Search

Tucker decomposition is a powerful tensor model to handle multi-aspect data. It demonstrates the low-rank property by decomposing the grid-structured data as interactions between a core tensor and a set of object representations (factors). A fundamental assumption of such decomposition is that there are finite objects in each aspect or mode, corresponding to discrete indexes of data entries. However, real-world data is often not naturally posed in this setting. For example, geographic data is represented as continuous indexes of latitude and longitude coordinates, and cannot fit tensor models directly. To generalize Tucker decomposition to such scenarios, we propose Functional Bayesian Tucker Decomposition (FunBaT). We treat the continuous-indexed data as the interaction between the Tucker core and a group of latent functions. We use Gaussian processes (GP) as functional priors to model the latent functions. Then, we convert each GP into a state-space prior by constructing an equivalent stochastic differential equation (SDE) to reduce computational cost. An efficient inference algorithm is developed for scalable posterior approximation based on advanced message-passing techniques. The advantage of our method is shown in both synthetic data and several real-world applications. We release the code of FunBaT at \url{https://github.com/xuangu-fang/Functional-Bayesian-Tucker-Decomposition}.

Text-to-3D generation has made remarkable progress recently, particularly with methods based on Score Distillation Sampling (SDS) that leverages pre-trained 2D diffusion models. While the usage of classifier-free guidance is well acknowledged to be crucial for successful optimization, it is considered an auxiliary trick rather than the most essential component. In this paper, we re-evaluate the role of classifier-free guidance in score distillation and discover a surprising finding: the guidance alone is enough for effective text-to-3D generation tasks. We name this method Classifier Score Distillation (CSD), which can be interpreted as using an implicit classification model for generation. This new perspective reveals new insights for understanding existing techniques. We validate the effectiveness of CSD across a variety of text-to-3D tasks including shape generation, texture synthesis, and shape editing, achieving results superior to those of state-of-the-art methods. Our project page is https://xinyu-andy.github.io/Classifier-Score-Distillation, Comment: Our project page is https://xinyu-andy.github.io/Classifier-Score-Distillation

We study the QCD equation of state and other thermodynamic observables including the isentropic trajectories and the speed of sound. These observables are of eminent importance for the understanding of experimental results in heavy ion collisions and also provide a QCD input for studies of the timeline of heavy-ion-collisions with hydrodynamical simulations. They can be derived from the quark propagator whose gap equation is solved within a minimal approximation to the Dyson-Schwinger equations of QCD at finite temperature and density. This minimal approximation aims at a combination of computational efficiency and simplification of the truncation scheme while maintaining quantitative precision. This minimal DSE scheme is confronted and benchmarked with results for correlation functions and observables from first principles QCD lattice at vanishing density and quantitative functional approaches at finite density., Comment: 14 pages, 13 figures

We study the interplay between measurement-induced dynamics and conditional unitary evolution in quantum systems. We numerically and analytically investigate commuting random measurement and feedforward (MFF) processes, and find a sharp transition in their ability to generate entanglement negativity as the number of MFF channels varies. We also establish a direct connection between these findings and transitions induced by random dephasing from an environment with broken time-reversal symmetry. In one variant of the problem, we employ free probability theory to rigorously prove the transition's existence. Furthermore, these MFF processes have dynamic circuit representations that can be experimentally explored on current quantum computing platforms., Comment: Close to final published version. 6 pages, 4 figures (main) + 14 pages, 5 figures (supplementary materials)

We construct the equation of state (EoS) of QCD based on the finite chemical potential information from the functional QCD approaches, with the assistance of the lattice QCD EoS. The obtained EoS is consistent with the up-to-date estimations of the QCD phase diagram, including a phase transition temperature at zero chemical potential of $T=155$ MeV, the curvature of the transition line $\kappa=0.016$ and also a critical end point at $(T,\mu_B)=(118, 600)$ MeV. In specific, the phase diagram mapping is achieved by incorporating the order parameters into the EoS, namely the dynamical quark mass for the chiral phase transition together with the Polyakov loop parameter for the deconfinement phase transition. We also implement the EoS in hydrodynamic simulations to compute the particle yields, ratios and collective flow, and find that our obtained EoS agrees well with the commonly used one based on the combination of lattice QCD simulation and hadron resonance gas model., Comment: 8 pages, 12 figures

