Target Application Research Articles

Synthetic cells in tissue engineering.

Tissue functions rely on complex structural, biochemical, and biomechanical cues that guide cellular behavior and organization. Synthetic cells, a promising new class of biomaterials, hold significant potential for mimicking these tissue properties using simplified, nonliving building blocks. Advanced synthetic cell models have already shown utility in biotechnology and immunology, including applications in cancer targeting and antigen presentation. Recent bottom-up approaches have also enabled synthetic cells to assemble into 3D structures with controlled intercellular interactions, creating tissue-like architectures. Despite these advancements, challenges remain in replicating multicellular behaviors and dynamic mechanical environments. Here, we review recent advancements in synthetic cell-based tissue formation and introduce a three-pillar framework to streamline the development of synthetic tissues. This approach, focusing on synthetic extracellular matrix integration, synthetic cell self-organization, and adaptive biomechanics, could enable scalable synthetic tissues engineering for regenerative medicine and drug development.

Advances in Drug Targeting, Drug Delivery, and Nanotechnology Applications: Therapeutic Significance in Cancer Treatment.

In the 21st century, thanks to advances in biotechnology and developing pharmaceutical technology, significant progress is being made in effective drug design. Drug targeting aims to ensure that the drug acts only in the pathological area; it is defined as the ability to accumulate selectively and quantitatively in the target tissue or organ, regardless of the chemical structure of the active drug substance and the method of administration. With drug targeting, conventional, biotechnological and gene-derived drugs target the body's organs, tissues, and cells that can be selectively transported to specific regions. These systems serve as drug carriers and regulate the timing of release. Despite having many advantageous features, these systems have limitations in thoroughly treating complex diseases such as cancer. Therefore, combining these systems with nanoparticle technologies is imperative to treat cancer at both local and systemic levels effectively. The nanocarrier-based drug delivery method involves encapsulating target-specific drug molecules into polymeric or vesicular systems. Various drug delivery systems (DDS) were investigated and discussed in this review article. The first part discussed active and passive delivery systems, hydrogels, thermoplastics, microdevices and transdermal-based drug delivery systems. The second part discussed drug carrier systems in nanobiotechnology (carbon nanotubes, nanoparticles, coated, pegylated, solid lipid nanoparticles and smart polymeric nanogels). In the third part, drug targeting advantages were discussed, and finally, market research of commercial drugs used in cancer nanotechnological approaches was included.

Hyaluronic Acid-Based Hybrid Nanoparticles as Promising Carriers for the Intranasal Administration of Dimethyl Fumarate.

Dimethyl fumarate (DMF), the first-line oral therapy for relapsing-remitting multiple sclerosis, is rapidly metabolized into monomethyl fumarate. The DMF oral administration provokes gastrointestinal discomfort causing treatment withdrawal. The present study aimed to develop an innovative formulation for DMF nasal administration. Lipid-polymer hybrid nanoparticles (LPNs) were developed to improve DMF stability, limiting gastrointestinal side effects and increasing brain bioavailability by nose-to-brain targeting application. DMF-loaded and unloaded LPNs with or without hyaluronic acid (HA) were prepared using the nanoprecipitation via magnetic/mechanical stirring technique. Particle morphology and surface properties were evaluated; drug content, viscosity, and mucoadhesion were determined. Physico-chemical stability of LPNs and DMF in the LPNs was also explored. In vitro DMF permeation experiments were performed utilizing the PermeaPad®. The cytotoxicity and cellular uptake studies were performed using RPMI 2650 and SK-N-BE2 cell lines. DMF nose-to-brain delivery was evaluated by intranasally administering DMF-loaded LPNs to rats. LPNs with average sizes of 120-250 nm and a negative zeta potential -17.3 to -43 mV were obtained, primarily influenced by the presence of HA. HA assured drug stability up to 60 days and promoting the in vitro permeation of DMF compared to the free-DMF. HA greatly improved the viscosity and mucoadhesive properties. LPNs with and without HA did not exhibit any cytotoxicity and showed a rapid cell uptake starting from 15min to 2h with a best internalization after 1h of treatment in both epithelial and neuronal cell lines. Nasal administration of DMF-loaded LPNs allowed to quantify up to about 12μg/mL of DMF in the rat cerebrospinal fluid. The results highlight the role of HA in improving LPNs properties and performance as carrier of DMF for nasal administration. In particular, LPNs appear able to enter neurons and monolayers of epithelial cells, allowing to promote the nose-to-brain DMF delivery.

