20505 research outputs found
Sort by
Embedding ethics up front in AI and robotics: evidence from future engineers
No full textAs artificial intelligence and robotics increasingly shape societies, ensuring that these technologies align with ethical and societal values is a pressing challenge. This paper presents survey findings from 98 MSc Robotics and Applied AI students at Cranfield University, offering rare empirical evidence of how future AI and robotics professionals perceive their ethical responsibilities. While students demonstrate strong awareness of key risks such as autonomous decision-making in warfare, surveillance, labour displacement, and emotional manipulation, they show limited engagement with professional codes of ethics or structured training. Instead, ethical reflection often occurs informally, through peer discussions or media exposure. These findings highlight a consistent gap between ethical awareness and institutionalised engagement, raising questions about how future engineers will navigate the ethical challenges of AI. To address this, the paper proposes an “ethics up front” model for ethics integration that embeds reflection early in the development lifecycle, supported by participatory design, professional education, and regulatory alignment. This paper provides empirical evidence on future AI engineers’ ethical orientations and proposes a practical model for early-stage ethics integration into the practice of AI and robotics engineering.This research was funded by EPSRC (Engineering and Physical Sciences Research Council) and ISCF (Industry Strategy Challenge Fund) under the Made Smarter scheme No EP/V062158/1 and by the Horizon Europe project AI-PRISM, grant number 101058589.AI and Ethic
First evidence of SO2-releasing bags controlling fungal growth, aflatoxins and cyclopiazonic acid contamination in unshelled peanuts
No full textPeanuts (also known as groundnuts; Arachis hypogaea L.) are a globally significant cash crop but are highly susceptible to fungal contamination, particularly by Aspergillus section Flavi, which can contaminate the product with mycotoxins, including aflatoxins. This contamination poses serious food safety concerns, especially in low- and middle-income countries, limiting access to international markets. This study assessed the efficacy of slow-release SO2-bags compared to plastic bags (control) in reducing fungal growth and mycotoxin contamination in stored unshelled peanuts. Naturally contaminated and Aspergillus flavus-inoculated peanuts were stored at water activity (aw) levels of 0.85 and 0.95 for 15 days at 25°C. Fungal populations were monitored, aflatoxins and cyclopiazonic acid were analysed using LC-MS/MS qTRAP. SO2-releasing bags completely inhibited fungal growth in naturally contaminated peanuts and significantly reduced A. flavus population with effectiveness dependent on water activity. Additionally, SO2-releasing bags suppressed mycotoxin production. This study demonstrates, for the first time, the effectiveness of SO2-releasing bags in preventing fungal spoilage and mycotoxin contamination in post-harvest peanuts, offering a promising solution for enhancing peanut safety and quality.This work was supported by the EWA-BELT (Grant agreement No 862848) funded by Horizon 2020 European Union Funding for Research & Innovation and NutriNuts funded by Innovate UK (project reference 105663).Letters in Applied Microbiolog
What successful corporate venture capital funds do differently
No full textCorporate venture capital is experiencing renewed momentum. Over the past decade, as digital disruption and new technologies have accelerated, CVC activity has grown to record levels and has been rediscovered as a way for large companies to stay close to innovation at the edge of the firm. Today, many large companies are launching venture funds alongside traditional R&D to spot emerging technologies, build startup relationships, and open new growth opportunities. Yet a paradox persists: even as more CVCs are launched with ambition and visibility, many struggle to sustain momentum and quietly fade within a few years. For today’s senior leaders, the question is no longer whether to launch a CVC, but how to keep it alive long enough to matter. This article offers research-backed advice on what successful CVCs do differently.Harvard Business Revie
Physics-informed state observer for unknown linear autonomous systems with noisy measurements
No full textState estimation is a pivotal element in navigation tasks of autonomous vehicles. This technique is mainly applied when either a required measurement is not available or when the amount of available sensors in the platform are limited. Most of the state estimation algorithms available in on-board control modules use kinematic models as prior model to estimate the states of the autonomous system. However, these simple kinematic models do not consider dynamic terms and physical properties which can lead to biased state estimates. To overcome this issue, this paper proposes a physics-informed state observer for unknown linear systems under partial and noisy measurements. The proposed approach fuses two complementary concepts for state estimation and dynamics identification. The proposed approach is capable to obtain reliable state estimates whilst attenuating the level of noise. Lyapunov stability is used to derive an appropriate update law for the construction of physics-informed estimate model. Simulation studies are given to show the advantages and challenges of the proposed approach.2025 11th International Conference on Control, Decision and Information Technologies (CoDIT
Long-term operational strategy of microalgal-bacterial granular sludge towards strengthened sulfadiazine antimicrobial resistance control