Practical tensor data is often along with time information. Most existing temporal decomposition approaches estimate a set of fixed factors for the objects in each tensor mode, and hence cannot capture the temporal evolution of the objects' representation. More important, we lack an effective approach to capture such evolution from streaming data, which is common in real-world applications. To address these issues, we propose Streaming Factor Trajectory Learning for temporal tensor decomposition. We use Gaussian processes (GPs) to model the trajectory of factors so as to flexibly estimate their temporal evolution. To address the computational challenges in handling streaming data, we convert the GPs into a state-space prior by constructing an equivalent stochastic differential equation (SDE). We develop an efficient online filtering algorithm to estimate a decoupled running posterior of the involved factor states upon receiving new data. The decoupled estimation enables us to conduct standard Rauch-Tung-Striebel smoothing to compute the full posterior of all the trajectories in parallel, without the need for revisiting any previous data. We have shown the advantage of SFTL in both synthetic tasks and real-world applications. The code is available at {https://github.com/xuangu-fang/Streaming-Factor-Trajectory-Learning}.

We study the dynamical behaviors of the criterion identifying the first-order phase transition in the matter generated by the relativistic heavy-ion collisions, by explicitly involving the dynamical effects based on the Fokker-Plank framework. The perspectives we taken into account range from phase transition scenarios, initial temperatures, volume effect, relaxation rates, and evolution trajectories. Our numerical calculations show that the dynamical signal of the first-order phase transition can be reserved in certain conditions. Besides the delaying effects due to a finite relaxation time, a larger initial temperature, a smaller volume, a larger relaxation rate, or bending of the trajectory will lead to reduction of the signal. Our discussions on the criterion offer valuable reference information for the experimental detection of the first-order phase transition signal., Comment: 9 pages, 7 figures

We take the Dyson-Schwinger equation (DSE) approach of QCD to study the phase transition and the equation of state of cold dense matter. Besides the bare vertex and Gauss gluon model, we take into account an improved truncation scheme, the CLRQ vertex and infrared-constant gluon model. For the dynamical chiral symmetry breaking solution of the DSE, we require that the emergence of quark number density to be at the chemical potential for the nuclear liquid-gas phase transition to take place, by incorporating a chemical potential dependent modification factor to the gluon model. The result shows that our modified scheme can not only describe the phase transition of the cold dense matter well but also the deduced equation of state of the matter can describe the recent astronomical observations consistently.

LiDAR-based 3D object detection has recently seen significant advancements through active learning (AL), attaining satisfactory performance by training on a small fraction of strategically selected point clouds. However, in real-world deployments where streaming point clouds may include unknown or novel objects, the ability of current AL methods to capture such objects remains unexplored. This paper investigates a more practical and challenging research task: Open World Active Learning for 3D Object Detection (OWAL-3D), aimed at acquiring informative point clouds with new concepts. To tackle this challenge, we propose a simple yet effective strategy called Open Label Conciseness (OLC), which mines novel 3D objects with minimal annotation costs. Our empirical results show that OLC successfully adapts the 3D detection model to the open world scenario with just a single round of selection. Any generic AL policy can then be integrated with the proposed OLC to efficiently address the OWAL-3D problem. Based on this, we introduce the Open-CRB framework, which seamlessly integrates OLC with our preliminary AL method, CRB, designed specifically for 3D object detection. We develop a comprehensive codebase for easy reproducing and future research, supporting 15 baseline methods (\textit{i.e.}, active learning, out-of-distribution detection and open world detection), 2 types of modern 3D detectors (\textit{i.e.}, one-stage SECOND and two-stage PV-RCNN) and 3 benchmark 3D datasets (\textit{i.e.}, KITTI, nuScenes and Waymo). Extensive experiments evidence that the proposed Open-CRB demonstrates superiority and flexibility in recognizing both novel and known classes with very limited labeling costs, compared to state-of-the-art baselines. Source code is available at \url{https://github.com/Luoyadan/CRB-active-3Ddet/tree/Open-CRB}.

Existing volumetric neural rendering techniques, such as Neural Radiance Fields (NeRF), face limitations in synthesizing high-quality novel views when the camera poses of input images are imperfect. To address this issue, we propose a novel 3D reconstruction framework that enables simultaneous optimization of camera poses, dubbed CBARF (Cascaded Bundle-Adjusting NeRF).In a nutshell, our framework optimizes camera poses in a coarse-to-fine manner and then reconstructs scenes based on the rectified poses. It is observed that the initialization of camera poses has a significant impact on the performance of bundle-adjustment (BA). Therefore, we cascade multiple BA modules at different scales to progressively improve the camera poses. Meanwhile, we develop a neighbor-replacement strategy to further optimize the results of BA in each stage. In this step, we introduce a novel criterion to effectively identify poorly estimated camera poses. Then we replace them with the poses of neighboring cameras, thus further eliminating the impact of inaccurate camera poses. Once camera poses have been optimized, we employ a density voxel grid to generate high-quality 3D reconstructed scenes and images in novel views. Experimental results demonstrate that our CBARF model achieves state-of-the-art performance in both pose optimization and novel view synthesis, especially in the existence of large camera pose noise.