Automatic Prediction of Molecular Properties Using Substructure Vector Embeddings within a Feature Selection Workflow.

Machine learning (ML) methods provide a pathway to accurately predict molecular properties, leveraging patterns derived from structure-property relationships within materials databases. This approach holds significant importance in drug discovery and materials design, where the rapid, efficient screening of molecules can accelerate the development of new pharmaceuticals and chemical materials for highly specialized target application. Unsupervised and self-supervised learning methods applied to graph-based or geometric models have garnered considerable traction. More recently, transformer-based language models have emerged as powerful tools. Nevertheless, their application entails considerable computational resources, owing to the need for an extensive pretraining process on a vast corpus of unlabeled chemical data sets. To this end, we present a semisupervised strategy that harnesses substructure vector embeddings in conjunction with a ML-based feature selection workflow to predict various molecular and drug properties. We evaluate the efficacy of our modeling methodology across a diverse range of data sets, encompassing both regression and classification tasks. Our findings demonstrate superior performance compared to most existing state-of-the-art algorithms, while offering advantages in terms of balancing model accuracy with computational requirements. Moreover, our approach provides deeper insights into feature interactions that are essential for model interpretability. A case study is conducted to predict the lipophilicity of chemical molecules, exemplifying the robustness of our strategy. The result underscores the importance of meticulous feature analysis and selection over a mere reliance on predictive modeling with a high degree of algorithmic complexity.

Unsupervised Temporal Adaptation in Skeleton-Based Human Action Recognition

With deep learning approaches, the fundamental assumption of data availability can be severely compromised when a model trained on a source domain is transposed to a target application domain where data are unlabeled, making supervised fine-tuning mostly impossible. To overcome this limitation, the present work introduces an unsupervised temporal-domain adaptation framework for human action recognition from skeleton-based data that combines Contrastive Prototype Learning (CPL) and Temporal Adaptation Modeling (TAM), with the aim of transferring the knowledge learned from a source domain to an unlabeled target domain. The CPL strategy, inspired by recent success in contrastive learning applied to skeleton data, learns a compact temporal representation from the source domain, from which the TAM strategy leverages the capacity for self-training to adapt the representation to a target application domain using pseudo-labels. The research demonstrates that simultaneously solving CPL and TAM effectively enables the training of a generalizable human action recognition model that is adaptive to both domains and overcomes the requirement of a large volume of labeled skeleton data in the target domain. Experiments are conducted on multiple large-scale human action recognition datasets such as NTU RGB+D, PKU MMD, and Northwestern–UCLA to comprehensively evaluate the effectiveness of the proposed method.

Development of Piezoelectric Inertial Rotary Motor for Free-Space Optical Communication Systems.

This paper presents the design, development, and investigation of a novel piezoelectric inertial motor whose target application is the low Earth orbit (LEO) temperature conditions. The motor utilizes the inertial stick-slip principle, driven by the first bending mode of three piezoelectric bimorph plates, and is compact and lightweight, with a total volume of 443 cm3 and a mass of 28.14 g. Numerical simulations and experimental investigations were conducted to assess the mechanical and electromechanical performance of the motor in a temperature range from -20 °C to 40 °C. The results show that the motor's resonant frequency decreases from 12,810 Hz at -20 °C to 12,640 Hz at 40 °C, with a total deviation of 170 Hz. The displacement amplitude increased from 12.61 μm to 13.31 μm across the same temperature range, indicating an improved mechanical response at higher temperatures. The motor achieved a maximum angular speed up to 1200 RPM and a stall torque of 13.1 N·mm at an excitation voltage amplitude of 180 Vp-p. The simple and scalable design, combined with its stability under varying temperature conditions, makes it well suited for small satellite applications, particularly in precision positioning tasks such as satellite orientation and free-space optical (FSO) communications.

Sensor Fusion Method for Object Detection and Distance Estimation in Assisted Driving Applications.

