arXiv subjects

Jie Jiang

Publications and source records attributed to Jie Jiang.

At least 127 records · Page 7Linked to original sources

arXiv2025

Hunyuan3D 2.5: Towards High-Fidelity 3D Assets Generation with Ultimate Details

In this report, we present Hunyuan3D 2.5, a robust suite of 3D diffusion models aimed at generating high-fidelity and detailed textured 3D assets. Hunyuan3D 2.5 follows two-stages pipeline of its previous version Hunyuan3D 2.0, while demonstrating substantial advancements in both shape and texture generation. In terms of shape generation, we introduce a new shape foundation model -- LATTICE, which is trained with scaled high-quality datasets, model-size, and compute. Our largest model reaches 10B parameters and generates sharp and detailed 3D shape with precise image-3D following while keeping mesh surface clean and smooth, significantly closing the gap between generated and handcrafted 3D shapes. In terms of texture generation, it is upgraded with phyiscal-based rendering (PBR) via a novel multi-view architecture extended from Hunyuan3D 2.0 Paint model. Our extensive evaluation shows that Hunyuan3D 2.5 significantly outperforms previous methods in both shape and end-to-end texture generation.

cs.CV↗

arXiv2025

Hunyuan3D 2.1: From Images to High-Fidelity 3D Assets with Production-Ready PBR Material

3D AI-generated content (AIGC) is a passionate field that has significantly accelerated the creation of 3D models in gaming, film, and design. Despite the development of several groundbreaking models that have revolutionized 3D generation, the field remains largely accessible only to researchers, developers, and designers due to the complexities involved in collecting, processing, and training 3D models. To address these challenges, we introduce Hunyuan3D 2.1 as a case study in this tutorial. This tutorial offers a comprehensive, step-by-step guide on processing 3D data, training a 3D generative model, and evaluating its performance using Hunyuan3D 2.1, an advanced system for producing high-resolution, textured 3D assets. The system comprises two core components: the Hunyuan3D-DiT for shape generation and the Hunyuan3D-Paint for texture synthesis. We will explore the entire workflow, including data preparation, model architecture, training strategies, evaluation metrics, and deployment. By the conclusion of this tutorial, you will have the knowledge to finetune or develop a robust 3D generative model suitable for applications in gaming, virtual reality, and industrial design.

cs.CV↗

arXiv2025

CAdam: Confidence-Based Optimization for Online Learning

Modern recommendation systems frequently employ online learning to dynamically update their models with freshly collected data. The most commonly used optimizer for updating neural networks in these contexts is the Adam optimizer, which integrates momentum ($m_t$) and adaptive learning rate ($v_t$). However, the volatile nature of online learning data, characterized by its frequent distribution shifts and presence of noise, poses significant challenges to Adam's standard optimization process: (1) Adam may use outdated momentum and the average of squared gradients, resulting in slower adaptation to distribution changes, and (2) Adam's performance is adversely affected by data noise. To mitigate these issues, we introduce CAdam, a confidence-based optimization strategy that assesses the consistency between the momentum and the gradient for each parameter dimension before deciding on updates. If momentum and gradient are in sync, CAdam proceeds with parameter updates according to Adam's original formulation; if not, it temporarily withholds updates and monitors potential shifts in data distribution in subsequent iterations. This method allows CAdam to distinguish between the true distributional shifts and mere noise, and to adapt more quickly to new data distributions. In various settings with distribution shift or noise, our experiments demonstrate that CAdam surpasses other well-known optimizers, including the original Adam. Furthermore, in large-scale A/B testing within a live recommendation system, CAdam significantly enhances model performance compared to Adam, leading to substantial increases in the system's gross merchandise volume (GMV).

cs.LG↗

arXiv2025

Scaling Laws for Floating Point Quantization Training

Low-precision training is considered an effective strategy for reducing both training and downstream inference costs. Previous scaling laws for precision mainly focus on integer quantization, which pay less attention to the constituents in floating-point (FP) quantization, and thus cannot well fit the LLM losses in this scenario. In contrast, while FP quantization training is more commonly implemented in production, it's research has been relatively superficial. In this paper, we thoroughly explore the effects of FP quantization targets, exponent bits, mantissa bits, and the calculation granularity of the scaling factor in FP quantization training performance of LLM models. In addition to an accurate FP quantization unified scaling law, we also provide valuable suggestions for the community: (1) Exponent bits contribute slightly more to the model performance than mantissa bits. We provide the optimal exponent-mantissa bit ratio for different bit numbers, which is available for future reference by hardware manufacturers; (2) We discover the formation of the critical data size in low-precision LLM training. Too much training data exceeding the critical data size will inversely bring in degradation of LLM performance; (3) The optimal FP quantization precision is directly proportional to the computational power, but within a wide computational power range. We estimate that the best cost-performance precision should lie between 4-8 bits.

cs.LG↗

arXiv2025

Voyager: Long-Range and World-Consistent Video Diffusion for Explorable 3D Scene Generation

