Search

Natural intelligences (NIs) thrive in a dynamic world - they learn quickly, sometimes with only a few samples. In contrast, artificial intelligences (AIs) typically learn with a prohibitive number of training samples and computational power. What design principle difference between NI and AI could contribute to such a discrepancy? Here, we investigate the role of weight polarity: development processes initialize NIs with advantageous polarity configurations; as NIs grow and learn, synapse magnitudes update, yet polarities are largely kept unchanged. We demonstrate with simulation and image classification tasks that if weight polarities are adequately set a priori, then networks learn with less time and data. We also explicitly illustrate situations in which a priori setting the weight polarities is disadvantageous for networks. Our work illustrates the value of weight polarities from the perspective of statistical and computational efficiency during learning., ICML camera-ready

We present a statistical simulator, scDesign3, to generate realistic single-cell and spatial omics data, including various cell states, experimental designs, and feature modalities, by learning interpretable parameters from real data. Using a unified probabilistic model for single-cell and spatial omics data, scDesign3 infers biologically meaningful parameters; assesses the goodness-of-fit of inferred cell clusters, trajectories, and spatial locations; and generates in siliconegative and positive controls for benchmarking computational tools. This folder contains the preprocessed data and the code for analysis in this study. Scripts with "R" in their names are added during the review process.

Clustering filtering is usually a practical method for light detection and ranging (LiDAR) point clouds filtering according to their characteristic attributes. However, the amount of point cloud data is extremely large in practice, making it impossible to cluster point clouds data directly, and the filtering error is also too large. Moreover, many existing filtering algorithms have poor classification results in discontinuous terrain. This article proposes a new fast classification filtering algorithm based on density clustering, which can solve the problem of point clouds classification in discontinuous terrain. Based on the spatial density of LiDAR point clouds, also the features of the ground object point clouds and the terrain point clouds, the point clouds are clustered firstly by their elevations, and then the plane point clouds are selected. Thus the number of samples and feature dimensions of data are reduced. Using the DBSCAN clustering filtering method, the original point clouds are finally divided into noise point clouds, ground object point clouds, and terrain point clouds. The experiment uses 15 sets of data samples provided by the International Society for Photogrammetry and Remote Sensing (ISPRS), and the results of the proposed algorithm are compared with the other eight classical filtering algorithms. Quantitative and qualitative analysis shows that the proposed algorithm has good applicability in urban areas and rural areas, and is significantly better than other classic filtering algorithms in discontinuous terrain, with a total error of about 10%. The results show that the proposed method is feasible and can be used in different terrains.

Why do brains have inhibitory connections? Why do deep networks have negative weights? We believe representing functions is the primary role of both (i) the brain in natural intelligence, and (ii) deep networks in artificial intelligence. Our answer to why there are inhibitory/negative weights is: to learn more functions. We prove that, in the absence of negative weights, neural networks with non-decreasing activation functions are not universal approximators. While this may be an intuitive result to some, to the best of our knowledge, there is no formal theory, in either machine learning or neuroscience, that demonstrates why negative weights are crucial in the context of representation capacity. Further, we provide insights on the geometric properties of the representation space that non-negative deep networks cannot represent. We expect these insights will yield a deeper understanding of more sophisticated inductive priors imposed on the distribution of weights that lead to more efficient biological and machine learning.

Recommender systems are often susceptible to well-crafted fake profiles, leading to biased recommendations. The wide application of recommender systems makes studying the defense against attack necessary. Among existing defense methods, data-processing-based methods inevitably exclude normal samples, while model-based methods struggle to enjoy both generalization and robustness. Considering the above limitations, we suggest integrating data processing and robust model and propose a general framework, Triple Cooperative Defense (TCD), which cooperates to improve model robustness through the co-training of three models. Specifically, in each round of training, we sequentially use the high-confidence prediction ratings (consistent ratings) of any two models as auxiliary training data for the remaining model, and the three models cooperatively improve recommendation robustness. Notably, TCD adds pseudo label data instead of deleting abnormal data, which avoids the cleaning of normal data, and the cooperative training of the three models is also beneficial to model generalization. Through extensive experiments with five poisoning attacks on three real-world datasets, the results show that the robustness improvement of TCD significantly outperforms baselines. It is worth mentioning that TCD is also beneficial for model generalizations., Comment: 15 pages, 4 figures, 5 tables

