arXiv subjects

Jun Gao

Publications and source records attributed to Jun Gao.

At least 253 records · Page 14Linked to original sources

arXiv2020

Learning Deformable Tetrahedral Meshes for 3D Reconstruction

3D shape representations that accommodate learning-based 3D reconstruction are an open problem in machine learning and computer graphics. Previous work on neural 3D reconstruction demonstrated benefits, but also limitations, of point cloud, voxel, surface mesh, and implicit function representations. We introduce Deformable Tetrahedral Meshes (DefTet) as a particular parameterization that utilizes volumetric tetrahedral meshes for the reconstruction problem. Unlike existing volumetric approaches, DefTet optimizes for both vertex placement and occupancy, and is differentiable with respect to standard 3D reconstruction loss functions. It is thus simultaneously high-precision, volumetric, and amenable to learning-based neural architectures. We show that it can represent arbitrary, complex topology, is both memory and computationally efficient, and can produce high-fidelity reconstructions with a significantly smaller grid size than alternative volumetric approaches. The predicted surfaces are also inherently defined as tetrahedral meshes, thus do not require post-processing. We demonstrate that DefTet matches or exceeds both the quality of the previous best approaches and the performance of the fastest ones. Our approach obtains high-quality tetrahedral meshes computed directly from noisy point clouds, and is the first to showcase high-quality 3D tet-mesh results using only a single image as input. Our project webpage: https://nv-tlabs.github.io/DefTet/

cs.CV↗

arXiv2020

Constructing higher-order topological states in higher dimension

Higher-order topological phase as a generalization of Berry phase attracts an enormous amount of research. The current theoretical models supporting higher-order topological phases, however, cannot give the connection between lower and higher-order topological phases when extending the lattice from lower to higher dimensions. Here, we theoretically propose and experimentally demonstrate a topological corner state constructed from the edge states in one dimensional lattice. The two-dimensional square lattice owns independent spatial modulation of coupling in each direction, and the combination of edge states in each direction come up to the higher-order topological corner state in two-dimensional lattice, revealing the connection of topological phase in lower and higher dimensional lattices. Moreover, the topological corner states in two-dimensional lattice can also be viewed as the dimension-reduction from a four-dimensional topological phase characterized by vector Chern number, considering two modulation phases as synthetic dimensions in Aubry-Andre-Harper model discussed as example here. Our work deeps the understanding to topological phases breaking through the lattice dimension, and provides a promising tool constructing higher topological phases in higher dimensional structures.

quant-ph↗

arXiv2020

New CTEQ global analysis of quantum chromodynamics with high-precision data from the LHC

We present the new parton distribution functions (PDFs) from the CTEQ-TEA collaboration, obtained using a wide variety of high-precision Large Hadron Collider (LHC) data, in addition to the combined HERA I+II deep-inelastic scattering data set, along with the data sets present in the CT14 global QCD analysis. New LHC measurements in single-inclusive jet production with the full rapidity coverage, as well as production of Drell-Yan pairs, top-quark pairs, and high-$p_T$ $Z$ bosons, are included to achieve the greatest sensitivity to the PDFs. The parton distributions are determined at NLO and NNLO, with each of these PDFs accompanied by error sets determined using the Hessian method. Fast PDF survey techniques, based on the Hessian representation and the Lagrange Multiplier method, are used to quantify the preference of each data set to quantities such as $α_s(m_Z)$, and the gluon and strange quark distributions. We designate the main resulting PDF set as CT18. The ATLAS 7 TeV precision $W/Z$ data are not included in CT18, due to their tension with other data sets in the global fit. Alternate PDF sets are generated including the ATLAS precision 7 TeV $W/Z$ data (CT18A), a new scale choice for low-$x$ DIS data (CT18X), or all of the above with a slightly higher choice for the charm mass (CT18Z). Theoretical calculations of standard candle cross sections at the LHC (such as the $gg$ fusion Higgs boson cross section) are presented.