No full textHarnessing the dynamic interactions within the microalgal-bacterial granular sludge (MBGS) systems represents an emerging biotechnological frontier in tackling antibiotic pollution. Nevertheless, effective long-term treatment targeting strengthened antimicrobial resistance remains challenging. This study explored the operational strategies of pre-accumulation with antibiotic-free cultivation and pre-acclimatization with 5 mg/L sulfadiazine (SDZ) exposure for long-term, effective high-concentration SDZ removal (stepwise increase up to 50 mg/L), while controlling antibiotic resistance genes. The results showed that pre-accumulation achieved a stable microbial structure and greater antibiotic removal (up to 99.9 %) under extremely high SDZ concentration (50 mg/L) compared to pre-acclimatization, while pre-acclimatization suffered more oxidative damage, leading to lower biomass and pigments accumulation, but higher SDZ removal under 30 mg/L. The amide bond hydrolysis was identified as the predominant degradation pathway, resulting in less toxic by-products. Pre-accumulation performed better in controlling relative abundance of target sulfonamide antibiotic resistance genes (ARGs, sul1 and sul2) compared to pre-acclimatization. Genes encoding metabolic pathways related to antibiotic removal remained active, with downregulated genes encoding growth. Chlorella exhibited a protective effect on bacteria and inhibited ARGs spread, whereas Nakamurella and Nitrosomonas might be the key bacteria in SDZ biodegradation, but also served as potential hosts of ARGs. The continuous and efficient removal of SDZ, along with the mitigation of ARGs spread and toxic by-product accumulation under ultra-high antibiotic exposure, can be attributed to the MBGS system's initial operational strategy. The initial operational strategy involving stress pre-acclimation and biomass pre-accumulation offers a robust and adaptable solution for treating antibiotic-containing wastewater.The present work was funded by Jinhua Science and Technology Program (2022–3–066). This work is also supported by the National Scholarship Fund.Journal of Hazardous Material
Software: Stochastic Prediction of Oil Spill Transport and Fate using Ap-proximation Methods or Machine Learning
No full textOil spills represent a persistent risk to marine ecosystems, coastal communities, and energy-related maritime operations, demanding predictive tools that areboth accurate and computationally efficient for real-time decision support. This paper presents CRANSLIK 3.2, a machine-learning-driven oil spill trajectoryforecasting system for the Mediterranean Sea that significantly advances CRANSLIK 3.1's capabilities through targeted optimisation and architecturalrefinement of deep neural networks. The system is validated against the established MEDSLIK-II model using a documented real-world oil spill event off thecoast of Algeria, demonstrating reliable operational performance. Key innovations include the application of the Levenberg–Marquardt optimisation al-gorithm and a comprehensive evaluation of Long Short-Term Memory (LSTM) network architectures, in which fourteen activation function combinations weresystematically tested. An LSTM configuration combining Exponential Linear Unit (ELU) activation with Sigmoid gating functions achieved the highestpredictive accuracy while pre-serving rapid inference times. The results highlight the ability of data-driven models to complement physics-based approaches,offering a robust, scalable, and time-critical forecasting tool for environmental protection and energy-sector risk mitigation.Python 3
Enhancing 3D Scan Quality and Automating Data Processing for Equine Back Analysis in Saddle Fitting Applications
No full textJulienne, Aurélie - Industrial Supervisor - Voltaire Group
Renard, Laurent - Industrial Supervisor - Voltaire GroupPoor saddle fit can cause pain, muscle atrophy, and reduced performance in horses, while back shape changes over time make regular reassessment essential. To address this, Voltaire Group developed a mobile app using iPhone 3D scanning, but the current system struggles with scan accuracy, robustness, and extraction of meaningful geometric features. This research improves the reconstruction and analysis pipeline to enable more precise and reliable saddle fitting. The original workflow relied on surfel-based reconstruction followed by Poisson meshing, which often produced imprecision and inconsistencies due to low-resolution depth data. In contrast, the proposed workflow filters the raw depth data and applies a volumetric Truncated Signed Distance Function (TSDF) representation combined with marching cubes meshing. Comparative analyses were conducted between methods, with evaluation metrics focused on reconstruction precision relative to reference scans. In addition, a novel technique for detecting the dominant symmetry plane—tailored to horse back geometry—was introduced, along with newgeometric measurements relevant to saddle fit. Results show that TSDF-based reconstruction increases precision by up to 30% while supporting real-time processing. The new pipeline also avoids the error accumulation observed in the previous point cloud–to–mesh workflow, ensuring consistent and reliable outcomes across configurations. The symmetry plane detection method proved highly effective, reducing orientation errors by an average of 2 cm. Overall, this work contributes a robust and precise 3D reconstruction pipeline for equine back scanning, supporting improved saddle fitting practices and demonstrating the potential of mobile 3D scanning in equestrian applications.MSc in Computational and Software Techniques in Engineerin
“We are not like them”. The generational shift in leadership perceptions in an emerging market