Offshore jacket platform is widely used as production or oil recovering platform in the shallow sea, and is also applied to the offshore wind turbine supporting structure in the recent years. The jacket structures are normally designed to be conservative and bulky according to various design codes. In this work, a structural optimization design method for jacket platform structure has been developed based on topology optimization theory. The topology optimization method is applicable at an early design stage, which can determine the initial structure and force transmission path. The whole design space is chosen as design variables, and the goal is to maximize the structural stiffness. A set of constraints based on multi-criteria design assessment is applied according to standard requirements, which includes stress, deformation, vibration and design variable constraints. The optimization results are compared with the original platform for static performance, dynamic performance and Ultimate Carrying Capacity (UCC). Results show that the optimized structure show a 13.7% reduction in the global mass, 46.31% reduction in the maximum equivalent stress, and large ultimate carrying capacity ability under the same environmental loads. It is demonstrated that the proposed topology optimization method is capable of effectively determining the optimal design of jacket platform structures.

The effects of the addition of Pr6O11 nanopowder on the critical current density (Jc), irreversible field (Hirr) and upper critical field (Hc2) of the MgB2 bulks prepared by a Mg-diffusion method were investigated. The results show that the Pr6O11 nanopowder doping can significantly enhance the Jc, Hirr and Hc2 of the MgB2 bulks, but has little effect on the superconducting transition temperature (Tc). At 20 K in a self-field, the Jc value of the 1 wt% Pr6O11-doped MgB2 sample is 3.61×105 A/cm2, which is nearly 5 times larger than that of the undoped sample. In addition, at 10 K, the Hc2 and Hirr of the bulk increase by 1.9 and 2.6 T, respectively. The effect of Pr6O11-doping on superconductivity and flux pinning mechanism was also discussed.

In lifelong learning, data are used to improve performance not only on the current task, but also on previously encountered, and as yet unencountered tasks. In contrast, classical machine learning which starts from a blank slate, or tabula rasa, uses data only for the single task at hand. While typical transfer learning algorithms can improve performance on future tasks, their performance on prior tasks degrades upon learning new tasks (called forgetting). Many recent approaches for continual or lifelong learning have attempted to maintain performance on old tasks given new tasks. But striving to avoid forgetting sets the goal unnecessarily low. The goal of lifelong learning should be to not only improve performance on future tasks (forward transfer) but also to improve performance on past tasks (backward transfer) with any new data. Our key insight is that we can synergistically ensemble representations that were learned independently on disparate tasks to enable both forward and backward transfer. This generalizes ensembling independently learned representations (like in decision forests) and complements ensembling dependent representations (like in gradient boosted trees). Moreover, we ensemble representations in quasilinear space and time. We demonstrate this insight with two algorithms: representation ensembles of (1) trees and (2) networks. Both algorithms demonstrate forward and backward transfer in a variety of simulated and benchmark data scenarios, including tabular, image, spoken, and adversarial tasks, including CIFAR-100, 5-dataset, Split Mini-Imagenet, and Food1k, as well as the spoken digit dataset. This is in stark contrast to the reference algorithms we compared to, most of which failed to transfer either forward or backward, or both, despite that many of them require quadratic space or time complexity.

Polymer-air multilayer (PAM) was developed to decrease the heat loss through window glass panes. A PAM consists of a few polymer films separated from each other by air gaps. Thanks to the excellent optical properties of the polymer films, the visual transmittance of PAM is higher than 70%, and the haze is less than 2%. PAM not only has mechanisms to reduce the conductive and convective heat transfer, but also can obstruct the radiative heat transfer. With a 4~6 mm thick PAM coating, the U-factor of a glass pane can be lowered from above 1 Btu/{{h}{ft}^2{{\deg}F}} to 0.5~0.6 Btu/{{h}{ft}^2{{\deg}F}} .PAM is resilient and robust, relevant to the window retrofitting applications., Comment: 21 pages, 7 figures

We report the first example of a cobalt PCcarbeneP pincer complex (1) featuring a central alkylidene carbon donor accessed through the dehydration of an alcoholic POP proligand., We report the first example of a cobalt PCcarbeneP pincer complex (1) featuring a central alkylidene carbon donor accessed through the dehydration of an alcoholic POP proligand. Complex 1 shares bonding similarities with cobalt PBP and PNP pincer complexes where the donor atom engages in π-bonding with the cobalt centre, and thus completes the PXP (X = B, C, N) pincer ligand series for cobalt (for X donors that partake in M–L π-bonding). As compared to PBP and PNP pincer complexes, which are known to be good hydride and proton acceptors (respectively), complex 1 is found to be an effective hydrogen atom acceptor. Complex 1 partakes in cooperative ligand reactivity, engaging in several small molecule activations with styrene, bromine, carbon disulphide, phenyl acetylene, acetonitrile, hydrogen, benzaldehyde and water (through microreversibility). The mechanism for the formation of complex 1 is studied through the isolation and computational analysis of key intermediates. The formation of 1 is found to avoid C–H activation of the proligand, and instead proceeds through a combination of O–H activation, hydrogen atom transfer, β-hydride elimination and hydrogen activation processes.