hep-ph↗

arXiv2020

Modeling of $t$-channel single top-quark production at the LHC

We study the modeling of $t$-channel single top-quark production at Large Hadron Collider (LHC) energies. We compare predictions at next-to-next-to-leading order in a 5-flavor scheme to those of next-to-leading order in a 4-flavor scheme, finding the two schemes agree within a few percent in general for the shape of kinematic distributions of the top quark. The predictions in the 5-flavor scheme show strong stability for both normalization and distributions, and are superior to those of the 4-flavor scheme at comparable orders. We present comparisons of the predictions with LHC data. Our findings provide clear theoretical guidance for precision studies of single top-quark physics at the LHC.

hep-ph↗

arXiv2020

Interactive Annotation of 3D Object Geometry using 2D Scribbles

Inferring detailed 3D geometry of the scene is crucial for robotics applications, simulation, and 3D content creation. However, such information is hard to obtain, and thus very few datasets support it. In this paper, we propose an interactive framework for annotating 3D object geometry from both point cloud data and RGB imagery. The key idea behind our approach is to exploit strong priors that humans have about the 3D world in order to interactively annotate complete 3D shapes. Our framework targets naive users without artistic or graphics expertise. We introduce two simple-to-use interaction modules. First, we make an automatic guess of the 3D shape and allow the user to provide feedback about large errors by drawing scribbles in desired 2D views. Next, we aim to correct minor errors, in which users drag and drop mesh vertices, assisted by a neural interactive module implemented as a Graph Convolutional Network. Experimentally, we show that only a few user interactions are needed to produce good quality 3D shapes on popular benchmarks such as ShapeNet, Pix3D and ScanNet. We implement our framework as a web service and conduct a user study, where we show that user annotated data using our method effectively facilitates real-world learning tasks. Web service: http://www.cs.toronto.edu/~shenti11/scribble3d.

cs.CV↗

arXiv2020

Large Deviation Principle for the Whittaker 2d Growth Model

The Whittaker 2d growth model is a triangular continuous Markov diffusion process that appears in many scientific contexts. It has been theoretically intriguing to establish a large deviation principle for this 2d process with a scaling factor. The main challenge is the spatiotemporal interactions and dynamics that may depend on potential sample-path intersections. We develop such a principle with a novel rate function. Our approach is mainly based on Schider's Theorem, contraction principle, and special treatment for intersecting sample paths.

math.PR↗

arXiv2020

Information Laundering for Model Privacy

In this work, we propose information laundering, a novel framework for enhancing model privacy. Unlike data privacy that concerns the protection of raw data information, model privacy aims to protect an already-learned model that is to be deployed for public use. The private model can be obtained from general learning methods, and its deployment means that it will return a deterministic or random response for a given input query. An information-laundered model consists of probabilistic components that deliberately maneuver the intended input and output for queries to the model, so the model's adversarial acquisition is less likely. Under the proposed framework, we develop an information-theoretic principle to quantify the fundamental tradeoffs between model utility and privacy leakage and derive the optimal design.

cs.CR↗

arXiv2020

Timestamp Boson Sampling

Quantum advantage, benchmarking the computational power of quantum machines outperforming all classical computers in a specific task, represents a crucial milestone in developing quantum computers and has been driving different physical implementations since the concept was proposed. Boson sampling machine, an analog quantum computer that only requires multiphoton interference and single-photon detection, is considered to be a promising candidate to reach this goal. However, the probabilistic nature of photon sources and inevitable loss in evolution network make the execution time exponentially increasing with the problem size. Here, we propose and experimentally demonstrate a timestamp boson sampling that can reduce the execution time by 2 orders of magnitude for any problem size. We theoretically show that the registration time of sampling events can be retrieved to reconstruct the probability distribution at an extremely low-flux rate. By developing a time-of-flight storage technique with a precision up to picosecond level, we are able to detect and record the complete time information of 30 individual modes out of a large-scale 3D photonic chip. We successfully validate boson sampling with only one registered event. We show that it is promptly applicable to fill the remained gap of realizing quantum advantage by timestamp boson sampling. The approach associated with newly exploited resource from time information can boost all the count-rate-limited experiments, suggesting an emerging field of timestamp quantum optics.

