Searcharxiv⌕ Search

arXiv subjects

Lei Zhao

Publications and source records attributed to Lei Zhao.

At least 127 records · Page 7Linked to original sources
arXiv2024

PNeSM: Arbitrary 3D Scene Stylization via Prompt-Based Neural Style Mapping

3D scene stylization refers to transform the appearance of a 3D scene to match a given style image, ensuring that images rendered from different viewpoints exhibit the same style as the given style image, while maintaining the 3D consistency of the stylized scene. Several existing methods have obtained impressive results in stylizing 3D scenes. However, the models proposed by these methods need to be re-trained when applied to a new scene. In other words, their models are coupled with a specific scene and cannot adapt to arbitrary other scenes. To address this issue, we propose a novel 3D scene stylization framework to transfer an arbitrary style to an arbitrary scene, without any style-related or scene-related re-training. Concretely, we first map the appearance of the 3D scene into a 2D style pattern space, which realizes complete disentanglement of the geometry and appearance of the 3D scene and makes our model be generalized to arbitrary 3D scenes. Then we stylize the appearance of the 3D scene in the 2D style pattern space via a prompt-based 2D stylization algorithm. Experimental results demonstrate that our proposed framework is superior to SOTA methods in both visual quality and generalization.

cs.CV↗
arXiv2024

Attack Deterministic Conditional Image Generative Models for Diverse and Controllable Generation

Existing generative adversarial network (GAN) based conditional image generative models typically produce fixed output for the same conditional input, which is unreasonable for highly subjective tasks, such as large-mask image inpainting or style transfer. On the other hand, GAN-based diverse image generative methods require retraining/fine-tuning the network or designing complex noise injection functions, which is computationally expensive, task-specific, or struggle to generate high-quality results. Given that many deterministic conditional image generative models have been able to produce high-quality yet fixed results, we raise an intriguing question: is it possible for pre-trained deterministic conditional image generative models to generate diverse results without changing network structures or parameters? To answer this question, we re-examine the conditional image generation tasks from the perspective of adversarial attack and propose a simple and efficient plug-in projected gradient descent (PGD) like method for diverse and controllable image generation. The key idea is attacking the pre-trained deterministic generative models by adding a micro perturbation to the input condition. In this way, diverse results can be generated without any adjustment of network structures or fine-tuning of the pre-trained models. In addition, we can also control the diverse results to be generated by specifying the attack direction according to a reference text or image. Our work opens the door to applying adversarial attack to low-level vision tasks, and experiments on various conditional image generation tasks demonstrate the effectiveness and superiority of the proposed method.

cs.CV↗
arXiv2024

Is Inverse Reinforcement Learning Harder than Standard Reinforcement Learning? A Theoretical Perspective

Inverse Reinforcement Learning (IRL) -- the problem of learning reward functions from demonstrations of an \emph{expert policy} -- plays a critical role in developing intelligent systems. While widely used in applications, theoretical understandings of IRL present unique challenges and remain less developed compared with standard RL. For example, it remains open how to do IRL efficiently in standard \emph{offline} settings with pre-collected data, where states are obtained from a \emph{behavior policy} (which could be the expert policy itself), and actions are sampled from the expert policy. This paper provides the first line of results for efficient IRL in vanilla offline and online settings using polynomial samples and runtime. Our algorithms and analyses seamlessly adapt the pessimism principle commonly used in offline RL, and achieve IRL guarantees in stronger metrics than considered in existing work. We provide lower bounds showing that our sample complexities are nearly optimal. As an application, we also show that the learned rewards can \emph{transfer} to another target MDP with suitable guarantees when the target MDP satisfies certain similarity assumptions with the original (source) MDP.