In the wheat nutrient deficiencies classification challenge, we present the DividE and EnseMble (DEEM) method for progressive test data predictions. We find that (1) test images are provided in the challenge; (2) samples are equipped with their collection dates; (3) the samples of different dates show notable discrepancies. Based on the findings, we partition the dataset into discrete groups by the dates and train models on each divided group. We then adopt the pseudo-labeling approach to label the test data and incorporate those with high confidence into the training set. In pseudo-labeling, we leverage models ensemble with different architectures to enhance the reliability of predictions. The pseudo-labeling and ensembled model training are iteratively conducted until all test samples are labeled. Finally, the separated models for each group are unified to obtain the model for the whole dataset. Our method achieves an average of 93.6\% Top-1 test accuracy~(94.0\% on WW2020 and 93.2\% on WR2021) and wins the 1$st$ place in the Deep Nutrient Deficiency Challenge~\footnote{https://cvppa2023.github.io/challenges/}.

The rich entanglement dynamics and transitions exhibited by monitored quantum systems typically only exist in the conditional state, making observation extremely difficult. In this work we construct a general recipe for mimicking the conditional entanglement dynamics of a monitored system in a corresponding measurement-free dissipative system involving directional interactions between the original system and a set of auxiliary register modes. This mirror setup autonomously implements a measurement-feedforward dynamics that effectively retains a small fraction of the information content in a typical measurement record. We illustrate our ideas in a bosonic system featuring a competition between entangling measurements and local unitary dynamics, and also discuss extensions to qubit systems and truly many-body systems., Comment: 7+11 pages, 4+7 figures

The application of 3D ViTs to medical image segmentation has seen remarkable strides, somewhat overshadowing the budding advancements in Convolutional Neural Network (CNN)-based models. Large kernel depthwise convolution has emerged as a promising technique, showcasing capabilities akin to hierarchical transformers and facilitating an expansive effective receptive field (ERF) vital for dense predictions. Despite this, existing core operators, ranging from global-local attention to large kernel convolution, exhibit inherent trade-offs and limitations (e.g., global-local range trade-off, aggregating attentional features). We hypothesize that deformable convolution can be an exploratory alternative to combine all advantages from the previous operators, providing long-range dependency, adaptive spatial aggregation and computational efficiency as a foundation backbone. In this work, we introduce 3D DeformUX-Net, a pioneering volumetric CNN model that adeptly navigates the shortcomings traditionally associated with ViTs and large kernel convolution. Specifically, we revisit volumetric deformable convolution in depth-wise setting to adapt long-range dependency with computational efficiency. Inspired by the concepts of structural re-parameterization for convolution kernel weights, we further generate the deformable tri-planar offsets by adapting a parallel branch (starting from $1\times1\times1$ convolution), providing adaptive spatial aggregation across all channels. Our empirical evaluations reveal that the 3D DeformUX-Net consistently outperforms existing state-of-the-art ViTs and large kernel convolution models across four challenging public datasets, spanning various scales from organs (KiTS: 0.680 to 0.720, MSD Pancreas: 0.676 to 0.717, AMOS: 0.871 to 0.902) to vessels (e.g., MSD hepatic vessels: 0.635 to 0.671) in mean Dice., Comment: 14 pages, the source code with our pre-trained model is available at this https://github.com/MASILab/deform-uxnet

Fourier Neural Operator (FNO) is a popular operator learning framework. It not only achieves the state-of-the-art performance in many tasks, but also is efficient in training and prediction. However, collecting training data for the FNO can be a costly bottleneck in practice, because it often demands expensive physical simulations. To overcome this problem, we propose Multi-Resolution Active learning of FNO (MRA-FNO), which can dynamically select the input functions and resolutions to lower the data cost as much as possible while optimizing the learning efficiency. Specifically, we propose a probabilistic multi-resolution FNO and use ensemble Monte-Carlo to develop an effective posterior inference algorithm. To conduct active learning, we maximize a utility-cost ratio as the acquisition function to acquire new examples and resolutions at each step. We use moment matching and the matrix determinant lemma to enable tractable, efficient utility computation. Furthermore, we develop a cost annealing framework to avoid over-penalizing high-resolution queries at the early stage. The over-penalization is severe when the cost difference is significant between the resolutions, which renders active learning often stuck at low-resolution queries and inferior performance. Our method overcomes this problem and applies to general multi-fidelity active learning and optimization problems. We have shown the advantage of our method in several benchmark operator learning tasks. The code is available at https://github.com/shib0li/MRA-FNO.