quant-ph↗

arXiv2020

Investigating Bottom-Quark Yukawa Interaction at Higgs Factory

Measuring the fermion Yukawa coupling constants is important for understanding the origin of the fermion masses and its relationship to the spontaneously electroweak symmetry breaking. On the other hand, some new physics models will change the Lorentz structure of the Yukawa interactions between the standard model (SM) fermions and the SM-like Higgs boson even in their decoupling limit. Thus the precisely measurement of the fermion Yukawa interactions is a powerful tool of new physics searching in the decoupling limit. In this work, we show the possibility of investigating the Lorentz structure of the bottom-quark Yukawa interaction with the 125GeV SM-like Higgs boson at future $e^+e^-$ colliders.

hep-ph↗

arXiv2020

EPI-based Oriented Relation Networks for Light Field Depth Estimation

Light field cameras record not only the spatial information of observed scenes but also the directions of all incoming light rays. The spatial and angular information implicitly contain geometrical characteristics such as multi-view or epipolar geometry, which can be exploited to improve the performance of depth estimation. An Epipolar Plane Image (EPI), the unique 2D spatial-angular slice of the light field, contains patterns of oriented lines. The slope of these lines is associated with the disparity. Benefiting from this property of EPIs, some representative methods estimate depth maps by analyzing the disparity of each line in EPIs. However, these methods often extract the optimal slope of the lines from EPIs while ignoring the relationship between neighboring pixels, which leads to inaccurate depth map predictions. Based on the observation that an oriented line and its neighboring pixels in an EPI share a similar linear structure, we propose an end-to-end fully convolutional network (FCN) to estimate the depth value of the intersection point on the horizontal and vertical EPIs. Specifically, we present a new feature-extraction module, called Oriented Relation Module (ORM), that constructs the relationship between the line orientations. To facilitate training, we also propose a refocusing-based data augmentation method to obtain different slopes from EPIs of the same scene point. Extensive experiments verify the efficacy of learning relations and show that our approach is competitive to other state-of-the-art methods. The code and the trained models are available at https://github.com/lkyahpu/EPI_ORM.git.

cs.CV↗

arXiv2020

A Dirac-type theorem for uniform hypergraphs

Dirac (1952) proved that every connected graph of order $n>2k+1$ with minimum degree more than $k$ contains a path of length at least $2k+1$. Erdős and Gallai (1959) showed that every $n$-vertex graph $G$ with average degree more than $k-1$ contains a path of length $k$. The hypergraph extension of the Erdős-Gallai Theorem have been given by Győri, Katona, Lemons~(2016) and Davoodi et al.~(2018). Füredi, Kostochka, and Luo (2019) gave a connected version of the Erdős-Gallai Theorem for hypergraphs. In this paper, we give a hypergraph extension of the Dirac's Theorem: Given positive integers $n,k$ and $r$, let $H$ be a connected $n$-vertex $r$-graph with no Berge path of length $2k+1$. We show that (1) If $k> r\ge 4$ and $n>2k+1$, then $δ_1(H)\le\binom{k}{r-1}$. Furthermore, the equality holds if and only if $S'_r(n,k)\subseteq H\subseteq S_r(n,k)$ or $H\cong S(sK_{k+1}^{(r)},1)$; (2) If $k\ge r\ge 2$ and $n>2k(r-1)$, then $δ_1(H)\le \binom{k}{r-1}$. The result is also a Dirac-type version of the result of Füredi, Kostochka, and Luo. As an application of (1), we give a better lower bound of the minimum degree than the ones in the Dirac-type results for Berge Hamiltonian cycle given by Bermond et al.~(1976) and Clemens et al. (2016), respectively.