stat.ML↗
arXiv2024

CogVLM: Visual Expert for Pretrained Language Models

We introduce CogVLM, a powerful open-source visual language foundation model. Different from the popular shallow alignment method which maps image features into the input space of language model, CogVLM bridges the gap between the frozen pretrained language model and image encoder by a trainable visual expert module in the attention and FFN layers. As a result, CogVLM enables deep fusion of vision language features without sacrificing any performance on NLP tasks. CogVLM-17B achieves state-of-the-art performance on 10 classic cross-modal benchmarks, including NoCaps, Flicker30k captioning, RefCOCO, RefCOCO+, RefCOCOg, Visual7W, GQA, ScienceQA, VizWiz VQA and TDIUC, and ranks the 2nd on VQAv2, OKVQA, TextVQA, COCO captioning, etc., surpassing or matching PaLI-X 55B. Codes and checkpoints are available at https://github.com/THUDM/CogVLM.

cs.CV↗
arXiv2024

Do as I can, not as I get

This paper proposes a model called TMR to mine valuable information from simulated data environments. We intend to complete the submission of this paper.

cs.AI↗
arXiv2023

Emulators in JINSP

JINSP(Jiutian Intelligence Network Simulation Platform) describes a series of basic emulators and their combinations, such as the simulation of the protocol stack for dynamic users in a real environment, which is composed of user behavior simulation, base station simulation, and terminal simulation. It is applied in specific business scenarios, such as multi-target antenna optimization, compression feedback, and so on. This paper provides detailed descriptions of each emulator and its combination based on this foundation, including the implementation process of the emulator, integration with the platform, experimental results, and other aspects.

cs.NI↗
arXiv2023

Implement services for business scenarios by combining basic emulators

This article mainly introduces how to use various basic emulators to form a combined emulator in the Jiutian Intelligence Network Simulation Platform to realize simulation service functions in different business scenarios. Among them, the combined emulator is included. The business scenarios include different practical applications such as multi-objective antenna optimization, high traffic of business, CSI (channel state information) compression feedback, etc.

cs.AI↗
arXiv2023

ArtBank: Artistic Style Transfer with Pre-trained Diffusion Model and Implicit Style Prompt Bank

Artistic style transfer aims to repaint the content image with the learned artistic style. Existing artistic style transfer methods can be divided into two categories: small model-based approaches and pre-trained large-scale model-based approaches. Small model-based approaches can preserve the content strucuture, but fail to produce highly realistic stylized images and introduce artifacts and disharmonious patterns; Pre-trained large-scale model-based approaches can generate highly realistic stylized images but struggle with preserving the content structure. To address the above issues, we propose ArtBank, a novel artistic style transfer framework, to generate highly realistic stylized images while preserving the content structure of the content images. Specifically, to sufficiently dig out the knowledge embedded in pre-trained large-scale models, an Implicit Style Prompt Bank (ISPB), a set of trainable parameter matrices, is designed to learn and store knowledge from the collection of artworks and behave as a visual prompt to guide pre-trained large-scale models to generate highly realistic stylized images while preserving content structure. Besides, to accelerate training the above ISPB, we propose a novel Spatial-Statistical-based self-Attention Module (SSAM). The qualitative and quantitative experiments demonstrate the superiority of our proposed method over state-of-the-art artistic style transfer methods.

cs.CV↗
arXiv2023

Geometric properties of integrable Kepler and Hooke billiards with conic section boundaries

We study the geometry of reflection of a massive point-like particle at conic section boundaries. Thereby the particle is subjected to a central force associated with either a Kepler or Hooke potential. The conic section is assumed to have a focus at the Kepler center, or have its center at the Hookian center respectively. When the particle hits the boundary it is ideally reflected according to the law of reflection. These systems are known to be integrable. We describe the consecutive billiard orbits in terms of their foci. We show that the second foci of these orbits always lie on a circle in the Kepler case. In the Hooke case, we show that the foci of the orbits lie on a Cassini oval. For both systems we analyze the envelope of the directrices of the orbits as well.

math.DS↗
arXiv2023

Principle of fundamental resonance in hypersonic boundary layers: an asymptotic viewpoint

