143174 research outputs found
Sort by
MagicGripper: a mini-magictac integrated gripper enabling multimodal perception in contact-rich manipulation
No full textContact-rich robotic manipulation in unstructured environments demands reliable multimodal perception. Here, we present MagicGripper, a multimodal robotic gripper built around mini-MagicTac, a compact variant of the MagicTac sensor. Mini-MagicTac embeds multi-layer grid structures in a 3D-printed elastomer, enabling visual, proximity, and tactile sensing in a gripper-compatible form factor. In this paper, we introduce the design and multimodal perception capabilities of mini-MagicTac, as well as two algorithmic frameworks for proximity and contact detection. Experimental evaluations show that mini-MagicTac achieves high spatial resolution, accurate contact localisation, and robust force estimation under mechanical and manufacturing variations. Autonomous grasping trials further validate MagicGripper’s reliable multimodal perception and adaptability to complex manipulation scenarios. These results demonstrate MagicGripper as a compact and versatile platform for embodied intelligence in contact-rich environments
Sexual dimorphism in benign adrenocortical tumours
No full textBenign adrenocortical tumours are the most common adrenal neoplasms. Evidence over the past few decades has highlighted sex differences in their prevalence, clinical characteristics, and treatment outcomes. Cortisol-producing adenomas causing either Cushing's syndrome, particularly those with PRKACA or GNAS somatic mutations associated with a more severe phenotype, or mild autonomous cortisol secretion (MACS) are more commonly observed in women. The mechanisms underpinning this sexual dimorphism remain to be fully elucidated. Studies in mice have revealed a protective role of androgens in males, leading to a decelerated growth rate of adrenocortical cells. Furthermore, evidence from human adrenal tumour tissue suggests that oestrogen, progesterone, and luteinising hormone/choriogonadotropin signalling in the adrenal cortex may play a role in adrenal tumourigenesis and steroid production. Clinically, this is supported by the increased incidence of cortisol-producing adrenocortical adenomas or nodular hyperplasia during puberty, pregnancy, and menopause. Notably, women with MACS seem to be more vulnerable to the harmful effects of cortisol excess and carry a higher mortality risk than men. Women with aldosterone-producing adenomas have a higher prevalence of somatic KCNJ5 mutations than men, and patients harbouring these mutations are likely to have more favourable clinical outcomes after adrenalectomy. In this review, we summarise the possible mechanisms behind the sexual dimorphism of benign adrenocortical tumours and provide an up-to-date overview of the sex-specific differences in their prevalence, clinical presentation, and outcomes, focusing on cortisol and aldosterone excess. Considering sexual dimorphism is crucial to guide diagnosis and management, and to counsel these patients for optimised care
Repurposing language models for FX volatility forecasting: a data-efficient and context-aware approach
No full textForecasting foreign exchange (FX) volatility is a critical task in finance. The high data demand of deep learning models creates a practical challenge, because data availability is limited by constantly shifting market dynamics. To address this problem, we introduce Vola-BERT, a novel predictive model that solves this data-efficiency problem by repurposing a pre-trained bidirectional language model. Our core insight is that the model’s value lies not in linguistic knowledge, but in its sophisticated capacity to model complex sequential patterns, which we adapt to the financial time-series domain. Vola-BERT is further enhanced by a new semantic conditioning framework that transparently integrates non-time-series data, such as high-impact economic events. This mechanism transforms the model from a black-box forecaster into an interpretable analytical tool capable of directly quantifying the impact of real-world market drivers. Our empirical study against 11 baseline models shows that Vola-BERT achieves state-of-the-art forecasting accuracy, consistently outperforming baselines under both data-rich and data-scarce conditions. This advantage is particularly pronounced in the data-scarce scenario, where it maintains high accuracy using only 10 percent of the training data. This work demonstrates that repurposing language models is a promising, data-efficient direction for context-aware numerical analysis
Cultural feasibility of conversational robots for dementia care in India: participatory design study
No full textBackground:
Dementia poses a significant challenge in India. The rising incidence rates, limited resources, and restricted clinician access have contributed to a staggering 90% gap in diagnosis and care. Conversational technology provides a natural user interface with the potential to promote the independence, well-being, and safety of people living with dementia at home. However, the feasibility of implementing such technology to support people living with dementia across diverse cultural and economic settings remains underexplored.