Large language models (LLMs) have demonstrated remarkable performance on a variety of natural language tasks based on just a few examples of natural language instructions, reducing the need for extensive feature engineering. However, most powerful LLMs are closed-source or limited in their capability for languages other than English. In this technical report, we present Baichuan 2, a series of large-scale multilingual language models containing 7 billion and 13 billion parameters, trained from scratch, on 2.6 trillion tokens. Baichuan 2 matches or outperforms other open-source models of similar size on public benchmarks like MMLU, CMMLU, GSM8K, and HumanEval. Furthermore, Baichuan 2 excels in vertical domains such as medicine and law. We will release all pre-training model checkpoints to benefit the research community in better understanding the training dynamics of Baichuan 2., Comment: Baichuan 2 technical report. Github: https://github.com/baichuan-inc/Baichuan2

Pound-Drever-Hall (PDH) laser frequency stabilization is a powerful technique widely used for building narrow-linewidth lasers. This technique is however ineffective in suppressing high-frequency (>100~kHz) laser phase noise detrimental for many applications. Here, we introduce an effective method which can greatly enhance its high-frequency performance. The idea is to recycle the residual PDH signal of a laser locked to a cavity, by feedforwarding it directly to the laser output field after a delay fiber. Using this straightforward method, we demonstrate a phase noise suppression capability about 4 orders of magnitude better than just using usual PDH feedback for phase noise around a few MHz. We further find that this method exhibits noise suppression performance equivalent to cavity filtering. The new method holds great promise for applications demanding highly stable lasers with diminished phase noise up to tens of MHz, e.g. precise and high-speed control of atomic and molecular quantum states.

Two-dimensional single-slice abdominal computed tomography (CT) provides a detailed tissue map with high resolution allowing quantitative characterization of relationships between health conditions and aging. However, longitudinal analysis of body composition changes using these scans is difficult due to positional variation between slices acquired in different years, which leading to different organs/tissues captured. To address this issue, we propose C-SliceGen, which takes an arbitrary axial slice in the abdominal region as a condition and generates a pre-defined vertebral level slice by estimating structural changes in the latent space. Our experiments on 2608 volumetric CT data from two in-house datasets and 50 subjects from the 2015 Multi-Atlas Abdomen Labeling Challenge dataset (BTCV) Challenge demonstrate that our model can generate high-quality images that are realistic and similar. We further evaluate our method's capability to harmonize longitudinal positional variation on 1033 subjects from the Baltimore Longitudinal Study of Aging (BLSA) dataset, which contains longitudinal single abdominal slices, and confirmed that our method can harmonize the slice positional variance in terms of visceral fat area. This approach provides a promising direction for mapping slices from different vertebral levels to a target slice and reducing positional variance for single-slice longitudinal analysis. The source code is available at: https://github.com/MASILab/C-SliceGen.

We present the dynamics study toward the G326.611+0.811 (G326) hub-filament-system (HFS) cloud using the new APEX observations of both $^{13}$CO and C$^{18}$O (J = 2-1). The G326 HFS cloud constitutes a central hub and at least four hub-composing filaments that are divided into a major branch of filaments (F1, and F2) and a side branch (F3-F5). The cloud holds ongoing high-mass star formation as characterised by three massive dense clumps (i.e., 370-1100 $M_{\odot}$ and 0.14-0.16 g cm$^{-2}$ for C1-C3) with the high clump-averaged mass infalling rates ($>10^{-3}$ $M_{\odot}$ yr$^{-1}$) within in the major filament branch, and the associated point sources bright at 70 $\mu$m typical of young protostars. Along the five filaments, the velocity gradients are found in both $^{13}$CO and C$^{18}$O (J = 2-1) emission, suggesting that the filament-aligned gravitational collapse toward the central hub (i.e., C2) is being at work for high-mass star formation therein. Moreover, a periodic velocity oscillation along the major filament branch is revealed in both $^{13}$CO and C$^{18}$O (J = 2-1) emission with a characteristic wavelength of $\sim$3.5 pc and an amplitude of $\sim$0.31-0.38 km s$^{-1}$. We suggest that this pattern of velocity oscillation in G326 could arise from the clump-forming gas motions induced by gravitational instability. Taking into account the prevalent velocity gradients, the fragmentation of the major branch of filaments, and the ongoing collapse of the three massive dense clumps, it is indicative that G326 is a HFS undergoing global collapse., Comment: 21 pages, 13 figures, 2 tables, Accepted for publication in ApJ