The fusion of multiple sensors' data in real-time is a crucial process for autonomous and assisted driving, where high-level controllers need classification of objects in the surroundings and estimation of relative positions. This paper presents an open-source framework to estimate the distance between a vehicle equipped with sensors and different road objects on its path using the fusion of data from cameras, radars, and LiDARs. The target application is an Advanced Driving Assistance System (ADAS) that benefits from the integration of the sensors' attributes to plan the vehicle's speed according to real-time road occupation and distance from obstacles. Based on geometrical projection, a low-level sensor fusion approach is proposed to map 3D point clouds into 2D camera images. The fusion information is used to estimate the distance of objects detected and labeled by a Yolov7 detector. The open-source pipeline implemented in ROS consists of a sensors' calibration method, a Yolov7 detector, 3D point cloud downsampling and clustering, and finally a 3D-to-2D transformation between the reference frames. The goal of the pipeline is to perform data association and estimate the distance of the identified road objects. The accuracy and performance are evaluated in real-world urban scenarios with commercial hardware. The pipeline running on an embedded Nvidia Jetson AGX achieves good accuracy on object identification and distance estimation, running at 5 Hz. The proposed framework introduces a flexible and resource-efficient method for data association from common automotive sensors and proves to be a promising solution for enabling effective environment perception ability for assisted driving.

PIMCoSim: Hardware/Software Co-Simulator for Exploring Processing-in-Memory Architectures

As the scope of artificial intelligence (AI) expands and the structure becomes more complex, the amount of data for inference and training has increased. In traditional computer architectures, the memory bandwidth limitations have intensified bottlenecks in AI systems, and processing-in-memory (PIM) architectures have been proposed to overcome this issue. PIM is an architecture that performs computations within memory, thereby reducing data movement between the CPU and memory. However, since PIM is difficult to optimize as a general-purpose architecture, it is essential to adopt an architecture suitable for the target application. While various simulators and emulators have been introduced for the design space exploration (DSE) of different PIM architectures, simulators are limited in debugging hardware operations, and emulators face challenges in flexibly modifying the system configuration, as emulators implement the entire architecture in hardware. Therefore, this paper introduces PIMCoSim, a comprehensive hardware–software co-simulator for the DSE of DRAM-PIM systems. This co-simulator partially emulates simplified hardware-implemented processing elements (PEs) and integrates software models for memory operations, facilitating the DSE of PIM systems. To validate PIMCoSim, we analyzed results for different computational workloads by varying PIM structures and operational policies, demonstrating the efficiency of DRAM-PIM systems. The co-simulation approach in PIMCoSim aims to contribute to analyzing DRAM-PIM configurations and adopting optimized structures.

Ring‐Opening Metathesis Polymerization of 1,2‐Dihydroazete Derivatives

AbstractFischer carbenes have recently found great utility in the construction of degradable metathesis materials, but investigations have been limited to oxygen‐containing enol ether monomers. Here, the ring‐opening metathesis polymerization of 1,2‐dihydroazetes is reported. The polymerization proceeds regioselectively, and the resulting molecular weights are targetable by adjusting the Grubbs initiator loading. Under acidic conditions, the resulting polymers degrade into 3‐aminopropanal derivatives through hydrolysis of the recurring enamide motifs in the polymer backbone. Additionally, the underlying kinetics and thermodynamics of the polymerization were studied through DFT calculations to elucidate the origins of metathesis regioselectivity. This work further expands the suite of monomers available to generate degradable metathesis materials and provides a flexible platform for target applications.

Preparation and characterisation of zirconia/hydroxyapatite bioactive composites as potential dental implants

In dental implants, zirconia is well-known as a crown material due to its excellent acid and base resistances and appearance close to natural teeth. In addition, its extraordinary mechanical properties render zirconia to be a potential candidate as an implant component of a whole implanted tooth, if its biocompatibility can be improved to promote adhesion to natural hard tissues. This study aims to enhance the bioactivity of zirconia with the aim of improving its integration with gum bone. Hydroxyapatite is the major component of natural bone and is thus selected as the modifier to improve the bioactivity of zirconia. A series of zirconia/hydroxyapatite composites with varied compositions were prepared under different conditions in order to find the optimal composites for the target application. Various analytical technologies and mechanical tests are employed to characterise the structure and properties of resultant composites. Results show that the component ratio and sintering temperature have a significant influence on the composite properties. An increase in hydroxyapatite component tends to enhance bioactivity but decline mechanical strength. Composites containing 10 wt% of hydroxyapatite maintain sufficient mechanical strength under the optimal sintering conditions whilst possessing excellent bioactivity, demonstrating that hydroxyapatite-modified zirconia has the potential as dental implant materials. Sintering results suggest that mechanical strength is obtained at 1400 °C for 2 h for the composite containing 10 wt% of hydroxyapatite.

Ring-Opening Metathesis Polymerization of 1,2-Dihydroazete Derivatives.

