Automatic methods to segment the vestibular schwannoma (VS) tumors and the cochlea from magnetic resonance imaging (MRI) are critical to VS treatment planning. Although supervised methods have achieved satisfactory performance in VS segmentation, the
y require full annotations by experts, which is laborious and time-consuming. In this work, we aim to tackle the VS and cochlea segmentation problem in an unsupervised domain adaptation setting. Our proposed method leverages both the image-level domain alignment to minimize the domain divergence and semi-supervised training to further boost the performance. Furthermore, we propose to fuse the labels predicted from multiple models via noisy label correction. Our results on the challenge validation leaderboard showed that our unsupervised method has achieved promising VS and cochlea segmentation performance with mean dice score of 0.8261 $pm$ 0.0416; The mean dice value for the tumor is 0.8302 $pm$ 0.0772. This is comparable to the weakly-supervised based method.
Data processing and analytics are fundamental and pervasive. Algorithms play a vital role in data processing and analytics where many algorithm designs have incorporated heuristics and general rules from human knowledge and experience to improve thei
r effectiveness. Recently, reinforcement learning, deep reinforcement learning (DRL) in particular, is increasingly explored and exploited in many areas because it can learn better strategies in complicated environments it is interacting with than statically designed algorithms. Motivated by this trend, we provide a comprehensive review of recent works focusing on utilizing deep reinforcement learning to improve data processing and analytics. First, we present an introduction to key concepts, theories, and methods in deep reinforcement learning. Next, we discuss deep reinforcement learning deployment on database systems, facilitating data processing and analytics in various aspects, including data organization, scheduling, tuning, and indexing. Then, we survey the application of deep reinforcement learning in data processing and analytics, ranging from data preparation, natural language interface to healthcare, fintech, etc. Finally, we discuss important open challenges and future research directions of using deep reinforcement learning in data processing and analytics.
Optical coherence tomography (OCT) is a non-invasive imaging technique widely used for ophthalmology. It can be extended to OCT angiography (OCT-A), which reveals the retinal vasculature with improved contrast. Recent deep learning algorithms produce
d promising vascular segmentation results; however, 3D retinal vessel segmentation remains difficult due to the lack of manually annotated training data. We propose a learning-based method that is only supervised by a self-synthesized modality named local intensity fusion (LIF). LIF is a capillary-enhanced volume computed directly from the input OCT-A. We then construct the local intensity fusion encoder (LIFE) to map a given OCT-A volume and its LIF counterpart to a shared latent space. The latent space of LIFE has the same dimensions as the input data and it contains features common to both modalities. By binarizing this latent space, we obtain a volumetric vessel segmentation. Our method is evaluated in a human fovea OCT-A and three zebrafish OCT-A volumes with manual labels. It yields a Dice score of 0.7736 on human data and 0.8594 +/- 0.0275 on zebrafish data, a dramatic improvement over existing unsupervised algorithms.
Ki67 is a significant biomarker in the diagnosis and prognosis of cancer, whose index can be evaluated by quantifying its expression in Ki67 immunohistochemistry (IHC) stained images. However, quantitative analysis on multi-source Ki67 images is yet
a challenging task in practice due to cross-domain distribution differences, which result from imaging variation, staining styles, and lesion types. Many recent studies have made some efforts on domain generalization (DG), whereas there are still some noteworthy limitations. Specifically in the case of Ki67 images, learning invariant representation is at the mercy of the insufficient number of domains and the cell categories mismatching in different domains. In this paper, we propose a novel method to improve DG by searching the domain-agnostic subnetwork in a domain merging scenario. Partial model parameters are iteratively pruned according to the domain gap, which is caused by the data converting from a single domain into merged domains during training. In addition, the model is optimized by fine-tuning on merged domains to eliminate the interference of class mismatching among various domains. Furthermore, an appropriate implementation is attained by applying the pruning method to different parts of the framework. Compared with known DG methods, our method yields excellent performance in multiclass nucleus recognition of Ki67 IHC images, especially in the lost category cases. Moreover, our competitive results are also evaluated on the public dataset over the state-of-the-art DG methods.
The structure and evolution of protoplanetary disks (PPDs) are largely governed by disk angular momentum transport, mediated by magnetic fields. In the most observable outer disk, PPD gas dynamics is primarily controlled by ambipolar diffusion as the
dominant non-ideal magnetohydrodynamic (MHD) effect. In this work, we study the gas dynamics in outer PPDs by conducting a set of global 3D non-ideal MHD simulations with ambipolar diffusion and net poloidal magnetic flux, using the Athena++ MHD code, with resolution comparable to local simulations. Our simulations demonstrate the co-existence of magnetized disk wind and turbulence driven by the magneto-rotational instability (MRI). While MHD wind dominates disk angular momentum transport, the MRI turbulence also contributes significantly. We observe that magnetic flux spontaneously concentrate into axisymmetric flux sheets, leading to radial variations in turbulence levels, stresses, and accretion rates. Annular substructures arise as a natural consequence of magnetic flux concentration. The flux concentration phenomena show diverse properties with different levels of disk magnetization and ambipolar diffusion. The disk generally loses magnetic flux over time, though flux sheets could prevent the leak of magnetic flux in some cases. Our results demonstrate the ubiquity of disk annular substructures in weakly MRI turbulent outer PPDs, and imply a stochastic nature of disk evolution.
