Search

Multi-modal medical image completion has been extensively applied to alleviate the missing modality issue in a wealth of multi-modal diagnostic tasks. However, for most existing synthesis methods, their inferences of missing modalities can collapse into a deterministic mapping from the available ones, ignoring the uncertainties inherent in the cross-modal relationships. Here, we propose the Unified Multi-Modal Conditional Score-based Generative Model (UMM-CSGM) to take advantage of Score-based Generative Model (SGM) in modeling and stochastically sampling a target probability distribution, and further extend SGM to cross-modal conditional synthesis for various missing-modality configurations in a unified framework. Specifically, UMM-CSGM employs a novel multi-in multi-out Conditional Score Network (mm-CSN) to learn a comprehensive set of cross-modal conditional distributions via conditional diffusion and reverse generation in the complete modality space. In this way, the generation process can be accurately conditioned by all available information, and can fit all possible configurations of missing modalities in a single network. Experiments on BraTS19 dataset show that the UMM-CSGM can more reliably synthesize the heterogeneous enhancement and irregular area in tumor-induced lesions for any missing modalities.

Knee osteoarthritis (OA) is the most common osteoarthritis and a leading cause of disability. Cartilage defects are regarded as major manifestations of knee OA, which are visible by magnetic resonance imaging (MRI). Thus early detection and assessment for knee cartilage defects are important for protecting patients from knee OA. In this way, many attempts have been made on knee cartilage defect assessment by applying convolutional neural networks (CNNs) to knee MRI. However, the physiologic characteristics of the cartilage may hinder such efforts: the cartilage is a thin curved layer, implying that only a small portion of voxels in knee MRI can contribute to the cartilage defect assessment; heterogeneous scanning protocols further challenge the feasibility of the CNNs in clinical practice; the CNN-based knee cartilage evaluation results lack interpretability. To address these challenges, we model the cartilages structure and appearance from knee MRI into a graph representation, which is capable of handling highly diverse clinical data. Then, guided by the cartilage graph representation, we design a non-Euclidean deep learning network with the self-attention mechanism, to extract cartilage features in the local and global, and to derive the final assessment with a visualized result. Our comprehensive experiments show that the proposed method yields superior performance in knee cartilage defect assessment, plus its convenient 3D visualization for interpretability., Comment: 10 pages, 4 figures

Functional Connectivity (FC) matrices measure the regional interactions in the brain and have been widely used in neurological brain disease classification. However, a FC matrix is neither a natural image which contains shape and texture information, nor a vector of independent features, which renders the extracting of efficient features from matrices as a challenging problem. A brain network, also named as connectome, could forma a graph structure naturally, the nodes of which are brain regions and the edges are interregional connectivity. Thus, in this study, we proposed novel graph convolutional networks (GCNs) to extract efficient disease-related features from FC matrices. Considering the time-dependent nature of brain activity, we computed dynamic FC matrices with sliding-windows and implemented a graph convolution based LSTM (long short term memory) layer to process dynamic graphs. Moreover, the demographics of patients were also used to guide the classification. However, unlike in conventional methods where personal information, i.e., gender and age were added as extra inputs, we argue that this kind of approach may not actually improve the classification performance, for such personal information given in dataset was usually balanced distributed. In this paper, we proposed to utilize the demographic information as extra outputs and to share parameters among three networks predicting subject status, gender and age, which serve as assistant tasks. We tested the performance of the proposed architecture in ADNI II dataset to classify Alzheimer's disease patients from normal controls. The classification accuracy, sensitivity and specificity reach 0.90, 0.92 and 0.89 on ADNI II dataset., Comment: Number of pages: 22 Word count in abstract: 252, word count in manuscript: 3910 This manuscript includes 3 tables and 9 figures

The material class of hybrid organic-inorganic perovskites has risen rapidly from a virtually unknown material in photovoltaic applications a short 7 years ago into a ~20% efficient thin-film solar cell material. As promising as this class of materials is, however, there are limitations associated with its poor long-term stability, non-optimal band gap, presence of environmentally-toxic Pb element, etc. We herein apply a functionality-directed theoretical materials selection approach as a filter for initial screening of the compounds that satisfy the desired intrinsic photovoltaic functionalities and might overcome the above limitations. First-principles calculations are employed to systemically study thermodynamic stability and photovoltaic-related properties of hundred of candidate hybrid perovskites. We have identified in this materials selection process fourteen Ge and Sn-based materials with potential superior bulk-material-intrinsic photovoltaic performance. A distinct class of compounds containing NH3COH$^+$ with the organic molecule derived states intriguingly emerging at band-edges is found. Comparison of various candidate materials offers insights on how composition variation and microscopic structural changes affect key photovoltaic relevant properties in this family of materials., Comment: 17 pages, 10 figures,Supplementary information

This paper presents a new Bayesian non-parametric model by extending the usage of Hierarchical Dirichlet Allocation to extract tree structured word clusters from text data. The inference algorithm of the model collects words in a cluster if they share similar distribution over documents. In our experiments, we observed meaningful hierarchical structures on NIPS corpus and radiology reports collected from public repositories.

The kinetics of the charge carrier recombination in dye molecule-doped multilayer organic light-emitting diodes (OLEDs) was quantified by transient electroluminescence (EL). Three sets of dye molecules, such as derivatives of naphthalimide and stilbene, were used as dopants in light-emission layer. Although the devices show almost the same EL spectra for each set of molecules, they show very different EL efficiency. The difference in EL efficiency was attributed to the difference in charge carrier recombination, as revealed by transient EL. The recombination coefficient ({\gamma}) was determined from the long-time component of the temporal decay of the EL intensity after a rectangular voltage pulse was turned off. It was found that {\gamma} and EL efficiency were both strongly dependent on the molecular structures of the dopants, and the donor groups and {\pi}-conjugated structure guaranteed high {\gamma} and EL efficiency in OLEDs., Comment: in press

Based on the framework of Kubo formulism, we develop the minimally entangled typical thermal state algorithm to study the temperature and time dependence of current-current correlation function in one-dimensional spinless fermion model, taking into account both the electron-electron (e-e) intersite interaction and the dynamic disorder induced by classical phonons. Without e-e interaction, the numerical results, showing an exponential decay of the time dependent correlation, could be precisely compared with that from the analytical derivation, namely, from the generalized Langevin equation. More importantly, when a strong enough e-e interaction is presence, we find a long-time correlation in the regime of small dynamic disorder, indicating the breakdown of thermal relaxation, which is a typical many-body effect. On the basis of this finding, we show that it might be applied to understand the metalliclike charge transport and the abnormal improvement of the conductivity with respect to the redoping experiment in K$_3$C$_{60}$, an organic superconducting material., Comment: 8 pages, 3 figures

We describe Information Forests, an approach to classification that generalizes Random Forests by replacing the splitting criterion of non-leaf nodes from a discriminative one -- based on the entropy of the label distribution -- to a generative one -- based on maximizing the information divergence between the class-conditional distributions in the resulting partitions. The basic idea consists of deferring classification until a measure of "classification confidence" is sufficiently high, and instead breaking down the data so as to maximize this measure. In an alternative interpretation, Information Forests attempt to partition the data into subsets that are "as informative as possible" for the purpose of the task, which is to classify the data. Classification confidence, or informative content of the subsets, is quantified by the Information Divergence. Our approach relates to active learning, semi-supervised learning, mixed generative/discriminative learning., Comment: Proceedings of the Information Theory and Applications (ITA) Workshop, 2/7/2012