Fischer carbenes have recently found great utility in the construction of degradable metathesis materials, but investigations have been limited to oxygen-containing enol ether monomers. Here, the ring-opening metathesis polymerization of 1,2-dihydroazetes is reported. The polymerization proceeds regioselectively, and the resulting molecular weights are targetable by adjusting the Grubbs initiator loading. Under acidic conditions, the resulting polymers degrade into 3-aminopropanal derivatives through hydrolysis of the recurring enamide motifs in the polymer backbone. Additionally, the underlying kinetics and thermodynamics of the polymerization were studied through DFT calculations to elucidate the origins of metathesis regioselectivity. This work further expands the suite of monomers available to generate degradable metathesis materials and provides a flexible platform for target applications.

Astigmatism Quantification for Depth Localization of Bubbles and Tracers across Curved Surfaces

Abstract The present combined theoretical/experimental study addresses the impact of astigmatism on the two-phase flow diagnostics across the curved surfaces of liquid test-rig containments. In the present context, the target application is the two phase leakage-flow diagnostics across the annular housing gaps of oil-injected rotary positive displacement compressors (RPDC). Earlier studies by the authors identified the Defocusing Particle Tracking Velocimetry (DPTV) and Interferometric Particle Imaging (IPI) as particularly promising combination of flow measurement techniques to investigate the liquid and disperse gas phases inside the annular housing gap of RPDCs. The test-rig-specific influence of astigmatism on the resulting optical transfer function for a quantitative evaluation of the recorded defocused particle images (PI) is first compared to the theoretically derived circular PI diameter upon pure defocussing and subsequently tested for both classes of PIs, i.e DPTV and IPI. To mimic the optical configuration of optically accessible lateral surfaces of typical RPDC test rigs, a circular beaker glass (CBG) of comparable diameter is chosen for the experimental campaign. The results are discussed and future efforts for advanced PI-evaluation strategies are outlined on the grounds of the drawn conclusions.

Asymmetric multi-task learning for interpretable gaze-driven grasping action forecasting.

This work tackles the problem of automatically predicting the grasping intention of humans observing their environment, with eye-tracker glasses and video cameras recording the scene view. Our target application is the assistance to people with motor disabilities and potential cognitive impairments, using assistive robotics. Our proposal leverages the analysis of human attention captured in the form of gaze fixations recorded by an eye-tracker on the first person video, as the anticipation of prehension actions is a well studied and well known phenomenon. We propose a multi-task system that simultaneously addresses the prediction of human attention in the near future, and the anticipation of grasping actions. In our model, visual attention is modeled as a competitive process between a discrete set of states, each one associated to a well-known gaze movement pattern from visual psychology. We additionally consider an asymmetric multitask problem, where attention modeling is an auxiliary task that helps to regularize the learning process of the main action prediction task, and propose a constrained multi-task loss that naturally deals with this asymmetry. Our model shows superior performance than other losses for dynamic multi-task learning, current dominant deep architectures for general action forecasting and particularly-tailored models for predicting grasping intention. In particular, it provides state-of-the-art performance in three datasets for egocentric action anticipation, with an average precision of 0.569 and 0.524 in GITW and Sharon datasets, respectively, and an accuracy of 89.2% and a success rate of 51.7% in Invisible dataset.

SeeSaw: Learning Soft Tissue Deformation from Laparoscopy Videos with GNNs.

A major challenge in image-guided laparoscopic surgery is that structures of interest often deform and go, even if only momentarily, out of view. Methods which rely on having an up-to-date impression of those structures, such as registration or localisation, are undermined in these circumstances. This is particularly true for soft-tissue structures that continually change shape - in registration, they must often be re-mapped. Furthermore, methods which require 'revisiting' of previously seen areas cannot in principle function reliably in dynamic contexts, drastically weakening their uptake in the operating room. We present a novel approach for learning to estimate the deformed states of previously seen soft tissue surfaces from currently observable regions, using a combined approach that includes a Graph Neural Network (GNN). The training data is based on stereo laparoscopic surgery videos, generated semi-automatically with minimal labelling effort. Trackable segments are first identified using a feature detection algorithm, from which surface meshes are produced using depth estimation and delaunay triangulation. We show the method can predict the displacements of previously visible soft tissue structures connected to currently visible regions with observed displacements, both on our own data and porcine data. Our innovative approach learns to compensate non-rigidity in abdominal endoscopic scenes directly from stereo laparoscopic videos through targeting a new problem formulation, and stands to benefit a variety of target applications in dynamic environments. Project page for this work: https://gitlab.com/nct_tso_public/seesaw-soft-tissue-deformation.