The vertical shear instability (VSI) is a robust phenomenon in irradiated protoplanetary disks (PPDs). While there is extensive literature on the VSI in the hydrodynamic limit, PPDs are expected to be magnetized and their extremely low ionization fra
ctions imply that non-ideal magneto-hydrodynamic (MHD) effects should be properly considered. To this end, we present linear analyses of the VSI in magnetized disks with Ohmic resistivity. We primarily consider toroidal magnetic fields, which are likely to dominate the field geometry in PPDs. We perform vertically global and radially local analyses to capture characteristic VSI modes with extended vertical structures. To focus on the effect of magnetism, we use a locally isothermal equation of state. We find that magnetism provides a stabilizing effect to dampen the VSI, with surface modes, rather than body modes, being the first to vanish with increasing magnetization. Subdued VSI modes can be revived by Ohmic resistivity, where sufficient magnetic diffusion overcome magnetic stabilization, and hydrodynamic results are recovered. We also briefly consider poloidal fields to account for the magnetorotational instability (MRI), which may develop towards surface layers in the outer parts of PPDs. The MRI grows efficiently at small radial wavenumbers, in contrast to the VSI. When resistivity is considered, we find the VSI dominates over the MRI for Ohmic Els{a}sser numbers $lesssim 0.09$ at plasma beta parameter $beta_Z sim 10^4$.
Alphas are stock prediction models capturing trading signals in a stock market. A set of effective alphas can generate weakly correlated high returns to diversify the risk. Existing alphas can be categorized into two classes: Formulaic alphas are sim
ple algebraic expressions of scalar features, and thus can generalize well and be mined into a weakly correlated set. Machine learning alphas are data-driven models over vector and matrix features. They are more predictive than formulaic alphas, but are too complex to mine into a weakly correlated set. In this paper, we introduce a new class of alphas to model scalar, vector, and matrix features which possess the strengths of these two existing classes. The new alphas predict returns with high accuracy and can be mined into a weakly correlated set. In addition, we propose a novel alpha mining framework based on AutoML, called AlphaEvolve, to generate the new alphas. To this end, we first propose operators for generating the new alphas and selectively injecting relational domain knowledge to model the relations between stocks. We then accelerate the alpha mining by proposing a pruning technique for redundant alphas. Experiments show that AlphaEvolve can evolve initial alphas into the new alphas with high returns and weak correlations.
Magnetic skyrmions are exciting candidates for energy-efficient computing due to their non-volatility, detectability,and mobility. A recent proposal within the paradigm of reversible computing enables large-scale circuits composed ofdirectly-cascaded
skyrmion logic gates, but it is limited by the manufacturing difficulty and energy costs associated withthe use of notches for skyrmion synchronization. To overcome these challenges, we therefore propose a skyrmion logicsynchronized via modulation of voltage-controlled magnetic anisotropy (VCMA). In addition to demonstrating theprinciple of VCMA synchronization through micromagnetic simulations, we also quantify the impacts of current den-sity, skyrmion velocity, and anisotropy barrier height on skyrmion motion. Further micromagnetic results demonstratethe feasibility of cascaded logic circuits in which VCMA synchronizers enable clocking and pipelining, illustrating afeasible pathway toward energy-efficient large-scale computing systems based on magnetic skyrmions.
Existing multi-camera SLAM systems assume synchronized shutters for all cameras, which is often not the case in practice. In this work, we propose a generalized multi-camera SLAM formulation which accounts for asynchronous sensor observations. Our fr
amework integrates a continuous-time motion model to relate information across asynchronous multi-frames during tracking, local mapping, and loop closing. For evaluation, we collected AMV-Bench, a challenging new SLAM dataset covering 482 km of driving recorded using our asynchronous multi-camera robotic platform. AMV-Bench is over an order of magnitude larger than previous multi-view HD outdoor SLAM datasets, and covers diverse and challenging motions and environments. Our experiments emphasize the necessity of asynchronous sensor modeling, and show that the use of multiple cameras is critical towards robust and accurate SLAM in challenging outdoor scenes. For additional information, please see the project website at: https://www.cs.toronto.edu/~ajyang/amv-slam
Artificial Intelligence (AI) promises to fundamentally transform society but faces multiple challenges in doing so. In particular, state-of-the-art neuromorphic devices used to implement AI typically lack processes like neuromodulation and neural osc
illations that are critical for enabling many advanced cognitive abilities shown by the brain. Here, we utilize smart materials, that adapt their structure and properties in response to external stimuli, to emulate the modulatory behaviour of neurons called neuromodulation. Leveraging these materials, we have designed and simulated the dynamics of a self-adaptive artificial neuron, which comprises five magnetic skyrmions hosted in a bilayer of thulium iron garnet (TmIG) and platinum (Pt). Micromagnetic simulations show that both the amplitudes and frequencies of neuronal dynamics can be modified by reconfiguring the skyrmion lattice, thereby actualizing neuromodulation. Further, we demonstrate that this neuron achieves a significant advancement over state-of-the-art by realizing the advanced cognitive abilities of context-awareness, cross-frequency coupling as well as information fusion, while utilizing ultra-low power and being ultra-compact. Building advanced cognition into AI can fundamentally transform a wide array of fields including personalized medicine, neuro-prosthesis, human-machine interaction and help realize the next-generation of context-aware AI.