Objective:
This study aimed to assess the cultural feasibility of conversational robots for dementia care in India, a culturally underserved context in robotics and artificial intelligence (AI) for aging and dementia care.
Methods:
We involved 29 stakeholders, including people living with dementia, caregivers, and dementia care professionals. We evaluated (1) the engagement of people living with dementia with 3 conversational robots with varying interactive modalities (a voice agent, a virtual affective robot, and an embodied robot), (2) robot acceptance, and (3) stakeholder perspectives on the benefits and challenges of deploying conversational AI in India.
Results:
People living with dementia were willing to engage in verbal dialogue with conversational robots. Stakeholders perceived the technology as beneficial for supporting daily tasks at home, reducing loneliness, and enhancing cognitive function. We identified design adaptations to address feasibility challenges in India, including the need to (1) adapt interaction style to use a kind tone, appreciative language, and customizable facial expressions; (2) improve speech recognition for local accents interpretation and noisy settings; and (3) introduce prototypes in local clinics to promote familiarity.
Conclusions:
This work offers novel insights into cultural acceptance, human-robot engagement, and perceived utility for dementia care, along with key design implications for integrating conversational AI into care settings in India
Permuting accumulation order for low-precision machine learning
No full textQuantization of weights and activations in neural networks is widely used to reduce data movement and the computational footprint of multipliers in arithmetic units. However, this increases the relative area contribution of adders. Most recent work in neural network quantization uses large floating-point accumulators due the large rounding and clipping errors incurred by smaller accumulators even if weights and activations are otherwise quantized to narrow data types. In this work, we propose a novel method of finding and applying permutations to weight and activation order in neural networks to reduce the error induced by small floating-point adders for the multiply accumulate (MAC) functions in matrix multiplications. Our method optimizes the order of accumulation with very low computational overhead by using a shared ideal order for a group of vectors instead of an ideal order for each vector, and using a static order rather than dynamically generating one at runtime. Our technique does not require quantization-aware training (QAT) or modification of weights, making it applicable to large language models (LLMs)
High-pressure hydrogen effects on thermoplastics: a comprehensive review of permeation, decompression failure, and mechanical properties
No full textHydrogen energy is widely regarded as a clean and sustainable alternative to fossil fuel. Among various hydrogen storage options, high-pressure gas cylinders, especially Type IV composite cylinders, are increasingly used due to their light weight and high storage efficiency. Since the thermoplastic liner plays a significant role as a permeation barrier of the total cylinder, the current research findings and gaps related to its properties under high-pressure hydrogen environments are reviewed. Firstly, the potential thermoplastics and processing techniques of the liner are presented. Then, the review focuses on three key properties of thermoplastic liners including permeability, decompression failure and mechanical properties under high-pressure hydrogen environments. The mechanism and key influencing factors of these properties are systematically discussed, followed by the proposal of targeted and valuable improvement strategies. Moreover, testing and standards, quantification and physical models of these three properties are also outlined to provide guidance and reference for future research and applications. In the end, the research gaps and future perspectives related to the thermoplastic liner are identified. This review provides a valuable reference for the performance optimization and engineering application of the thermoplastic liners of Type IV cylinders
The impact of infrastructure on development outcomes: a meta-analysis
No full textWe perform a meta-analysis of the infrastructure research done over several decades, using over a thousand estimates from 221 papers reporting outcome elasticities. We include the transport, energy, and digital or information and communication technology (ICT) sectors, and the whole set of outcomes covered in the literature, including output, employment and wages, inequality and poverty, trade, education and health, population, and environmental aspects. Our results update the underlying parameters of interest, the “true” underlying infrastructure elasticities, accounting for publication bias, as well as for heterogeneity stemming from study design and context, with a particular focus on developing countries. Our results support a positive and significant impact of infrastructure on development outcomes, but also important heterogeneity across sectors and countries. While effects appear larger in developed countries for digital and energy studies, cross-sectoral and transport studies based on developing countries produce larger elasticities
