Multi-phase Liver Tumor Segmentation with Spatial Aggregation and Uncertain Region Inpainting

Multi-phase computed tomography (CT) images provide crucial complementary information for accurate liver tumor segmentation (LiTS). State-of-the-art multi-phase LiTS methods usually fused cross-phase features through phase-weighted summation or channel-attention based concatenation. However, these methods ignored the spatial (pixel-wise) relationships between different phases, hence leading to insufficient feature integration. In addition, the performance of existing methods remains subject to the uncertainty in segmentation, which is particularly acute in tumor boundary regions. In this work, we propose a novel LiTS method to adequately aggregate multi-phase information and refine uncertain region segmentation. To this end, we introduce a spatial aggregation module (SAM), which encourages per-pixel interactions between different phases, to make full use of cross-phase information. Moreover, we devise an uncertain region inpainting module (URIM) to refine uncertain pixels using neighboring discriminative features. Experiments on an in-house multi-phase CT dataset of focal liver lesions (MPCT-FLLs) demonstrate that our method achieves promising liver tumor segmentation and outperforms state-of-the-arts.

Signal Transformer: Complex-valued Attention and Meta-Learning for Signal Recognition

Deep neural networks have been shown as a class of useful tools for addressing signal recognition issues in recent years, especially for identifying the nonlinear feature structures of signals. However, this power of most deep learning techniques heavily relies on an abundant amount of training data, so the performance of classic neural nets decreases sharply when the number of training data samples is small or unseen data are presented in the testing phase. This calls for an advanced strategy, i.e., model-agnostic meta-learning (MAML), which is able to capture the invariant representation of the data samples or signals. In this paper, inspired by the special structure of the signal, i.e., real and imaginary parts consisted in practical time-series signals, we propose a Complex-valued Attentional MEta Learner (CAMEL) for the problem of few-shot signal recognition by leveraging attention and meta-learning in the complex domain. To the best of our knowledge, this is also the first complex-valued MAML that can find the first-order stationary points of general nonconvex problems with theoretical convergence guarantees. Extensive experiments results showcase the superiority of the proposed CAMEL compared with the state-of-the-art methods.

Observation of orbital order in the Van der Waals material 1T-TiSe2

Besides magnetic and charge order, regular arrangements of orbital occupation constitute a fundamental order parameter of condensed matter physics. Even though orbital order is difficult to identify directly in experiments, its presence was firmly established in a number of strongly correlated, three-dimensional Mott insulators. Here, reporting resonant X-ray scattering experiments on the layered Van der Waals compound $1T$-TiSe$_2$, we establish the emergence of orbital order in a weakly correlated, quasi-two-dimensional material. Our experimental scattering results are consistent with first-principles calculations that bring to the fore a generic mechanism of close interplay between charge redistribution, lattice displacements, and orbital order. It demonstrates the essential role that orbital degrees of freedom play in TiSe$_2$, and their importance throughout the family of correlated Van der Waals materials.

Factorization for entropy production of the Eikonal equation and regularity

The Eikonal equation arises naturally in the limit of the second order Aviles-Giga functional whose $Gamma$-convergence is a long standing challenging problem. The theory of entropy solutions of the Eikonal equation plays a central role in the variational analysis of this problem. Establishing fine structures of entropy solutions of the Eikonal equation, e.g. concentration of entropy measures on $mathcal{H}^1$-rectifiable sets in $2$D, is arguably the key missing part for a proof of the full $Gamma$-convergence of the Aviles-Giga functional. In the first part of this work, for $pin left(1,frac{4}{3}right]$ we establish an $L^p$ version of the main theorem of Ghiraldin and Lamy [Comm. Pure Appl. Math. 73 (2020), no. 2, 317-349]. Specifically we show that if $m$ is a solution to the Eikonal equation, then $min B^{frac{1}{3}}_{3p,infty,loc}$ is equivalent to all entropy productions of $m$ being in $L^p_{loc}$. This result also shows that as a consequence of a weak form of the Aviles-Giga conjecture (namely the conjecture that all solutions to the Eikonal equation whose entropy productions are in $L^p_{loc}$ are rigid) - the rigidity/flexibility threshold of the Eikonal equation is exactly the space $ B^{frac{1}{3}}_{3,infty,loc}$. In the second part of this paper, under the assumption that all entropy productions are in $L^p_{loc}$, we establish a factorization formula for entropy productions of solutions of the Eikonal equation in terms of the two Jin-Kohn entropies. A consequence of this formula is control of all entropy productions by the Jin-Kohn entropies in the $L^p$ setting - this is a strong extension of an earlier result of the authors [Annales de lInstitut Henri Poincar{e}. Analyse Non Lin{e}aire 35 (2018), no. 2, 481-516].

