Portail HAL UHA (Université de Haute-Alsace)
Not a member yet
17715 research outputs found
Sort by
French in Harmony? The Representation of Language in BnF fr. 1588
No full textRelecture par Jean-Yves BartInternational audienc
Optimized Scheduling for Electric Vehicle Charging: A Multi-Objective Approach to Grid Stability and User Satisfaction
No full textInternational audienc
Look into the LITE in deep learning for time series classification
No full textInternational audienceDeep learning models have been shown to be a powerful solution for Time Series Classification (TSC). State-of-the-art architectures, while producing promising results on the UCR and the UEA archives , present a high number of trainable parameters. This can lead to long training with high CO2 emission, power consumption and possible increase in the number of FLoating-point Operation Per Second (FLOPS). In this paper, we present a new architecture for TSC, the Light Inception with boosTing tEchnique (LITE) with only 2.34% of the number of parameters of the state-of-the-art InceptionTime model, while preserving performance. This architecture, with only 9, 814 trainable parameters due to the usage of DepthWise Separable Convolutions (DWSC), is boosted by three techniques: multiplexing, custom filters, and dilated convolution. The LITE architecture, trained on the UCR, is 2.78 times faster than InceptionTime and consumes 2.79 times less CO2 and power, while achieving an average accuracy of 84.62% compared to 84.91% with In-ceptionTime. To evaluate the performance of the proposed architecture on multivariate time series data, we adapt LITE to handle multivariate time series, we call this version LITEMV. To bring theory into application, we also conducted experiments using LITEMV on multivariate time series representing human rehabilitation movements, showing that LITEMV not only is the most efficient model but also the best performing for this application on the Kimore dataset, a skeleton based human rehabilitation exercises dataset. Moreover, to address the interpretability of LITEMV, we present a study using Class Activation Maps to understand the classification decision taken by the model during evaluation
Spatially controlled plasmon-induced photopolymerization at the apex of gold nanobipyramids: paving the way for 3D nanofabrication
No full textInternational audienceIn this study, we investigate near-field photopolymerization at the tips of gold nanobipyramids (AuNBPs) under linearly polarized irradiation conditions with femtosecond pulses in the near infrared region. TEM was used to visualize the localisation and spatial extend of the polymer resulting from the near-field photopolymerization. The single-tilt mode in TEM provides 3D information of the position of the nanoparticle on the substrate and the polymer lobes. In particular, we demonstrate that photopolymerization is localized mainly at the tips of AuNBPs, in line with the electromagnetic field distribution expected by simulations. We observed experimentally that when AuNBPs lie on the substrate via a facet, asymmetrical polymerization occurred: while polymerization is constrained by the substrate on the side in contact with the surface, we showed the possibility of fabricating a nano-sized polymer volume in 3D via near-field photopolymerization. These results illustrate the great control of photopolymerization at the nanoscale and represent a major step towards 3D fabrication at nanometer resolution
Two-Minute Onsite Detection of Agricultural Water Nitrogen Pollution with an Ammonia Gas Sensing System
No full textInternational audienceNitrogen pollution from agriculture is one of the major environmental problems. A high proportion of the fertilizers leached into the environment by runoff water, causing eutrophication and biodiversity damage. This study employed a semiconductor ammonia gas sensor to detect total ammonia nitrogen (TAN) pollution through fertilization as ammonia emissions correlate with the aqueous ammonium ion concentration. The sensor has a vertical conduction channel and a high sensitivity. The gas method is compared with the conventional indophenol method. It shows consistent results from the drinking water criterion of 0.5 mg/L to the acute criterion of 17 mg/L nitrogen pollution. The gas method is more convenient. For a given nitrogen dose, by comparison of two kinds of popular fertilizers, ammonium sulfate and urea, the lower ammonium concentration for urea is discovered. This gas sensor was applied to track the TAN runoff of the rice field to reveal that after base fertilization, there is an ammonium shock exceeding regulations. Therefore, real-time monitoring is necessary to assess the time-dependent environmental impacts and develop optimized fertilization schedules considering the cost of fertilization and the environment. The gas sensor is an ideal tool for these approaches to controlling nitrogen pollution
A polydihydroxystyrene (PDHS) mussel-mimetic binder: toward sustainable hard carbon electrodes with outstanding performance in sodium-ion batteries
No full textInternational audienceA fluorine-free binder (polydihydroxystyrene, PDHS) used in low amounts, enables sustainable production of hard carbon anodes for Na-ion batteries. High initial coulombic efficiency, excellent reversible capacity and superior cycling stability are reached compared to PVDF/CMC binders
Rethinking Wine through No-Low Perspectives
No full textThème du congrès : « Resilience and Transitions »International audienc
Binder‐free self‐standing hard carbon electrodes graphite‐coated by PE‐CVD to boost the efficiency of Na‐ion batteries
No full textInternational audienceThis work explores the use of plasma-enhanced chemical vapor deposition (PE-CVD) to deposit a C-coating on binder-free, selfstanding electrodes (SSEs). The C-coating consists of a nanocrystalline graphite thin film with crystalline domains measuring %14 nm. Regardless of the fabrication pyrolysis temperature of SSE, an increase in crystallite size and a reduction in interlayer space and defects is observed after C-coating and post-treatment at 1500 °C. Pyrolysis of the SSE at 900 °C induce better coverage with the nanographitic layer during PE-CVD , but prevents the development of closed pores during post-treatment at 1500 °C. In contrast, a large number of closed pores form when the SSE is pyrolyzed at 1500 °C. However, no difference in performance is observed between the C-coated SSE pyrolyzed at 900 and 1500 °C and post-annealed at 1500 °C, therefore, lower temperature can be advantageously used for the pyrolysis step. Nevertheless, post-treatment at 1500 °C is necessary to enhance performance. For both materials, the graphite coating minimized the undesirable reactions with the electrolyte, leading to more stable and conductive solid electrolyte interphase, which improves iCE (from 91.0% to 92.5%). A high reversible capacity of 320 mAh g À1 is also obtained, and the higher coating's conductivity is beneficial for rate capability