Diffusion model-based generation of three-dimensional multiphase pore-scale images
No full textWe propose a diffusion model-based machine learning method for generating three-dimensional images of both the pore space of rocks and the fluid phases within it. This approach overcomes the limitations of current methods, which are restricted to generating only the pore space. Our reconstructed images accurately reproduce multiphase fluid pore-scale details in water-wet Bentheimer sandstone, matching experimental images in terms of two-point correlation, porosity, and fluid flow parameters. This method outperforms generative adversarial networks with a broader and more accurate parameter range. By enabling the generation of multiphase fluid pore-scale images of any size subject to computational constraints, this machine learning technique provides researchers with a powerful tool to understand fluid distribution and movement in porous materials without the need for costly experiments or complex simulations. This approach has wide-ranging potential applications, including carbon dioxide and underground hydrogen storage, the design of electrolyzers, and fuel cells
Levelised cost of demand response: estimating the cost-competitiveness of flexible demand
No full textTo make well-informed investment decisions, energy system stakeholders require reliable cost frameworks for demand response and storage technologies. While the levelised cost of storage permits comprehensive cost comparisons between different storage technologies, no generic cost measure for the comparison of different demand response schemes exists. This paper introduces the levelised cost of demand response, which is an analogous measure to the levelised cost of storage but crucially differs from it by considering consumer reward payments. Additionally, the value factor from cost estimations of variable renewable energy is adapted to account for the variable availability of demand response. The levelised cost of demand response is estimated for four direct load control schemes and twelve storage applications, and then contrasted against literature values for the levelised cost of the most competitive storage technologies. The direct load control schemes are vehicle-to-grid, smart charging, smart heat pumps, and heat pumps with thermal storage. The results show that only heat pumps with thermal storage consistently outcompete storage technologies, with EV-based schemes being competitive for some applications. The results and the underlying methodology offer a tool for energy system stakeholders to assess the competitiveness of demand response schemes even with limited user data
Tailoring composite hydrogel performance via controlled integration of norbornene-functionalised Pluronic micelles
No full textIncorporating micelles into polymeric hydrogels offers a powerful route to combine the tuneable mechanical and structural properties of hydrogels with the precise drug-
loading and release capabilities of nanocarriers. However, the method of micelle incorporation and its influence on hydrogel performance have yet to be studied in detail. Here, we present a modular strategy to tailor gelatin–norbornene hydrogels by integrating Pluronic® F127 micelles either physically or via covalent incorporation using norbornene-functionalised Pluronic (Pl_Nb). Pl_Nb was synthesised via Steglich esterification with >95% terminal functionalisation, forming stable, thermo-responsive micelles (2.5-15% w/v) with doxorubicin encapsulation efficiency of ~80%, comparable to unmodified Pluronic. Micelles were either physically entrapped or chemically integrated into gelatin–norbornene networks via bioorthogonal thiol–ene crosslinking. The incorporation route dictated network mechanics and dynamics: chemical crosslinking conferred temperature-dependent behaviour and enhanced stress relaxation compared to physical crosslinking, whereas both incorporation routes reduced stiffness relative to neat hydrogels and slowed drug release compared to direct loading. All hydrogels were cytocompatible, and the released doxorubicin retained its bioactivity, reducing cancer cell viability. These findings establish micelle–hydrogel coupling as a versatile design approach for engineering biomaterials with potential in controlled therapeutic delivery and regenerative medicine