High-entropy ceramics generally exhibit reduced thermal conductivity, but little is known about what controls this suppression and which descriptor can predict it. Herein, 18 medium- and high-entropy pyrochlores were synthesized to measure their thermal conductivity and Young's modulus. Up to 35% reductions in thermal conductivity were achieved with retained moduli, thereby attaining insulative yet stiff properties for potential thermal barrier coating applications. Notably, the measured thermal conductivity correlates well with a modified size disorder parameter. Thus, this modified size disorder parameter is suggested as a useful descriptor for designing thermally-insulative medium- and high-entropy ceramics (broadly defined as "compositionally-complex ceramics").

A 3-D numerical wave tank (NWT) is proposed in this paper to calculate the wave propagation and hydrodynamics force based on the incompressible Navier–Stokes equations by using the commercial computation fluid dynamics (CFD) software. The volume of fluid method is used to capture free surface. CFD code to generate waves in the NWT is first validated by analytical theory to check its accuracy. Next, a cylinder structure model is introduced to study the interaction of the structure with the wave load. The experiment included a cylinder structure, which was set up in a flume, to compare the wave load against numerical results. The performance of the numerical model is validated and verified by several cases, and it shows very good agreement with the experimental test.

Potential shortages in the reserves of phosphorus (P) rock and low soil P availability are among the main limiting factors for agricultural development. Recent studies show that biochar plays a crucial role in regulating soil physicochemical properties and crop development. To further verify the contribution of biochar to availability of soil P and nutrient efficiency of crop P, a maize continuous cropping field trial was started in 2013 to investigate the effects of biochar (0 and 2.63 t ha−1) combined with P fertilizer (0 and 26 kg ha−1) on soil biologically-based P and maize P use efficiency and yield. The treatments included mineral NPK fertilizer (NPK), mineral NK fertilizer (NK), mineral NPK fertilizer in combination with biochar (NPK + BC) and mineral NK fertilizer in combination with biochar (NK + BC). The following results were obtained in 2017 and 2018: biochar increased the pH, electrical conductivity (EC), total C (TC) and total P (TP) of the soil. Biochar can increase soil biologically-based P (CaCl2-P, enzyme-P, citrate-P and HCl-P). The grain yield of the NPK + BC treatment reached the highest point and improved maize grain yield by 4.2 %–12.5 % compared with NPK fertilization alone. Moreover, the application of biochar (2.63 t ha−1) instead of phosphate fertilizer in these experimental conditions did not lead to a decrease in the yield of maize. The addition of biochar increased the contents and uptake of P in maize grains, while negatively affecting the content and uptake of P in maize stover. Increases in P fertilizer use efficiencies (PUE) (12.4 %–31.4 %) and agronomic P efficiencies (AEP) (21.6 %–51.2 %) occurred after the application of biochar. The yield of maize and PUE positively correlated with the soil biologically available P. The structural equation modeling (SEM) results suggest that the increases in maize yield could be ascribed to improvements in soil biologically-based P and soil properties. Therefore, the addition of biochar is an effective approach to enhance the yield of maize, which is not only relevant to the improvement of soil biologically-based P but also increases in P use efficiency.

This study examines the causal relationship between the monetary policy and the stock price in Pakistan. We use bootstrap Granger full-sample causality test and sub-sample rolling window estimation test for the period of 1992:01 to 2015:06. The empirical result shows that there is a unidirectional causality from monetary policy to the stock price. However, considering the structural changes, we assess the parameter stability by using the full sample for short run relationship and find the short run instability for the full sample. This recommends that full-sample data cannot depend upon for the precise result. To address the structural changes issue, we propose rolling window approach is used to revisit the causal relationship between the two series and the result reports the bi-directional causal relationship between the monetary policy and the stock market. The relationship between the stock market and the money supply is not consistent, and the upward trend may cause serious policy implication on the stock market for future investment and profit.

