuming component to inform decisions. Beyond drug efficacy, the probability of success and efficiency of research and development are highly dependent on the approaches used for designing, analyzing, and interpreting clinical trials. Deep understanding of statistical methodology and quantitative approaches is therefore essential. Consequently, dedicated methodology groups have emerged in mid-size and large pharmaceutical companies and CROs. Their remit is to lead the conception and implementation of innovative quantitative methodologies in order to improve drug development, often by addressing complexities or offering more efficient designs. To achieve this, they collaborate internally and externally (e.g., with academics, regulators) to identify common challenges and tear down silos in order to invest in methods with the highest impact on efficiency and value to the portfolio. Given the immense financial stakes of drug development -- where delays carry massive implications -- these groups represent a critical strategic investment. Ho

We present a Japanese domain-specific language model for the pharmaceutical field, developed through continual pretraining on 2 billion Japanese pharmaceutical tokens and 8 billion English biomedical tokens. To enable rigorous evaluation, we introduce three new benchmarks: YakugakuQA, based on national pharmacist licensing exams; NayoseQA, which tests cross-lingual synonym and terminology normalization; and SogoCheck, a novel task designed to assess consistency reasoning between paired statements. We evaluate our model against both open-source medical LLMs and commercial models, including GPT-4o. Results show that our domain-specific model outperforms existing open models and achieves competitive performance with commercial ones, particularly on terminology-heavy and knowledge-based tasks. Interestingly, even GPT-4o performs poorly on SogoCheck, suggesting that cross-sentence consistency reasoning remains an open challenge. Our benchmark suite offers a broader diagnostic lens for pharmaceutical NLP, covering factual recall, lexical variation, and logical consistency. This work demonstrates the feasibility of building practical, secure, and cost-effective language models for Japanes

In this work we present, for the first time, a computational fluid dynamics tool for the simulation of the metered discharge in a pressurized metered dose inhaler. The model, based on open-source software, adopts the Volume-Of-Fluid method for the representation of the multiphase flow inside the device and a cavitation model to explicitly account for the onset of flashboiling upon actuation. Experimental visualizations of the flow inside the device and measurements of the mixture density and liquid and vapor flow rates at the nozzle orifice are employed to validate the model and assess the sensitivity of numerical results to modeling parameters. The results obtained for a standard device geometry show that the model is able to quantitatively predict several aspects of the dynamics and thermodynamics of the metered discharge. We conclude by showing how, by allowing to reproduce and understand the fluid dynamics upstream of the atomizing nozzle, our computational tool enables systematic design and optimization of the actuator geometry.

Controlling the size of powder particles is pivotal in the design of many pharmaceutical forms and the related manufacturing processes and plants. One of the most common techniques for particle size reduction in process industry is powder milling, whose efficiency relates to the mechanical properties of powder particles themselves. In this work, we first characterize the elastic and plastic response of different pharmaceutical powders by measuring their Young modulus, the hardness and the brittleness index via nano-indentation. Subsequently, we analyze the behavior of those powder samples during comminution via jet-mill at different process conditions. Finally, the correlation between single particle mechanical properties and milling process results is illustrated; the possibility to build a predictive model for powder grindability, based on nano-indentation data, is critically discussed

Pharmaceutical research and development has accumulated vast and heterogeneous archives of data. Much of this knowledge stems from discontinued programs, and reusing these archives is invaluable for reverse translation. However, in practice, such reuse is often infeasible. In this work, we introduce DiscoVerse, a multi-agent co-scientist designed to support pharmaceutical research and development at Roche. Designed as a human-in-the-loop assistant, DiscoVerse enables domain-specific queries by delivering evidence-based answers: it retrieves relevant data, links across documents, summarises key findings and preserves institutional memory. We assess DiscoVerse through expert evaluation of source-linked outputs. Our evaluation spans a selected subset of 180 molecules from Roche's research and development repositories, encompassing over 0.87 billion BPE tokens and more than four decades of research. To our knowledge, this represents the first agentic framework to be systematically assessed on real pharmaceutical data for reverse translation, enabled by authorized access to confidential archives covering the full lifecycle of drug development. Our contributions include: role-specialized

The pharmaceutical supply chain faces escalating cybersecurity challenges threatening patient safety and operational continuity. This paper examines the transformative potential of zero trust architecture for enhancing security and resilience within this critical ecosystem. We explore the challenges posed by data breaches, counterfeiting, and disruptions and introduce the principles of continuous verification, least-privilege access, and data-centric security inherent in zero trust. Real-world case studies illustrate successful implementations. Benefits include heightened security, data protection, and adaptable resilience. As recognized by researchers and industrialists, a reliable drug tracing system is crucial for ensuring drug safety throughout the pharmaceutical production process. One of the most pivotal domains within the pharmaceutical industry and its associated supply chains where zero trust can be effectively implemented is in the management of narcotics, high-health-risk drugs, and abusable substances. By embracing zero trust, the pharmaceutical industry fortifies its supply chain against constantly changing cyber threats, ensuring the trustworthiness of critical medica

Many pharmaceutical companies face concerns with the maintenance of desired revenue levels. Sales forecasts for the current portfolio of products and projects may indicate a decline in revenue as the marketed products approach patent expiry. To counteract the potential downturn in revenue, and to establish revenue growth, an in-flow of new projects into the development phases is required. In this article, we devise an approach with which the in-flow of new projects could be optimized, while adhering to the objectives and constraints set on revenue targets, budget limitations and strategic considerations on the composition of the company's portfolio.