The fundamental resonance (FR) in the nonlinear phase of the boundary-layer transition to turbulence appears when a dominant planar instability mode reaches a finite amplitude, and the low-amplitude oblique traveling modes with the same frequency as the dominant mode, together with the stationary streak modes, undergo the strongest amplification among all the Fourier components. This regime may be the most efficient means to trigger the natural transition in hypersonic boundary layers. In this paper, we aim to reveal the intrinsic mechanism of the FR in the weakly nonlinear framework based on the large-Reynolds-number asymptotic technique. It is found that the FR is in principle a triad resonance among a dominant planar fundamental mode, a streak mode and an oblique mode. In the major part of the boundary layer, the nonlinear interaction of the fundamental mode and the streak mode seeds for the growth of the oblique mode, whereas the interaction of the oblique mode and the fundamental mode drives the roll components (transverse and lateral velocity) of the streak mode, which leads to a stronger amplification of the streamwise component of the streak mode due to the lift-up mechanism. The main-layer solution of the streamwise velocity, spanwise velocity and temperature of both the streak and the oblique modes become singular as the wall is approached, and so a viscous wall layer appears underneath. The wall layer produces an outflux velocity to the main-layer solution, inclusion of which leads to an improved asymptotic theory, whose accuracy is confirmed by comparing with the calculations of the nonlinear parabolised stability equations (NPSE) at moderate Reynolds numbers and the secondary instability analysis (SIA) at sufficiently high Reynolds numbers.

physics.flu-dyn↗
arXiv2023

Boltzmann's Billiard Systems: Computation of the Billiard Mapping and Some Numerical Results

L. Boltzmann proposed a billiard model with a planar central force problem reflected against a line not passing through the center. He asserted that such a system is ergodic, which thus illustrates his ergodic hypothesis. However, it has been recently shown that when the underlying central force problem is the Kepler problem then the system is actually integrable by Gallavotti-Jauslin. This raises the question of whether Boltzmann's assertion holds true for some central force problems that he considered. In this paper, we present some geometrical and numerical analysis on the dynamics of several of these systems. As indicated by the numerics, many of these systems show chaotic dynamics and a system seems to be ergodic.

math.DS↗
arXiv2023

Concave Toric Domains in Stark-type Mechanical Systems

It has been shown in [Frauenfelder, 2023] that the bounded component of the energy surface of the planer Stark problem after the Levi-Civita transformation are concave toric domains. In this paper, we present a different approach on the problem of determining concave toric domains in a family of integrable natural mechanical problems in the plane based on the computation of action-variables. We give criteria on the potentials for the bounded components of the energy hypersurfaces to be concave toric domains and apply these criteria to a class of problems.

math.DS↗
arXiv2023

Kustaanheimo-Stiefel Transformation, Birkhoff-Waldvogel Transformation and Integrable Mechanical Billiards

The three-dimensional Kepler problem is related to the four-dimensional isotropic harmonic oscillators by the Kustaanheimo-Stiefel Transformations. In the first part of this paper, we study how certain integrable mechanical billiards are related by this transformation. This in part illustrates the rotation-invariance of integrable reflection walls in the three-dimensional Kepler billiards found till so far. The second part of this paper deals with Birkhoff-Waldvogel's Transformation of the three-dimensional problem with two Kepler centers. In particular, we establish an analogous theory of Levi-Civita planes for Birkhoff-Waldvogel's Transformation and showed the integrability of certain three-dimensional 2-center billiards via a different approach.

math.DS↗
arXiv2023

An unprecedented synergy of high-temperature tensile strength and ductility in a NiCoCrAlTi high-entropy alloy

The present work reported a novel L12-strengthening NiCoCrAlTi high entropy alloy (HEA) with an outstanding synergy of tensile strength and ductility at both ambient and high temperatures. Transmission electron microscopy (TEM) characterization revealed a high density of rod-like and spheroidal L12 precipitates distributing in the micro/nanograins and non-recrystallized regions in the annealed specimens. The tremendously high yield stress, ultimate tensile stress (UTS), and ductility of the HEA at 600 C were ~1060 MPa, 1271 MPa, and 25%, respectively, which were significantly superior to most reported HEAs and Co- and Ni-based superalloys to date. Systematic TEM analysis unveiled that the cooperation among L12 precipitation, extensive stacking faults (SFs), deformation twins (DTs), immobile Lomer-Cottrell (L-C) locks formed from interactions between SFs and SFs/DTs, hierarchical SFs/DTs networks, as well as hetero-deformation-induced strengthening dominated the plastic deformation at 600 C. Such a unique deformation mechanism enabled extremely high tensile strength and sustained ductility of the HEA at a high temperature.

