arXiv subjects

Yang Zhou

Publications and source records attributed to Yang Zhou.

At least 487 records · Page 27Linked to original sources

arXiv2019

ETNet: Error Transition Network for Arbitrary Style Transfer

Numerous valuable efforts have been devoted to achieving arbitrary style transfer since the seminal work of Gatys et al. However, existing state-of-the-art approaches often generate insufficiently stylized results under challenging cases. We believe a fundamental reason is that these approaches try to generate the stylized result in a single shot and hence fail to fully satisfy the constraints on semantic structures in the content images and style patterns in the style images. Inspired by the works on error-correction, instead, we propose a self-correcting model to predict what is wrong with the current stylization and refine it accordingly in an iterative manner. For each refinement, we transit the error features across both the spatial and scale domain and invert the processed features into a residual image, with a network we call Error Transition Network (ETNet). The proposed model improves over the state-of-the-art methods with better semantic structures and more adaptive style pattern details. Various qualitative and quantitative experiments show that the key concept of both progressive strategy and error-correction leads to better results. Code and models are available at https://github.com/zhijieW94/ETNet.

cs.CV↗

arXiv2019

Platoon trajectories generation: A unidirectional interconnected LSTM-based car following model

Car following models have been widely applied and made remarkable achievements in traffic engineering. However, the traffic micro-simulation accuracy of car following models in a platoon level, especially during traffic oscillations, still needs to be enhanced. Rather than using traditional individual car following models, we proposed a new trajectory generation approach to generate platoon level trajectories given the first leading vehicle's trajectory. In this paper, we discussed the temporal and spatial error propagation issue for the traditional approach by a car following block diagram representation. Based on the analysis, we pointed out that error comes from the training method and the model structure. In order to fix that, we adopt two improvements on the basis of the traditional LSTM based car following model. We utilized a scheduled sampling technique during the training process to solve the error propagation in the temporal dimension. Furthermore, we developed a unidirectional interconnected LSTM model structure to extract trajectories features from the perspective of the platoon. As indicated by the systematic empirical experiments, the proposed novel structure could efficiently reduce the temporal and spatial error propagation. Compared with the traditional LSTM based car following model, the proposed model has almost 40% less error. The findings will benefit the design and analysis of micro-simulation for platoon level car following models.

cs.LG↗

arXiv2019

Optimal Dynamic Futures Portfolio in a Regime-Switching Market Framework

We study the problem of dynamically trading futures in a regime-switching market. Modeling the underlying asset price as a Markov-modulated diffusion process, we present a utility maximization approach to determine the optimal futures trading strategy. This leads to the analysis of the associated system of Hamilton-Jacobi-Bellman (HJB) equations, which are reduced to a system of linear ODEs. We apply our stochastic framework to two models, namely, the Regime-Switching Geometric Brownian Motion (RS-GBM) model and Regime-Switching Exponential Ornstein-Uhlenbeck (RS-XOU) model. Numerical examples are provided to illustrate the investor's optimal futures positions and portfolio value across market regimes.

q-fin.PM↗

arXiv2019

Generalizations of Reflected Entropy and the Holographic Dual

We introduce a new class of quantum and classical correlation measures by generalizing the reflected entropy to multipartite states. We define the new measures for quantum systems in one spatial dimension. For quantum systems having gravity duals, we show that the holographic duals of these new measures are various types of minimal surfaces consist of different entanglement wedge cross sections. One special generalized reflected entropy is $Δ_R$, with the holographic dual proportional to the so called multipartite entanglement wedge cross section $Δ_W$ defined before. We then perform a large $c$ computation of $Δ_R$ and find precise agreement with the holographic computation of 2$Δ_{W}$. This agreement shows another candidate $Δ_R$ as the dual of $Δ_W$ and also supports our holographic conjecture of the new class of generalized reflected entropies.

hep-th↗

arXiv2019

A Top-Down Approach for the Multiple Exercises and Valuation of Employee Stock Options