Carbon and SiC doped MgB 2 bulks and wires were fabricated by a two-step reaction method. The MgB 4 samples were sintered at 900 °C for 2 h in closed argon atmosphere, and then an appropriate amount of Mg powder was mixed with the ground MgB 4 powder to get MgB 2 green compacts, which were sintered at 750 °C for 2 h to get bulk samples or at 680 °C for 2 h to get wire samples in closed argon atmosphere. Carbon or SiC doped MgB 2 bulk samples were also prepared by conventional solid state sintering for comparison. The phase evolution and the microstructure of MgB 2 samples were characterized by means of X-ray diffraction (XRD) and scanning electron microscope (SEM). The SEM results indicate that the compactedness and the dopant distribution homogeneity of MgB 2 samples prepared by the two-step reaction method are improved obviously. The carbon doping can more effectively introduce the dopants into MgB 2 crystalline lattice and grain boundaries compared with SiC doping.

Produce safety continues to be a challenge because produce undergoes minimal processing prior to consumption and existing sanitizers are not effective in inactivating pathogens. Novel decontamination technologies for produce are required as alternatives to traditional methods. In this project, two non-thermal process techniques were developed based on the interaction of UV light and gallic acid (GA) to enhance the safety of fresh produce. The first technique is the simultaneous application of UV-A light and GA (UVA+GA). UVA+GA treatment was effective against E. coli O157:H7, and the mechanism behind the synergistic antimicrobial effect was associated with the cellular uptake of GA, generation of reactive oxidative species (ROS), inactivation of enzymes superoxide dismutase, and damage to the bacterial membrane. In the second technique, the antimicrobial activity of GA was enhanced by its prior UV-C exposure (UVC-GA) against E. coli O157:H7 and was persistent for at least 4 weeks. The antimicrobial activity was affected by solution pH and the wavelength of UV-C exposure. The generation of ROS during UV light exposure and photo-oxidized compounds of GA such as quinone contributed to the antimicrobial activity of the UVC-GA solution. Both UVA+GA and UVC-GA treatments can enhance the inactivation of inoculated E. coli O157:H7 on produce such as spinach leaves and tomatoes without affecting the color and firmness. Common environmental stresses could confer complex cross-stress response in E. coli O157:H7 towards UVA+GA and UVC-GA treatments in that both resistance and sensitization can be induced depending on the stress applied and the technology studied. Repeated exposure to moderate UVA+GA or UVC-GA treatment can also select for sub-population that demonstrates higher resistance towards these treatments as well as cross-resistance to other lethal stress such as heat and acid. ROS scavenging enzymes and alternative sigma factor RpoS are highly likely to be associated with the adaptive response process. In conclusion, both UVA+GA and UVC-GA treatments are promising novel non-thermal techniques that are potential alternative methods for fresh produce disinfection. For future work, a better understanding of the inactivation mechanisms, optimizing of processing parameters, and the development of adaptive response associated with the two treatments need to be explored.

Traditional fault diagnosis of optical transmission system board mainly relies on experience, resulting in lack of common fault diagnosis method. In response to this, PQ1 board of OptiX2500+ optical transmission system is selected as a research object. After principle analyzing and fault tree analysis, a fault tree model is constructed, which can be used in fault diagnosis and maintenance. Several diagnosis examples are adopted to validate the fault tree. The result indicates that it is efficient, reliable and accurate.

The effects of Ti 3 SiC 2 doping on lattice parameter ( a ), microstructure, critical temperature ( T c ) and critical current density ( J c ) of MgB 2 superconductors were studied. Carbon substitution for boron could be confirmed by the shrinkage of the lattice parameter a with increasing of Ti 3 SiC 2 doping level. A reduction of T c from 37.15 K to 36.55 K with increasing of Ti 3 SiC 2 doping was observed. The J c values were determined by M-H hysteresis loops. Results show that in low magnetic field region, the J c value for the undoped sample is higher than those for the doped samples. However, with the increase of the magnetic field, the J c for the properly doped samples will be enhanced.

The phase composition and microstructure evolution of pure and carbon-doped MgB 2 bulks exposed to hydrochloric acid (pH=2) for different time were investigated by X-ray diffraction (XRD) and Scanning electron microscopy (SEM) systematically. XRD results show the decomposition of pure and carbon-doped MgB 2 bulks exposed to acid is rapid decay with exposure time. The decomposition reaction will finish completely after 5 h exposure and the main solid products are B(OH) 3 and MgCl 2 (6H 2 O). The SEM results demonstrate that the reaction among pure and carbon-doped MgB 2 as well as HCl occurs at grain boundaries. But the grain connection of carbon-doped MgB 2 is better than that of pure MgB 2 after acid exposure. The T c of MgB 2 decreases after 10 min acid exposure, but T c is almost unchanged in carbon-doped MgB 2 after the same exposure. This result indicates that substitution of C for boron in MgB 2 can improve of the resistance to the corrosion of acid.