Solving stochastic optimal control problem via stochastic maximum principle with deep learning method

In this paper, we aim to solve the high dimensional stochastic optimal control problem from the view of the stochastic maximum principle via deep learning. By introducing the extended Hamiltonian system which is essentially an FBSDE with a maximum condition, we reformulate the original control problem as a new one. Three algorithms are proposed to solve the new control problem. Numerical results for different examples demonstrate the effectiveness of our proposed algorithms, especially in high dimensional cases. And an important application of this method is to calculate the sub-linear expectations, which correspond to a kind of fully nonlinear PDEs.

A Survey of Adversarial Learning on Graphs

Deep learning models on graphs have achieved remarkable performance in various graph analysis tasks, e.g., node classification, link prediction and graph clustering. However, they expose uncertainty and unreliability against the well-designed inputs, i.e., adversarial examples. Accordingly, a line of studies have emerged for both attack and defense addressed in different graph analysis tasks, leading to the arms race in graph adversarial learning. Despite the booming works, there still lacks a unified problem definition and a comprehensive review. To bridge this gap, we investigate and summarize the existing works on graph adversarial learning tasks systemically. Specifically, we survey and unify the existing works w.r.t. attack and defense in graph analysis tasks, and give appropriate definitions and taxonomies at the same time. Besides, we emphasize the importance of related evaluation metrics, investigate and summarize them comprehensively. Hopefully, our works can provide a comprehensive overview and offer insights for the relevant researchers. More details of our works are available at https://github.com/gitgiter/Graph-Adversarial-Learning.

Three algorithms for solving high-dimensional fully-coupled FBSDEs through deep learning

Recently, the deep learning method has been used for solving forward-backward stochastic differential equations (FBSDEs) and parabolic partial differential equations (PDEs). It has good accuracy and performance for high-dimensional problems. In this paper, we mainly solve fully coupled FBSDEs through deep learning and provide three algorithms. Several numerical results show remarkable performance especially for high-dimensional cases.

Voltage-induced high-speed DW motion in a synthetic antiferromagnet

Voltage-induced motion of a magnetic domain wall (DW) has potential in developing novel devices with ultralow dissipation. However, the speed for the voltage-induced DW motion (VIDWM) in a single ferromagnetic layer is usually very low. In this work, we proposed VIDWM with high speed in a synthetic antiferromaget (SAF). The velocity for the coupled DWs in the SAF is significantly higher than its counterpart in a single ferromagnetic layer. Strong interlayer antiferromagnetic exchange coupling plays a critical role for the high DW velocity since it inhibits the tilting of DW plane with strong Dzyaloshinskii-Moriya interaction. On the other hand, the Walker breakdown of DW motion is also inhibited due to the stabilization of moment orientation under a strong interlayer antiferromagnetic coupling. In theory, the voltage-induced gradient of magnetic anisotropy is proved to be equal to an effective magnetic field that drives DW.

Dispersive charge density wave excitations and temperature dependent commensuration in Bi2Sr2CaCu2O8+{delta}

Experimental evidence on high-Tc cuprates reveals ubiquitous charge density wave (CDW) modulations, which coexist with superconductivity. Although the CDW had been predicted by theory, important questions remain about the extent to which the CDW influences lattice and charge degrees of freedom and its characteristics as functions of doping and temperature. These questions are intimately connected to the origin of the CDW and its relation to the mysterious cuprate pseudogap. Here, we use ultrahigh resolution resonant inelastic x-ray scattering (RIXS) to reveal new CDW character in underdoped Bi2Sr2CaCu2O8+{delta} (Bi2212). At low temperature, we observe dispersive excitations from an incommensurate CDW that induces anomalously enhanced phonon intensity, unseen using other techniques. Near the pseudogap temperature T*, the CDW persists, but the associated excitations significantly weaken and the CDW wavevector shifts, becoming nearly commensurate with a periodicity of four lattice constants. The dispersive CDW excitations, phonon anomaly, and temperature dependent commensuration provide a comprehensive momentum space picture of complex CDW behavior and point to a closer relationship with the pseudogap state.

Three-dimensional single gyroid photonic crystals with a mid-infrared bandgap

A gyroid structure is a distinct morphology that is triply periodic and consists of minimal isosurfaces containing no straight lines. We have designed and synthesized amorphous silicon (a-Si) mid-infrared gyroid photonic crystals that exhibit a complete bandgap in infrared spectroscopy measurements. Photonic crystals were synthesized by deposition of a-Si/Al2O3 coatings onto a sacrificial polymer scaffold defined by two-photon lithography. We observed a 100% reflectance at 7.5 mum for single gyroids with a unit cell size of 4.5 mum, in agreement with the photonic bandgap position predicted from full-wave electromagnetic simulations, whereas the observed reflection peak shifted to 8 um for a 5.5 mum unit cell size. This approach represents a simulation-fabrication-characterization platform to realize three-dimensional gyroid photonic crystals with well-defined dimensions in real space and tailored properties in momentum space.