Search

We present Flat'n'Fold, a novel large-scale dataset for garment manipulation that addresses critical gaps in existing datasets. Comprising 1,212 human and 887 robot demonstrations of flattening and folding 44 unique garments across 8 categories, Flat'n'Fold surpasses prior datasets in size, scope, and diversity. Our dataset uniquely captures the entire manipulation process from crumpled to folded states, providing synchronized multi-view RGB-D images, point clouds, and action data, including hand or gripper positions and rotations. We quantify the dataset's diversity and complexity compared to existing benchmarks and show that our dataset features natural and diverse manipulations of real-world demonstrations of human and robot demonstrations in terms of visual and action information. To showcase Flat'n'Fold's utility, we establish new benchmarks for grasping point prediction and subtask decomposition. Our evaluation of state-of-the-art models on these tasks reveals significant room for improvement. This underscores Flat'n'Fold's potential to drive advances in robotic perception and manipulation of deformable objects. Our dataset can be downloaded at https://cvas-ug.github.io/flat-n-fold

Molecular orbital (MO) systems with clusters of heavy transition metal (TM) ions are one of the most important classes of model materials for studying the interplay between local physics and effects of itinerancy. Despite a large number of candidates identified in the family of 4d TM materials, an understanding of their physics from competing \textit{microscopic} energy scales is still missing. We bridge this gap by reporting the first resonant inelastic X-ray scattering (RIXS) measurement on a well-known series of Ru dimer systems with a 6H-perovskite structure, Ba$_3$MRu$_2$O$_9$ (M$^{3+}$=In$^{3+}$, Y$^{3+}$, La$^{3+}$). Our RIXS measurements reveal an extremely fragile MO state in these Ru dimer compounds, evidenced by an abrupt change in the RIXS spectrum accompanying a tiny change in the local structure tuned by the M-site ion. By modelling the RIXS spectra, we attribute the enhanced electronic instability in Ba$_3$MRu$_2$O$_9$ to the combined effect of a large hopping and a small spin-orbit coupling in the Ru dimers. The unique combination of energy scales uncovered in the present study make Ru MO systems ideal model systems for studying quantum phase transitions with molecular orbitals., Comment: 7 pages, 3 figures, Supplemental Materials available upon request

Single-photon emitters serve as building blocks for many emerging concepts in quantum photonics. The recent identification of bright, tunable, and stable emitters in hexagonal boron nitride (hBN) has opened the door to quantum platforms operating across the infrared to ultraviolet spectrum. While it is widely acknowledged that defects are responsible for single-photon emitters in hBN, crucial details regarding their origin, electronic levels, and orbital involvement remain unknown. Here, we employ a combination of resonant inelastic X-ray scattering and photoluminescence spectroscopy in defective hBN unveiling an elementary excitation at 285 meV that gives rise to a plethora of harmonics correlated with single-photon emitters. We discuss the importance of N $\pi^*$ antibonding orbitals in shaping the electronic states of the emitters. The discovery of the elementary excitations of hBN provides new fundamental insights into quantum emission in low-dimensional materials, paving the way for future investigations in other platforms.

Fusing multi-modal data can improve the performance of deep learning models. However, missing modalities are common for medical data due to patients' specificity, which is detrimental to the performance of multi-modal models in applications. Therefore, it is critical to adapt the models to missing modalities. This study aimed to develop an efficient multi-modal fusion architecture for medical data that was robust to missing modalities and further improved the performance on disease diagnosis.X-ray chest radiographs for the image modality, radiology reports for the text modality, and structured value data for the tabular data modality were fused in this study. Each modality pair was fused with a Transformer-based bi-modal fusion module, and the three bi-modal fusion modules were then combined into a tri-modal fusion framework. Additionally, multivariate loss functions were introduced into the training process to improve model's robustness to missing modalities in the inference process. Finally, we designed comparison and ablation experiments for validating the effectiveness of the fusion, the robustness to missing modalities and the enhancements from each key component. Experiments were conducted on MIMIC-IV, MIMIC-CXR with the 14-label disease diagnosis task. Areas under the receiver operating characteristic curve (AUROC), the area under the precision-recall curve (AUPRC) were used to evaluate models' performance. The experimental results demonstrated that our proposed multi-modal fusion architecture effectively fused three modalities and showed strong robustness to missing modalities. This method is hopeful to be scaled to more modalities to enhance the clinical practicality of the model.