No full textThis study examines the evolving leadership identity among a new generation of Russian business leaders, challenging the temporal stability of static cross-cultural frameworks. Based on qualitative interviews with 30 MBA graduates in Moscow, Russia, the findings propose a model of hybrid leadership characterized by confidence, result-driven inspiration, relational care and ingenious concept of otvetstvennost (ownership and moral responsibility). The research reveals a significant shift in leadership perceptions, where emerging leaders favor participative, humane-oriented, and transformational leadership styles over traditional authoritarian approaches. However, this transition is marked by a sociolinguistic tension: a persistent reluctance to internalize the term “leader” despite demonstrating leadership behaviors. This generational shift is theorized as an identity struggle between entrenched national prototypes and global managerial exemplars. The implications suggest that leadership development in transitional economies functions as an “identity laboratory,” requiring programs that reconcile local cultural nuances with evolving global standards.International Journal of Cross Cultural Managemen
Inferring wind velocity from informal environmental objects using optical flow informed recurrent neural networks
No full textDue to their flexibility and wide range of applications, UAVs are expected to play an important role in complex urban airspace in the future. However, unpredictable low-level air currents caused by the complexity and variability of local urban design can pose significant risks to the take-off and landing phases. Current high-quality wind profile radars are expensive and only deployed in major airports. The alternative is to conduct large-scale urban modelling of wind using computation fluid dynamics, which relies on a large volume of accurate city and wind profile data. This undermines the future business model of distributed air mobility, e.g., takeoff and land in ad-hoc locations across a city. Therefore, it is crucial to create an approach that is data-efficient and economical. To achieve this, we leverage the abundance of environmental objects that naturally interact with wind, such as trees, flags, and clothing. This initial pilot study aims to address this challenge by examining tree movement using two consecutive techniques: (1) optical flow to extract the natural movement vectors, and (2) deep recurrent neural networks to translate the vectors into wind velocity. The proposed CNN-ConvLSTM model, trained on a video dataset encompassing diverse environmental conditions with ground wind speeds from 0 to 14.6 m/s, extracted visual and motion features from RGB and optical flow images, achieving an 87.42% prediction accuracy in capturing spatiotemporal wind-induced motion patterns. These results suggest the possibility of extending visual anemometer technology to broader scenarios and diverse natural objects, guaranteeing safer UAV operation in complex environments.2025 11th International Conference on Control, Decision and Information Technologies (CoDIT
Exploring artificial intelligence for third-party logistics service providers: a dynamic capabilities perspective
No full textPurpose:
Artificial Intelligence (AI) is increasingly considered a transformative force within the logistics industry, improving efficiency and enhancing effectiveness for many organisations, including third-party logistics service providers (3PLs). However, the academic literature reveals a limited understanding of how 3PLs can approach the opportunities offered by AI. To fill this gap, we leverage the dynamic capabilities theory to explore AI adoption within the 3PL industry.
Design/methodology/approach:
We developed a single case study focusing on a leading British 3PL that introduced AI to improve warehousing operations and planning activities. We collected data through two rounds of qualitative interviews with managerial-level stakeholders from different departments and two on-site visits. Through abductive reasoning, we iteratively compared empirics with the available theoretical knowledge to illuminate how 3PLs can approach AI opportunities through their sensing, seizing and reconfiguring dynamic capabilities.
Findings:
Findings illustrate several micro-foundations underpinning higher-order dynamic capabilities. Sensing AI opportunities critically depends on building internal AI awareness as well as involving customers to embed their perspectives, leading to prioritising AI use cases. Seizing starts with aligning use cases with the business strategy, then procuring AI solutions and assessing their security and ethical implications before embedding different AI tools. These initiatives foster resource reconfiguration by embracing a cultural shift involving 3PLs and their customers and developing a robust data infrastructure to support AI efforts. Building on these findings, we suggest evolutionary patterns for dynamic capabilities through AI adoption.
Originality/value:
Existing research has yet to fully explore how 3PLs can approach AI adoption. The study contextualises the dynamic capabilities theory for AI-driven opportunities, elaborating on earlier studies to identify micro-foundations for 3PLs' higher-order dynamic capabilities. It proposes a set of research propositions and offers a research agenda to foster future exploration about embedding AI into logistics operations. By focusing on 3PLs in the context of rising digitalisation, the study highlights how firms can navigate the complexities of AI adoption, offering original insights to leverage the synergies among human workforce, technological tools and physical assets.International Journal of Physical Distribution & Logistics Managemen