Thermal treatment is the most widely used method of food preservation. However, it also results in losses in some quality due to undesirable chemical reactions. Sugars such as fructose in aqueous solution have been shown to accelerate the degradation of some food components such as ascorbic acid and anthocyanin (Cao et al., 2009; Rojas & Gerschenson, 2001), possibly due to oxidation. However, the mechanism of underlying this deleterious effect of sugars during thermal treatment is unclear. Previous studies have shown that monosaccharaides such as fructose and glucose can be autoxidized and generate reactive oxidative species (ROS) under physiological condition in the presence of oxygen and metal ions (P. Thornalley, Wolff, Crabbe, & Stern, 1984). However, information on ROS generation at elevated temperature is limited. The overall hypothesis for this project is to determine that sugars generate similar oxidative compounds at elevated temperatures. The objective of this study is to identify the generation of oxidative species during thermal treatment of sugars, and to understand the mechanism of the oxidative reactivity. To achieve this purpose, I measured the levels of oxidative species such as hydrogen peroxide and singlet oxygen generated from sugar solutions during thermal treatment. The effects of various factors such as types of sugars, fructose, glucose and sucrose; temperature, pH and the presence of phosphate buffer were investigated. To study the role of oxygen in oxidative species generation during thermal treatment of sugars, the consumption of dissolved oxygen by sugar solution during thermal treatment was also investigated. Further, acids generation during thermal treatment of sugar solution were measured and confirmed by pH measurement and acid-base titration. This study demonstrated that ROS are generated from sugar solutions during thermal treatment. The results suggested that dissolved oxygen in solution is crucial for ROS generation. Elevated temperature, lower pH and the presence of phosphate buffer showed a significant promotion effect on ROS generation from sugars during thermal treatment. The results from this study indicated that ROS generation from thermal treatment of sugars shared a similar mechanism as sugar autoxidation. We believe that the results of this study will provide methods to understand better the oxidation mechanism of sugar solutions and will be useful to improve the thermal treatment of food products containing sugars.

The MgB 2 wires were prepared by in-situ Powder-in-Tube (PIT) method with Nb/Cu compounds tube as the sheath and then were heat treated at various temperatures for 2 h in flowing argon. The resistivity ( R-T ), magnetic moment ( M-T ), transport critical current ( I c - B ) and magneto optical (MO) properties of MgB 2 wires were analyzed. The results show that there is a diffusion layer between MgB 2 superconducting core and Nb barrier layer when the heat-treatment temperature is higher than 750 °C, and this diffusion layer obstructs the current transfer into MgB 2 superconductor core from the Cu sheath. This indicates that the best heat-treatment temperature for the MgB 2 wires with Nb as the barrier layer should be lower than 750 °C.

The effects of amorphous carbon doping on superconducting properties of MgB 2 /Nb/Cu wires fabricated by the in-situ powder-in-tube (PIT) method have been investigated. Using Nb as the barrier and Cu as the stabilizer, the MgB 2 wires were fabricated by cold drawing. The wires of MgB 2- x C x ( x = 0.0, 0.05, 0.08, 0.10, 0.15) have been investigated to realize the effect of amorphous carbon doping on the phase formation, microstructure and superconducting properties. Characterization of MgB 2 /Nb/Cu wires was carried out using XRD, SQUID, SEM/EDS, and I c measurements. XRD result shows that the high purity MgB 2 phase is acquired after heat-treatment at 700 °C. The microstructure observation for wires shows that the MgB 2 /Nb/Cu wires have better grain connectivity after doping amorphous carbon, and Energy Dispersive X-ray (EDS) analysis confirms the uniform distribution of carbon in the MgB 2 superconducting phase. The critical current density ( J c ) of wires was measured in different magnetic fields by a standard four probes method, and the transport J c values as high as 1.4 × 10 5 A/cm 2 (4.2 K, 5 T) and 3.3 × 10 4 A/cm 2 (4.2 K, 10 T) have been achieved.