Whole brain segmentation with magnetic resonance imaging (MRI) enables the non-invasive measurement of brain regions, including total intracranial volume (TICV) and posterior fossa volume (PFV). Enhancing the existing whole brain segmentation methodology to incorporate intracranial measurements offers a heightened level of comprehensiveness in the analysis of brain structures. Despite its potential, the task of generalizing deep learning techniques for intracranial measurements faces data availability constraints due to limited manually annotated atlases encompassing whole brain and TICV/PFV labels. In this paper, we enhancing the hierarchical transformer UNesT for whole brain segmentation to achieve segmenting whole brain with 133 classes and TICV/PFV simultaneously. To address the problem of data scarcity, the model is first pretrained on 4859 T1-weighted (T1w) 3D volumes sourced from 8 different sites. These volumes are processed through a multi-atlas segmentation pipeline for label generation, while TICV/PFV labels are unavailable. Subsequently, the model is finetuned with 45 T1w 3D volumes from Open Access Series Imaging Studies (OASIS) where both 133 whole brain classes and TICV/PFV labels are available. We evaluate our method with Dice similarity coefficients(DSC). We show that our model is able to conduct precise TICV/PFV estimation while maintaining the 132 brain regions performance at a comparable level. Code and trained model are available at: https://github.com/MASILab/UNesT/tree/main/wholebrainSeg.

Learning from bounding-boxes annotations has shown great potential in weakly-supervised 3D point cloud instance segmentation. However, we observed that existing methods would suffer severe performance degradation with perturbed bounding box annotations. To tackle this issue, we propose a complementary image prompt-induced weakly-supervised point cloud instance segmentation (CIP-WPIS) method. CIP-WPIS leverages pretrained knowledge embedded in the 2D foundation model SAM and 3D geometric prior to achieve accurate point-wise instance labels from the bounding box annotations. Specifically, CP-WPIS first selects image views in which 3D candidate points of an instance are fully visible. Then, we generate complementary background and foreground prompts from projections to obtain SAM 2D instance mask predictions. According to these, we assign the confidence values to points indicating the likelihood of points belonging to the instance. Furthermore, we utilize 3D geometric homogeneity provided by superpoints to decide the final instance label assignments. In this fashion, we achieve high-quality 3D point-wise instance labels. Extensive experiments on both Scannet-v2 and S3DIS benchmarks demonstrate that our method is robust against noisy 3D bounding-box annotations and achieves state-of-the-art performance.

Despite its fully unitary dynamics, the bosonic Kitaev chain (BKC) displays key hallmarks of non-Hermitian physics including non-reciprocal transport and the non-Hermitian skin effect. Here we demonstrate another remarkable phenomena: the existence of an entanglement phase transition (EPT) in a variant of the BKC that occurs as a function of a Hamiltonian parameter g, and which coincides with a transition from a reciprocal to a non-reciprocal phase. As g is reduced below a critical value, the post-quench entanglement entropy of a subsystem of size l goes from a volume-law phase where it scales as l to a super-volume law phase where it scales like lN with N the total system size. This EPT occurs for a system undergoing purely unitary evolution and does not involve measurements, post-selection, disorder or dissipation. We derive analytically the entanglement entropy out of and at the critical point for the $l=1$ and $l/N \ll 1$ case.

While image diffusion models have made significant progress in text-driven 3D content creation, they often fail to accurately capture the intended meaning of text prompts, especially for view information. This limitation leads to the Janus problem, where multi-faced 3D models are generated under the guidance of such diffusion models. In this paper, we propose a robust high-quality 3D content generation pipeline by exploiting orthogonal-view image guidance. First, we introduce a novel 2D diffusion model that generates an image consisting of four orthogonal-view sub-images based on the given text prompt. Then, the 3D content is created using this diffusion model. Notably, the generated orthogonal-view image provides strong geometric structure priors and thus improves 3D consistency. As a result, it effectively resolves the Janus problem and significantly enhances the quality of 3D content creation. Additionally, we present a 3D synthesis fusion network that can further improve the details of the generated 3D contents. Both quantitative and qualitative evaluations demonstrate that our method surpasses previous text-to-3D techniques. Project page: https://efficientdreamer.github.io.