The training of graph neural networks (GNNs) is extremely time consuming because sparse graph-based operations are hard to be accelerated by hardware. Prior art explores trading off the computational precision to reduce the time complexity via sampling-based approximation. Based on the idea, previous works successfully accelerate the dense matrix based operations (e.g., convolution and linear) with negligible accuracy drop. However, unlike dense matrices, sparse matrices are stored in the irregular data format such that each row/column may have different number of non-zero entries. Thus, compared to the dense counterpart, approximating sparse operations has two unique challenges (1) we cannot directly control the efficiency of approximated sparse operation since the computation is only executed on non-zero entries; (2) sub-sampling sparse matrices is much more inefficient due to the irregular data format. To address the issues, our key idea is to control the accuracy-efficiency trade off by optimizing computation resource allocation layer-wisely and epoch-wisely. Specifically, for the first challenge, we customize the computation resource to different sparse operations, while limit

The survival of unorganized pharmacies is increasingly challenging in the face of growing competition from organized and e-pharmaceutical retail channels in emerging economies. A theoretical model is developed to capture the triple-channel interactions among unorganized, organized and e-retailing in emerging markets, taking into account the essential features of the pharmaceutical retail landscape, consumers, retailers and pharmaceutical products. Given the retailer and customer-specific factors, the price-setting game between the triple-channel retailers yielded the optimal prices for these retailers. The analysis found that the product category level demand has no influence on optimal pricing strategies of the retailers. The analysis also reveals counterintuitive results, for instance, (i) an increase in customer acceptance of unorganized retailers will result in a decrease in profits of both unorganized and organized retailers; (ii) as the distance and transportation cost to unorganized retailers increases for the consumers, the profit of the unorganized retailer increases; and (iii) consumers marginal utility of money has no influence on the optimal price, but have an influence

The present study considers the rural pharmaceutical retail sector in India, where the arrival of organized retailers and e-retailers is testing the survival strategies of unorganized retailers. Grounded in a field investigation of the Indian pharmaceutical retail sector, this study integrates primary data collection, consumer conjoint analysis and design of experiments to develop an empirically grounded agent-based simulation of multi-channel competition among unorganized, organized and e-pharmaceutical retailers. The results of the conjoint analysis reveal that store attributes of price discount, quality of products offered, variety of assortment, and degree of personalized service, and customer attributes of distance, degree of mobility, and degree of emergency are key determinants of optimal store choice strategies. The primary insight obtained from the agent-based modeling is that the attribute levels of each individual retailer have some effect on other retailers performance. The field-calibrated simulation also evidenced counterintuitive behavior that an increase in unorganized price discounts initially leads to an increase in average footprint at unorganized retailers, but

Data scarcity in pharmaceutical research has led to reliance on labour-intensive trial-and-error approaches for development rather than data-driven methods. While Machine Learning offers a solution, existing datasets are often small and noisy, limiting their utility. To address this, we developed a Variationally Encoded Conditional Tabular Generative Adversarial Network (VECT-GAN), a novel generative model specifically designed for augmenting small, noisy datasets. We introduce a pipeline where data is augmented before regression model development and demonstrate that this consistently and significantly improves performance over other state-of-the-art tabular generative models. We apply this pipeline across six pharmaceutical datasets, and highlight its real-world applicability by developing novel polymers with medically desirable mucoadhesive properties, which we made and experimentally characterised. Additionally, we pre-train the model on the ChEMBL database of drug-like molecules, leveraging knowledge distillation to enhance its generalisability, making it readily available for use on pharmaceutical datasets containing small molecules, an extremely common pharmaceutical task. W

We present PharmaShip, a real-world Chinese dataset of scanned pharmaceutical shipping documents designed to stress-test pre-trained text-layout models under noisy OCR and heterogeneous templates. PharmaShip covers three complementary tasks-sequence entity recognition (SER), relation extraction (RE), and reading order prediction (ROP)-and adopts an entity-centric evaluation protocol to minimize confounds across architectures. We benchmark five representative baselines spanning pixel-aware and geometry-aware families (LiLT, LayoutLMv3-base, GeoLayoutLM and their available RORE-enhanced variants), and standardize preprocessing, splits, and optimization. Experiments show that pixels and explicit geometry provide complementary inductive biases, yet neither alone is sufficient: injecting reading-order-oriented regularization consistently improves SER and EL and yields the most robust configuration, while longer positional coverage stabilizes late-page predictions and reduces truncation artifacts. ROP is accurate at the word level but challenging at the segment level, reflecting boundary ambiguity and long-range crossings. PharmaShip thus establishes a controlled, reproducible benchmark

We summarize the results on patterns of flavour violation in a Randall-Sundrum model with custodial protection (RSc) and compare them with those identified in the Littlest Higgs Model with T--parity (LHT) and in a number of SUSY Flavour Models. While K decays play in this presentation a prominent role, the inclusion of B physics and lepton flavour violation is crucial in the distinction between these three popular extensions of the Standard Model (SM) by means of flavour physics.