A comparison between black-, gray- and white-box modeling for the bidirectional Raman amplifier optimization

Designing and optimizing optical amplifiers to maximize system performance is becoming increasingly important as optical communication systems strive to increase throughput. Offline optimization of optical amplifiers relies on models ranging from white-box models deeply rooted in physics to black-box data-driven and physics-agnostic models. Here, we compare the capabilities of white-, gray- and black-box models on the challenging test case of optimizing a bidirectional distributed Raman amplifier to achieve a target frequency-distance signal power profile. We show that any of the studied methods can achieve similar frequency and distance flatness of between 1 and 3.6 dB (depending on the definition of flatness) over the C-band in an 80-km span. Then, we discuss the models’ applicability, advantages, and drawbacks based on the target application scenario, in particular in terms of flexibility, optimization speed, and access to training data.

Implications of Charge and Heteroatom Dopants on the Thermodynamics and Kinetics of Redox Reactions in Keggin-Type Polyoxometalates.

The utilization of polyoxometalate-based materials is largely dictated by their redox properties. Detailed understanding of the thermodynamic and kinetic efficiency of charge transfer is therefore essential to the development of polyoxometalate-based systems for target applications. Toward this end, we report electrochemical studies of a series of heteroatom-doped Keggin-type polyoxotungstate clusters [PW12O40]3- (PW 12 ), [VW12O40]3- (V in W 12 ), [P(VW11)O40]4- (PV out W 11 ), and [V(VW11)O40]4- (V in V out W 11 ) to elucidate the role of the identity and spatial location of heteroatoms and overall cluster charge on the rate constants of electron transfer and redox reaction entropies. Electrochemical analyses of the polyoxotungstates reveal that the kinetics of electron transfer for W-based redox processes change as a function of the redox activity of the heteroatom, whereas the spatial location of the heteroatom dopant does not significantly impact the electrokinetics. Variable temperature cyclic voltammetry measurements in organic solutions containing noncoordinating electrolyte ions establish that redox reaction entropies are primarily dictated by the overall charge of the clusters. Specifically, the redox entropy exhibits a good linear relationship with the dielectric continuum function Z ox 2 - Z red 2 (Z ox = charge of oxidized species, Z red = charge of reduced species). Finally, our experimental data do not show a prominent correlation between the kinetics of electron transfer and redox entropy, implying that the charge-transfer kinetics are not solely governed by structural reorganization. Taken together, these results highlight how structural and electronic parameters can influence the kinetics and thermodynamics of charge transfer in polyoxotungstates and provide insights into the design of polyoxometalate compounds with target redox properties.

Pyrochlore-like Local Structure in Globally Disordered Y2Zr2O7: Evidence and Reasoning by Combined Theoretical and Experimental Study.

Even though technological relevance for nuclear and oxide fuel cell application has persuaded a large number of investigations on the Y2Zr2O7 system, its local structure still remains ambiguous and debatable. While diffraction-based investigations claim a lack of local ordering, the MAS NMR and Raman spectroscopic investigations speculate the presence of local ordering. Besides, a correlation between the local and global structures of Y2Zr2O7 is also missing to date. Present work attempts to get deeper insights into the local structure of Y2Zr2O7 and correlate the local structure with global disordering. Complementary investigation using multiple strategies (XRD, RAMAN, electron microscopy, DFT calculations) revealed the presence of pyrochlore-like features in globally disordered Y2Zr2O7. Globally distributed "local probe ions" (Sn4+ by MAS NMR and Eu3+ by photoluminescence (PL)) were utilized for establishing the correlation between local and global structures. The 119Sn MAS NMR spectra possess the shoulder peak bearing chemical shift values, which are improbable under complete cationic randomization. The emission and decay profiles of the Eu3+ ion recommend positioning of few probe ions at the centrosymmetric site, which was further reinforced by crystal field splitting and lifetime values. Thus, both NMR and PL results reaffirm the existence of a pyrochlore-like atomic arrangement in predominantly defective fluorite Y2Zr2O7 phase. These discernible pyrochlore-like features have been envisaged to be emanating from unevenness in bond strength and ionicity/covalency of Y-O and Zr-O bonds. Cationic and anionic randomization energetically endorses global disordering, but the bonding disparity induces short-range ordering. It is believed that the presented findings will guide future research in regulating the physicochemical properties of Y2Zr2O7 for the target applications.

Boosting the Microbial Electrosynthesis of Formate by Shewanella oneidensis MR-1 with an Ionic Liquid Cosolvent.