cond-mat.mtrl-sci↗
arXiv2023

X-HRNet: Towards Lightweight Human Pose Estimation with Spatially Unidimensional Self-Attention

High-resolution representation is necessary for human pose estimation to achieve high performance, and the ensuing problem is high computational complexity. In particular, predominant pose estimation methods estimate human joints by 2D single-peak heatmaps. Each 2D heatmap can be horizontally and vertically projected to and reconstructed by a pair of 1D heat vectors. Inspired by this observation, we introduce a lightweight and powerful alternative, Spatially Unidimensional Self-Attention (SUSA), to the pointwise (1x1) convolution that is the main computational bottleneck in the depthwise separable 3c3 convolution. Our SUSA reduces the computational complexity of the pointwise (1x1) convolution by 96% without sacrificing accuracy. Furthermore, we use the SUSA as the main module to build our lightweight pose estimation backbone X-HRNet, where `X' represents the estimated cross-shape attention vectors. Extensive experiments on the COCO benchmark demonstrate the superiority of our X-HRNet, and comprehensive ablation studies show the effectiveness of the SUSA modules. The code is publicly available at https://github.com/cool-xuan/x-hrnet.

cs.CV↗
arXiv2023

Design of JiuTian Intelligent Network Simulation Platform

This paper introduced the JiuTian Intelligent Network Simulation Platform, which can provide wireless communication simulation data services for the Open Innovation Platform. The platform contains a series of scalable simulator functionalities, offering open services that enable users to use reinforcement learning algorithms for model training and inference based on simulation environments and data. Additionally, it allows users to address optimization tasks in different scenarios by uploading and updating parameter configurations. The platform and its open services were primarily introduced from the perspectives of background, overall architecture, simulator, business scenarios, and future directions.

cs.NI↗
arXiv2023

The trigger system for the CSR external-target experiment

A trigger system has been designed and implemented for the HIRFL-CSR external target experiment (CEE), the spectrometer for studying nuclear matter properties with heavy ion collisions in the GeV energy region. The system adopts master-slave structure and serial data transmission mode using optical fiber to deal with different types of detectors and long-distance signal transmission. The trigger logic can be accessed based on command register and controlled by a remote computer. The overall field programmable gate array (FPGA) logic can be flexibly reconfigured online to match the physical requirements of the experiment. The trigger system has been tested in beam experiment. It is demonstrated that the trigger system functions correctly and meets the physical requirements of CEE.

physics.ins-det↗
arXiv2023

TSSAT: Two-Stage Statistics-Aware Transformation for Artistic Style Transfer

Artistic style transfer aims to create new artistic images by rendering a given photograph with the target artistic style. Existing methods learn styles simply based on global statistics or local patches, lacking careful consideration of the drawing process in practice. Consequently, the stylization results either fail to capture abundant and diversified local style patterns, or contain undesired semantic information of the style image and deviate from the global style distribution. To address this issue, we imitate the drawing process of humans and propose a Two-Stage Statistics-Aware Transformation (TSSAT) module, which first builds the global style foundation by aligning the global statistics of content and style features and then further enriches local style details by swapping the local statistics (instead of local features) in a patch-wise manner, significantly improving the stylization effects. Moreover, to further enhance both content and style representations, we introduce two novel losses: an attention-based content loss and a patch-based style loss, where the former enables better content preservation by enforcing the semantic relation in the content image to be retained during stylization, and the latter focuses on increasing the local style similarity between the style and stylized images. Extensive qualitative and quantitative experiments verify the effectiveness of our method.

cs.CV↗