The mono-filamentary and multi-filamentary MgB2 wires sheathed with Nb/Cu composite tube were fabricated by in-situ powder-in-tube (PIT) technique with the carbon and TiC doping. The results show that there is a diffusion layer at the interfacial region between the Nb sheath and MgB2, which is non-superconducting phase for the sample heat-treated at high temperature, and this diffusion layer obstructs the current transmission at high heat-treatment temperatures and the best heat-treatment temperature for the MgB2 wires with Nb as the barrier layer should be lower than 750°C. Effects of amorphous carbon and TiC doped MgB2 wires were investigated by XRD, SEM and magnetization. The results show that substitution of B for C enhances the flux pinning but depresses Tc slightly. By controlling the processing parameters, an optimized Jc(H) performance is achieved under a partial amorphous C substitution. The Cu-clad 6-, 12-, and 36-filamentary MgB2 wires with Nb buffer layer also have been fabricated by the in-situ PIT method. To improve the strength of wires, the Nb-core was used as the central filament. The results show that the in-situ PIT technique can be used to produce the certain long length MgB2/Nb/Cu wires. The superconducting filaments of MgB2/Nb/Cu composite tubes have well workability suitable for the MgB2 wires long length and multi-filament.; Des conducteurs mono- et multi-filamentaires de MgB2 dans une gaine de Nb/Cu ont été élaborés par la technique PIT (powder in tube) avec des dopages de carbone et de TiC. Les résultats montrent qu'il y a une couche de diffusion non supraconductrice à l'interface entre le Nb et MgB2 pour les échantillons traités à haute température, couche qui empêche la pénétration du courant dans le conducteur. Les traitements thermiques doivent être inférieurs à 750°C. Les effets des dopages au carbone amorphe et au TiC ont été étudiés par XRD,MEB et aimantation. La substitution du bore par du carbone diminue légèrement la Tc mais augmente la piégeage des vortex, conduisant à un optimum du courant critique. Des multi-filaments de 6, 12 et 36 filaments sans dopage ont été élaborés par la technique PIT. Les propriétés mécaniques de ces conducteurs ont été renforcées en utilisant un filament central en Nb. L'assemblage MgB2/Nb/Cu est très adapté pour obtenir de grandes longueurs de conducteurs par la méthode PIT.

The mono-filamentary and multi-filamentary MgB2 wires sheathed with Nb/Cu composite tube were fabricated by in-situ powder-in-tube (PIT) technique with the carbon and TiC doping. The results show that there is a diffusion layer at the interfacial region between the Nb sheath and MgB2, which is non-superconducting phase for the sample heat-treated at high temperature, and this diffusion layer obstructs the current transmission at high heat-treatment temperatures and the best heat-treatment temperature for the MgB2 wires with Nb as the barrier layer should be lower than 750°C. Effects of amorphous carbon and TiC doped MgB2 wires were investigated by XRD, SEM and magnetization. The results show that substitution of B for C enhances the flux pinning but depresses Tc slightly. By controlling the processing parameters, an optimized Jc(H) performance is achieved under a partial amorphous C substitution. The Cu-clad 6-, 12-, and 36-filamentary MgB2 wires with Nb buffer layer also have been fabricated by the in-situ PIT method. To improve the strength of wires, the Nb-core was used as the central filament. The results show that the in-situ PIT technique can be used to produce the certain long length MgB2/Nb/Cu wires. The superconducting filaments of MgB2/Nb/Cu composite tubes have well workability suitable for the MgB2 wires long length and multi-filament., Des conducteurs mono- et multi-filamentaires de MgB2 dans une gaine de Nb/Cu ont été élaborés par la technique PIT (powder in tube) avec des dopages de carbone et de TiC. Les résultats montrent qu'il y a une couche de diffusion non supraconductrice à l'interface entre le Nb et MgB2 pour les échantillons traités à haute température, couche qui empêche la pénétration du courant dans le conducteur. Les traitements thermiques doivent être inférieurs à 750°C. Les effets des dopages au carbone amorphe et au TiC ont été étudiés par XRD,MEB et aimantation. La substitution du bore par du carbone diminue légèrement la Tc mais augmente la piégeage des vortex, conduisant à un optimum du courant critique. Des multi-filaments de 6, 12 et 36 filaments sans dopage ont été élaborés par la technique PIT. Les propriétés mécaniques de ces conducteurs ont été renforcées en utilisant un filament central en Nb. L'assemblage MgB2/Nb/Cu est très adapté pour obtenir de grandes longueurs de conducteurs par la méthode PIT.