We propose a new framework to value employee stock options (ESOs) that captures multiple exercises of different quantities over time. We also model the ESO holder's job termination risk and incorporate its impact on the payoffs of both vested and unvested ESOs. Numerical methods based on Fourier transform and finite differences are developed and implemented to solve the associated systems of PDEs. In addition, we introduce a new valuation method based on maturity randomization that yields analytic formulae for vested and unvested ESO costs. We examine the cost impact of job termination risk, exercise intensity, and various contractual features.

q-fin.PR↗

arXiv2019

Predicting Animation Skeletons for 3D Articulated Models via Volumetric Nets

We present a learning method for predicting animation skeletons for input 3D models of articulated characters. In contrast to previous approaches that fit pre-defined skeleton templates or predict fixed sets of joints, our method produces an animation skeleton tailored for the structure and geometry of the input 3D model. Our architecture is based on a stack of hourglass modules trained on a large dataset of 3D rigged characters mined from the web. It operates on the volumetric representation of the input 3D shapes augmented with geometric shape features that provide additional cues for joint and bone locations. Our method also enables intuitive user control of the level-of-detail for the output skeleton. Our evaluation demonstrates that our approach predicts animation skeletons that are much more similar to the ones created by humans compared to several alternatives and baselines.

cs.CV↗

arXiv2019

Entanglement on multiple $S^2$ boundaries in Chern-Simons theory

Topological entanglement structure amongst disjoint torus boundaries of three manifolds have already been studied within the context of Chern-Simons theory. In this work, we study the topological entanglement due to interaction between the quasiparticles inside three-manifolds with one or more disjoint $S^2$ boundaries in SU($N$) Chern-Simons theory. We focus on the world-lines of quasiparticles (Wilson lines), carrying SU($N$) representations, creating four punctures on every $S^2$. We compute the entanglement entropy by partial tracing some of the boundaries. In fact, the entanglement entropy depends on the SU($N$) representations on these four-punctured $S^2$ boundaries. Further, we observe interesting features on the GHZ-like and W-like entanglement structures. Such a distinction crucially depends on the multiplicity of the irreducible representations in the tensor product of SU($N$) representations.

hep-th↗

arXiv2019

SceneGraphNet: Neural Message Passing for 3D Indoor Scene Augmentation

In this paper we propose a neural message passing approach to augment an input 3D indoor scene with new objects matching their surroundings. Given an input, potentially incomplete, 3D scene and a query location, our method predicts a probability distribution over object types that fit well in that location. Our distribution is predicted though passing learned messages in a dense graph whose nodes represent objects in the input scene and edges represent spatial and structural relationships. By weighting messages through an attention mechanism, our method learns to focus on the most relevant surrounding scene context to predict new scene objects. We found that our method significantly outperforms state-of-the-art approaches in terms of correctly predicting objects missing in a scene based on our experiments in the SUNCG dataset. We also demonstrate other applications of our method, including context-based 3D object recognition and iterative scene generation.

cs.CV↗

arXiv2019

Higher-genus wall-crossing in Landau-Ginzburg theory

For a Fermat quasi-homogeneous polynomial, we study the associated weighted Fan-Jarvis-Ruan-Witten theory with narrow insertions. We prove a wall-crossing formula in all genera via localization on a master space, which is constructed by introducing an additional tangent vector to the moduli problem. This is a Landau-Ginzburg theory analogue of the higher-genus quasi-map wall-crossing formula proved by Ciocan-Fontanine and Kim. It generalizes the genus-$0$ result by Ross-Ruan and the genus-$1$ result by Guo-Ross.

math.AG↗

arXiv2019

$3d$ One-form Mixed Anomaly and Entanglement Entropy

We study mixed anomaly between $G_1$ and $G_2$ of one-form finite symmetry $G_1\times G_2$ in $3d$ Chern-Simons theories. We assign a quantum entanglement structure to two linked $G$-symmetry lines (Wilson loops) and compute the entanglement entropy $S[G]$. We find a measure of the mixed anomaly by computing $S[G_1\times G_2]-S[G_1]-S[G_2]$.