Strongly correlated materials often respond sensitively to the external perturbations. In the recently discovered superconducting infinite-layer nickelates, the superconducting transition temperature can be dramatically enhanced via only ~1% compressive strain-tuning enabled by substrate design. However, the root of such enhancement remains elusive. While the superconducting pairing mechanism is still not settled, magnetic Cooper pairing - similar to the cuprates has been proposed. Using resonant inelastic x-ray scattering, we investigate the magnetic excitations in infinite-layer PrNiO2 thin films for different strain conditions. The magnon bandwidth of PrNiO2 shows only marginal response to strain-tuning, in sharp contrast to the striking enhancement of the superconducting transition temperature Tc in the doped superconducting samples. These results suggest the enhancement of Tc is not mediated by spin excitations and thus provide important empirics for the understanding of superconductivity in infinite-layer nickelates., Comment: 17 pages, 4 figures

The binary ruthenate, RuO$_2$, has been the subject of intense interest due to its itinerant antiferromagnetism and strain-induced superconductivity. The strain mechanism and its effect on the microscopic electronic states leading to the normal and superconducting state, however, remain undisclosed. Here, we investigate highly-strained epitaxial (110) RuO$_2$ films using polarization-dependent oxygen K-edge X-ray absorption spectroscopy (XAS). Through the detection of pre-edge peaks, arising from O:$2p$ - Ru:$4d$ hybridization, we uncover the effects of epitaxial strain on the orbital/electronic structure near the Fermi level. Our data show robust strain-induced shifts of orbital levels and a reduction of hybridization strength. Furthermore, we reveal a pronounced in-plane anisotropy of the electronic structure along the $[110]/[1\bar{1}0]$ directions naturally stemming from the symmetry-breaking epitaxial strain of the substrate. The $B_{2g}$ symmetry component of the epitaxially-enforced strain breaks a sublattice degeneracy, resulting in an increase of the density of states at the Fermi level ($E_F$), possibly paving the way to superconductivity. These results underscore the importance of the effective reduction from tetragonal to orthorhombic lattice symmetry in (110) RuO$_2$ films and its relevance towards the superconducting and magnetic properties.

Hafnia (HfO2) is a promising material for emerging chip applications due to its high-k dielectric behaviour, suitability for negative capacitance heterostructures, scalable ferroelectricity, and silicon compatibility. The lattice dynamics along with phononic properties such as thermal conductivity, contraction, and heat capacity are under-explored, primarily due to the absence of high quality single crystals. Herein, we report the vibrational properties of a series of HfO2 crystals stabilized with yttrium (chemical formula HfO2:xY, where x = 20, 12, 11, 8, and 0%) and compare our findings with a symmetry analysis and lattice dynamics calculations. We untangle the effects of Y by testing our calculations against the measured Raman and infrared spectra of the cubic, antipolar orthorhombic, and monoclinic phases and then proceed to reveal the signature modes of polar orthorhombic hafnia. This work provides a spectroscopic fingerprint for several different phases of HfO2 and paves the way for an analysis of mode contributions to high-k dielectric and ferroelectric properties for chip technologies.

Limited by the performance factor, it is arduous to recognize target object and locate it in Augmented Reality (AR) scenes on low-end mobile devices, especially which using monocular cameras. In this paper, we proposed an algorithm which can quickly locate the target object through image object detection in the circumstances of having very sparse feature points. We introduce YOLOv3-Tiny to our algorithm as the object detection module to filter the possible points and using Principal Component Analysis (PCA) to determine the location. We conduct the experiment in a manually designed scene by holding a smartphone and the results represent high positioning speed and accuracy of our method.

This paper constructs a novel intelligent medical diagnosis system, which can realize automatic communication and breast cancer pathological image recognition. This system contains two main parts, including a pre-training chatbot called M-Chatbot and an improved neural network model of EfficientNetV2-S named EfficientNetV2-SA, in which the activation function in top layers is replaced by ACON-C. Using information retrieval mechanism, M-Chatbot instructs patients to send breast pathological image to EfficientNetV2-SA network, and then the classifier trained by transfer learning will return the diagnosis results. We verify the performance of our chatbot and classification on the extrinsic metrics and BreaKHis dataset, respectively. The task completion rate of M-Chatbot reached 63.33\%. For the BreaKHis dataset, the highest accuracy of EfficientNetV2-SA network have achieved 84.71\%. All these experimental results illustrate that the proposed model can improve the accuracy performance of image recognition and our new intelligent medical diagnosis system is successful and efficient in providing automatic diagnosis of breast cancer.