math.CO↗

arXiv2020

Anchor: Locating Android Framework-specific Crashing Faults

Android framework-specific app crashes are hard to debug. Indeed, the callback-based event-driven mechanism of Android challenges crash localization techniques that are developed for traditional Java programs. The key challenge stems from the fact that the buggy code location may not even be listed within the stack trace. For example, our empirical study on 500 framework-specific crashes from an open benchmark has revealed that 37 percent of the crash types are related to bugs that are outside the stack traces. Moreover, Android programs are a mixture of code and extra-code artifacts such as the Manifest file. The fact that any artifact can lead to failures in the app execution creates the need to position the localization target beyond the code realm. In this paper, we propose Anchor, a two-phase suspicious bug location suggestion tool. Anchor specializes in finding crash-inducing bugs outside the stack trace. Anchor is lightweight and source code independent since it only requires the crash message and the apk file to locate the fault. Experimental results, collected via cross-validation and in-the-wild dataset evaluation, show that Anchor is effective in locating Android framework-specific crashing faults.

cs.SE↗

arXiv2020

Quantum Go Machine

Go has long been considered as a testbed for artificial intelligence. By introducing certain quantum features, such as superposition and collapse of wavefunction, we experimentally demonstrate a quantum version of Go by using correlated photon pairs entangled in polarization degree of freedom. The total dimension of Hilbert space of the generated states grows exponentially as two players take turns to place the stones in time series. As nondeterministic and imperfect information games are more difficult to solve using nowadays technology, we excitedly find that the inherent randomness in quantum physics can bring the game nondeterministic trait, which does not exist in the classical counterpart. Some quantum resources, like coherence or entanglement, can also be encoded to represent the state of quantum stones. Adjusting the quantum resource may vary the average imperfect information (as comparison classical Go is a perfect information game) of a single game. We further verify its non-deterministic feature by showing the unpredictability of the time series data obtained from different classes of quantum state. Finally, by comparing quantum Go with a few typical games that are widely studied in artificial intelligence, we find that quantum Go can cover a wide range of game difficulties rather than a single point. Our results establish a paradigm of inventing new games with quantum-enabled difficulties by harnessing inherent quantum features and resources, and provide a versatile platform for the test of new algorithms to both classical and quantum machine learning.

quant-ph↗

arXiv2020

Two-Dimensional Quantum Walk of Correlated Photons

Quantum walks in an elaborately designed graph, is a powerful tool simulating physical and topological phenomena, constructing analog quantum algorithms and realizing universal quantum computing. Integrated photonics technology has emerged as a versatile platform to implement various quantum information tasks and a promising candidate to perform large-scale quantum walks. Both extending physical dimensions and involving more particles will increase the complexity of the evolving systems and the desired quantum resources. Pioneer works have demonstrated single particle walking on two-dimensional (2D) lattices and multiple walkers interfering on a one-dimensional structure. However, 2D multi-particle quantum walk, genuinely being not classically simulatable, has been a vacancy for nearly ten years. Here, we present a genuine 2D quantum walk with correlated photons on a triangular photonic lattice, which can be mapped to a state space up to 37X37 dimensions. This breaks through the physically restriction of single-particle evolution, which can encode information in a large space and constitute high-dimensional graphs indeed beneficial to quantum information processing. A site-by-site addressing between the chip facet and the 2D fanout interface enables an observation of over 600 non-classical interferences simultaneously, violating a classical limit up to 57 standard deviations. Our platform offers a promising prospect for multi-photon quantum walks in a large-scale 2D arrangement, paving the way for practical quantum simulation and quantum computation beyond classical regime.

quant-ph↗

arXiv2020

Fast Correlated-Photon Imaging Enhanced by Deep Learning

Correlated photon pairs, carrying strong quantum correlations, have been harnessed to bring quantum advantages to various fields from biological imaging to range finding. Such inherent non-classical properties support extracting more valid signals to build photon-limited images even in low flux-level, where the shot noise becomes dominant as light source decreases to single-photon level. Optimization by numerical reconstruction algorithms is possible but require thousands of photon-sparse frames, thus unavailable in real time. Here, we present an experimental fast correlated-photon imaging enhanced by deep learning, showing an intelligent computational strategy to discover deeper structure in big data. Convolutional neural network is found being able to efficiently solve image inverse problems associated with strong shot noise and background noise (electronic noise, scattered light). Our results fill the key gap in incompatibility between imaging speed and image quality by pushing low-light imaging technique to the regime of real-time and single-photon level, opening up an avenue to deep leaning-enhanced quantum imaging for real-life applications.