Microbial electrosynthesis (MES) is a rapidly growing technology at the forefront of sustainable chemistry, leveraging the ability of microorganisms to catalyze electrochemical reactions to synthesize valuable compounds from renewable energy sources. The reduction of CO2 is a major target application for MES, but research in this area has been stifled, especially with the use of direct electron transfer (DET)-based microbial systems. The major fundamental hurdle that needs to be overcome is the low efficiency of CO2 reduction largely attributed to minimal microbial access to CO2 owing to its low solubility in the electrolyte. With their tunable physical properties, ionic liquids present a potential solution to this challenge and have previously shown promise in facilitating efficient CO2 electroreduction by increasing the CO2 solubility. However, the use of ionic liquids in MES remains unexplored. In this study, we investigated the role of 1-ethyl-3-methylimidazolium acetate ([EMIM][Ac]) using Shewanella oneidensis MR-1 as a model DET strain. Electrochemical investigations demonstrated the ability of S. oneidensis MR-1 biocathodes to directly convert CO2 to formate with a faradaic efficiency of 34.5 ± 26.1%. The addition of [EMIM][Ac] to the system significantly increased cathodic current density and enhanced the faradaic efficiency to 94.5 ± 4.3% while concurrently amplifying the product yield from 34 ± 23 μM to 366 ± 34 μM. These findings demonstrate that ionic liquids can serve as efficient, biocompatible cosolvents for microbial electrochemical reduction of CO2 to value-added products, holding promise for more robust applications of MES.

Supramolecular Janus Nanocylinders: Controlling Their Characteristics by the Self-Assembly Process.

Janus NanoRods (JNR) are anisotropic and non-symmetrical colloids with two faces of different chemical composition. They are difficult to prepare because of their nanometric dimensions and strong anisotropy. Recently, a versatile strategy was developed, allowing the formation of JNR relying on the self-assembly in aqueous medium of two polymers end-functionalized with non-symmetrical and complementary hydrogen bonding stickers. However, the supramolecular JNR prepared following this strategy are out-of-equilibrium (frozen) and therefore their characteristics depend on the self-assembly process. The present study elucidates the formation mechanism of the JNR and the parameters of the self-assembly process influencing their characteristics. The polymers are initially dissolved as unimers in DMSO. Dropwise addition of water triggers the rapid assembly of more and more unimers into long nanocylinders that are unable to grow anymore once formed. Consequently, increasing the dropwise addition rate of water hardly impacts the process, whereas lowering the initial polymer concentration in DMSO reduces both the length and proportion of nanocylinders. Increasing temperature during water addition weakens hydrogen bonds, triggering the formation of a mixture of spheres and nanocylinders. Many supramolecular polymer assemblies are frozen in solution and these findings should help understanding how to control their characteristics, allowing to adapt them to a target application.

Multi-Resolution Real-Time Deep Pose-Space Deformation

We present a hard-real-time multi-resolution mesh shape deformation technique for skeleton-driven soft-body characters. Producing mesh deformations at multiple levels of detail is very important in many applications in computer graphics. Our work targets applications where the multi-resolution shapes must be generated at fast speeds ("hard-real-time", e.g., a few milliseconds at most and preferably under 1 millisecond), as commonly needed in computer games, virtual reality and Metaverse applications. We assume that the character mesh is driven by a skeleton, and that high-quality character shapes are available in a set of training poses originating from a high-quality (slow) rig such as volumetric FEM simulation. Our method combines multi-resolution analysis, mesh partition of unity, and neural networks, to learn the pre-skinning shape deformations in an arbitrary character pose. Combined with linear blend skinning, this makes it possible to reconstruct the training shapes, as well as interpolate and extrapolate them. Crucially, we simultaneously achieve this at hard real-time rates and at multiple mesh resolution levels. Our technique makes it possible to trade deformation quality for memory and computation speed, to accommodate the strict requirements of modern real-time systems. Furthermore, we propose memory layout and code improvements to boost computation speeds. Previous methods for realtime approximations of quality shape deformations did not focus on hard real-time, or did not investigate the multi-resolution aspect of the problem. Compared to a "naive" approach of separately processing each hierarchical level of detail, our method offers a substantial memory reduction as well as computational speedups. It also makes it possible to construct the shape progressively level by level and interrupt the computation at any time, enabling graceful degradation of the deformation detail. We demonstrate our technique on several examples, including a stylized human character, human hands, and an inverse-kinematics-driven quadruped animal.