hep-th↗

arXiv2018

Entanglement of Purification for Multipartite States and its Holographic Dual

We introduce a new information-theoretic measure of multipartite quantum/classical correlations $Δ_P$, by generalizing the entanglement of purification to multipartite states. We provide proofs of its various properties, focusing on several entropic inequalities, in generic quantum systems. In particular, it turns out that the multipartite entanglement of purification gives an upper bound on multipartite mutual information, which is a generalization of quantum mutual information in the spirit of relative entropy. After that, motivated by a tensor network description of the AdS/CFT correspondence, we also define a holographic dual of multipartite entanglement of purification $Δ_W$, as a sum of minimal areas of codimension-2 surfaces which divide the entanglement wedge into multi-pieces. We prove that this geometrical quantity satisfies all properties we proved for the multipartite entanglement of purification. These agreements strongly support the $Δ_{P}=Δ_{W}$ conjecture. We also show that the multipartite entanglement of purification gives an upper bound on multipartite squashed entanglement, which is a promising measure of multipartite quantum entanglement. We discuss potential saturation of multipartite squashed entanglement onto multipartite mutual information in holographic CFTs and its applications.

hep-th↗

arXiv2018

Simplest Little Higgs Revisited: Hidden Mass Relation, Unitarity and Naturalness

We analyze the scalar potential of the Simplest Little Higgs (SLH) model in an approach consistent with the spirit of continuum effective field theory (CEFT). By requiring correct electroweak symmetry breaking (EWSB) with the $125\,\text{GeV}$ Higgs boson, we are able to derive a relation between the pseudo-axion mass $m_η$ and the heavy top mass $m_T$, which serves as a crucial test of the SLH mechanism. By requiring $m_η^2>0$ an upper bound on $m_T$ can be obtained for any fixed SLH global symmetry breaking scale $f$. We also point out that an absolute upper bound on $f$ can be obtained by imposing partial wave unitarity constraint, which in turn leads to absolute upper bounds of $m_T\lesssim 19\,\text{TeV}, m_η\lesssim 1.5\,\text{TeV}$ and $m_{Z'}\lesssim 48\,\text{TeV}$. We present the allowed region in the three-dimensional parameter space characterized by $f,t_β,m_T$, taking into account the requirement of valid EWSB and the constraint from perturbative unitarity. We also propose a strategy of analyzing the fine-tuning problem consistent with the spirit of CEFT and apply it to the SLH. We suggest that the scalar potential and fine-tuning analysis strategies adopted here should also be applicable to a wide class of Little Higgs and Twin Higgs models, which may reveal interesting relations as crucial tests of the related EWSB mechanism and provide a new perspective on assessing their degree of fine-tuning.

hep-ph↗

arXiv2018

DeepMove: Learning Place Representations through Large Scale Movement Data

Understanding and reasoning about places and their relationships are critical for many applications. Places are traditionally curated by a small group of people as place gazetteers and are represented by an ID with spatial extent, category, and other descriptions. However, a place context is described to a large extent by movements made from/to other places. Places are linked and related to each other by these movements. This important context is missing from the traditional representation. We present DeepMove, a novel approach for learning latent representations of places. DeepMove advances the current deep learning based place representations by directly model movements between places. We demonstrate DeepMove's latent representations on place categorization and clustering tasks on large place and movement datasets with respect to important parameters. Our results show that DeepMove outperforms state-of-the-art baselines. DeepMove's representations can provide up to 15% higher than competing methods in matching rate of place category and result in up to 39% higher silhouette coefficient value for place clusters. DeepMove is spatial and temporal context aware. It is scalable. It outperforms competing models using much smaller training dataset (a month or 1/12 of data). These qualities make it suitable for a broad class of real-world applications.