quant-ph↗

arXiv2020

Protecting Quantum Superposition and Entanglement with Photonic Higher-Order Topological Crystalline Insulator

Higher-order topological insulator, as a newly found non-trivial material and structure, possesses a topological phase beyond the bulk-boundary correspondence. Here, we present an experimental observation of photonic higher-order topological crystalline insulator and its topological protection to quantum superposition and entanglement in a two-dimensional lattice. By freely writing the insulator structure with femtosecond laser and directly measuring evolution dynamics with single-photon imaging techniques, we are able to observe the distinct features of the topological corner states in C_4 and C_2 photonic lattice symmetry. Especially, we propose and experimentally identify the topological corner states by exciting the photonic lattice with single-photon superposition state, and we examine the protection impact of topology on quantum entanglement for entangled photon states. The single-photon dynamics and the protected entanglement reveal an intrinsic topological protection mechanism isolating multi-partite quantum states from diffusion-induced decoherence. The higher-order topological crystalline insulator, built-in superposition state generation, heralded single-photon imaging and quantum entanglement demonstrated here link topology, material, and quantum physics, opening the door to wide investigations of higher-order topology and applications of topological enhancement in genuine quantum regime.

quant-ph↗

arXiv2020

128 Identical Quantum Sources Integrated on a Single Silica Chip

Quantum technology is playing an increasingly important role due to the intrinsic parallel processing capabilities endorsed by quantum superposition, exceeding upper limits of classical performances in diverse fields. Integrated photonic chip offers an elegant way to construct large-scale quantum systems in a physically scalable fashion, however, nonuniformity of quantum sources prevents all the elements from being connected coherently for exponentially increasing Hilbert space. Here, we experimentally demonstrate 128 identical quantum sources integrated on a single silica chip. By actively controlling the light-matter interaction in femtosecond laser direct writing, we are able to unify the properties of waveguides comprehensively and therefore the spontaneous four-wave mixing process for quantum sources. We verify the indistinguishability of the on-chip sources by a series of heralded two-source Hong-Ou-Mandel interference, with all the dip visibilities above 90%. In addition, the brightness of the sources is found easily reaching MHz and being applicable to both discrete-variable and continuous-variable platform, showing either clear anti-bunching feature or large squeezing parameter under different pumping regimes. The demonstrated scalability and uniformity of quantum sources, together with integrated photonic network and detection, will enable large-scale all-on-chip quantum processors for real-life applications.

quant-ph↗

arXiv2020

Decoy-State Quantum Key Distribution over a Long-Distance High-Loss Underwater Free-Space Channel

Atmospheric free space and fiber have been widely exploited as the channels for quantum communication, and have enabled inter-continent and inter-city applications. Air-sea free-space channel, being capable of linking the satellite-based quantum resource and underwater vehicle, has now become the last piece of the puzzle in building global quantum communication network. However, long-distance quantum communication penetrating water up to tens to hundreds of meters is extremely challenging due to the inevitable high loss. Here, we present an experimental demonstration of underwater decoy-state quantum key distribution against high loss, meanwhile keep a low quantum bit error rate less than 2.5% for different distances. By directly modulating blue-green lasers at a high speed of 50MHz and decoy-state protocol, we are able to for the first time reach a long-distance quantum key distribution that is unconditionally secure and can enable real-life air-sea quantum communication tasks. The demonstrated distance, even in coastal water of Jerlov types 2C, is up to 30 meters, about one-order improvement over the proof-in-principle demonstrations in previous experiments, and the channel loss is equivalent to 345-meter-long clean seawater of Jerlov type I, representing a key step forward to practical air-sea quantum communication.

quant-ph↗