cs.LG↗

arXiv2018

VisemeNet: Audio-Driven Animator-Centric Speech Animation

We present a novel deep-learning based approach to producing animator-centric speech motion curves that drive a JALI or standard FACS-based production face-rig, directly from input audio. Our three-stage Long Short-Term Memory (LSTM) network architecture is motivated by psycho-linguistic insights: segmenting speech audio into a stream of phonetic-groups is sufficient for viseme construction; speech styles like mumbling or shouting are strongly co-related to the motion of facial landmarks; and animator style is encoded in viseme motion curve profiles. Our contribution is an automatic real-time lip-synchronization from audio solution that integrates seamlessly into existing animation pipelines. We evaluate our results by: cross-validation to ground-truth data; animator critique and edits; visual comparison to recent deep-learning lip-synchronization solutions; and showing our approach to be resilient to diversity in speaker and language.

cs.GR↗

arXiv2018

Full 3D Reconstruction of Transparent Objects

Numerous techniques have been proposed for reconstructing 3D models for opaque objects in past decades. However, none of them can be directly applied to transparent objects. This paper presents a fully automatic approach for reconstructing complete 3D shapes of transparent objects. Through positioning an object on a turntable, its silhouettes and light refraction paths under different viewing directions are captured. Then, starting from an initial rough model generated from space carving, our algorithm progressively optimizes the model under three constraints: surface and refraction normal consistency, surface projection and silhouette consistency, and surface smoothness. Experimental results on both synthetic and real objects demonstrate that our method can successfully recover the complex shapes of transparent objects and faithfully reproduce their light refraction properties.

cs.GR↗

arXiv2018

Non-Stationary Texture Synthesis by Adversarial Expansion

The real world exhibits an abundance of non-stationary textures. Examples include textures with large-scale structures, as well as spatially variant and inhomogeneous textures. While existing example-based texture synthesis methods can cope well with stationary textures, non-stationary textures still pose a considerable challenge, which remains unresolved. In this paper, we propose a new approach for example-based non-stationary texture synthesis. Our approach uses a generative adversarial network (GAN), trained to double the spatial extent of texture blocks extracted from a specific texture exemplar. Once trained, the fully convolutional generator is able to expand the size of the entire exemplar, as well as of any of its sub-blocks. We demonstrate that this conceptually simple approach is highly effective for capturing large-scale structures, as well as other non-stationary attributes of the input exemplar. As a result, it can cope with challenging textures, which, to our knowledge, no other existing method can handle.

cs.GR↗

arXiv2018

Triplet-Center Loss for Multi-View 3D Object Retrieval

Most existing 3D object recognition algorithms focus on leveraging the strong discriminative power of deep learning models with softmax loss for the classification of 3D data, while learning discriminative features with deep metric learning for 3D object retrieval is more or less neglected. In the paper, we study variants of deep metric learning losses for 3D object retrieval, which did not receive enough attention from this area. First , two kinds of representative losses, triplet loss and center loss, are introduced which could learn more discriminative features than traditional classification loss. Then, we propose a novel loss named triplet-center loss, which can further enhance the discriminative power of the features. The proposed triplet-center loss learns a center for each class and requires that the distances between samples and centers from the same class are closer than those from different classes. Extensive experimental results on two popular 3D object retrieval benchmarks and two widely-adopted sketch-based 3D shape retrieval benchmarks consistently demonstrate the effectiveness of our proposed loss, and significant improvements have been achieved compared with the state-of-the-arts.

cs.CV↗

arXiv2018

Entanglement on linked boundaries in Chern-Simons theory with generic gauge groups

We study the entanglement for a state on linked torus boundaries in $3d$ Chern-Simons theory with a generic gauge group and present the asymptotic bounds of Rényi entropy at two different limits: (i) large Chern-Simons coupling $k$, and (ii) large rank $r$ of the gauge group. These results show that the Rényi entropies cannot diverge faster than $\ln k$ and $\ln r$, respectively. We focus on torus links $T(2,2n)$ with topological linking number $n$. The Rényi entropy for these links shows a periodic structure in $n$ and vanishes whenever $n = 0 \text{ (mod } \textsf{p})$, where the integer $\textsf{p}$ is a function of coupling $k$ and rank $r$. We highlight that the refined Chern-Simons link invariants can remove such a periodic structure in $n$.

hep-th↗