Real-world applications like video gaming and virtual reality often demand the ability to model 3D scenes that users can explore along custom camera trajectories. While significant progress has been made in generating 3D objects from text or images, creating long-range, 3D-consistent, explorable 3D scenes remains a complex and challenging problem. In this work, we present Voyager, a novel video diffusion framework that generates world-consistent 3D point-cloud sequences from a single image with user-defined camera path. Unlike existing approaches, Voyager achieves end-to-end scene generation and reconstruction with inherent consistency across frames, eliminating the need for 3D reconstruction pipelines (e.g., structure-from-motion or multi-view stereo). Our method integrates three key components: 1) World-Consistent Video Diffusion: A unified architecture that jointly generates aligned RGB and depth video sequences, conditioned on existing world observation to ensure global coherence 2) Long-Range World Exploration: An efficient world cache with point culling and an auto-regressive inference with smooth video sampling for iterative scene extension with context-aware consistency, and 3) Scalable Data Engine: A video reconstruction pipeline that automates camera pose estimation and metric depth prediction for arbitrary videos, enabling large-scale, diverse training data curation without manual 3D annotations. Collectively, these designs result in a clear improvement over existing methods in visual quality and geometric accuracy, with versatile applications.

cs.CV↗

arXiv2025

Solar Surface Magnetic Field Simulation from 2010 to 2024 and Anomalous Southern Poleward Flux Transport in Cycle 24

The solar surface magnetic field is fundamental for modeling the coronal magnetic field, studying the solar dynamo, and predicting solar cycle strength. We perform a continuous simulation of the surface magnetic field from 2010 to 2024, covering solar cycle 24 and the ongoing cycle 25, using the surface flux transport model with assimilated observed active regions (ARs) as the source. The simulation reproduces the evolution of the axial dipole strength, polar field reversal timing, and magnetic butterfly diagram in good agreement with SDO/HMI observations. Notably, these results are achieved without incorporating radial diffusion or cyclic variations in meridional flow speed, suggesting their limited impact. Poleward surges of the following polarity typically dominate throughout the cycle, but in the southern hemisphere during cycle 24, they are limited to a short period from 2011 to 2016. This anomalous pattern arises from intermittent AR emergence, with about 46% of total unsigned flux contributed by ARs emerging during Carrington Rotations 2141-2160 (September 2013 - February 2015). These ARs show a strong active longitude at Carrington longitudes 200-260 degree and a weaker one at 80-100 degree. After 2016, poleward migrations of leading-polarity flux become dominant, despite most ARs following Joy's and Hale's laws. This reversal is likely due to prolonged intervals between AR emergences, which allow leading-polarity flux to distribute across a broad latitude range before cancellation by subsequent ARs. These findings highlight the importance of the temporal interval of AR emergence in driving the flux transport pattern.

astro-ph.SR↗

arXiv2025

Hunyuan-Game: Industrial-grade Intelligent Game Creation Model

Intelligent game creation represents a transformative advancement in game development, utilizing generative artificial intelligence to dynamically generate and enhance game content. Despite notable progress in generative models, the comprehensive synthesis of high-quality game assets, including both images and videos, remains a challenging frontier. To create high-fidelity game content that simultaneously aligns with player preferences and significantly boosts designer efficiency, we present Hunyuan-Game, an innovative project designed to revolutionize intelligent game production. Hunyuan-Game encompasses two primary branches: image generation and video generation. The image generation component is built upon a vast dataset comprising billions of game images, leading to the development of a group of customized image generation models tailored for game scenarios: (1) General Text-to-Image Generation. (2) Game Visual Effects Generation, involving text-to-effect and reference image-based game visual effect generation. (3) Transparent Image Generation for characters, scenes, and game visual effects. (4) Game Character Generation based on sketches, black-and-white images, and white models. The video generation component is built upon a comprehensive dataset of millions of game and anime videos, leading to the development of five core algorithmic models, each targeting critical pain points in game development and having robust adaptation to diverse game video scenarios: (1) Image-to-Video Generation. (2) 360 A/T Pose Avatar Video Synthesis. (3) Dynamic Illustration Generation. (4) Generative Video Super-Resolution. (5) Interactive Game Video Generation. These image and video generation models not only exhibit high-level aesthetic expression but also deeply integrate domain-specific knowledge, establishing a systematic understanding of diverse game and anime art styles.

cs.CV↗

arXiv2025

Sample Path Large Deviations for Random Walks on Regular Trees

This paper investigates the large deviation problem in the sample path space of the nearest-neighbor random walks on regular trees. We establish the sample path large deviation principle for the law of the distance from a nearest random walk on a regular tree to the root with a good convex rate function. Furthermore, we derive an implicit expression for the rate function via the Fenchel-Legendre transform of the log-moment generating function.

math.PR↗

arXiv2025

Observation-Based Iterative Map for Solar Cycles. I. Nature of Solar Cycle Variability

Inter-cycle variations in the series of 11-year solar activity cycles have a significant impact on both the space environment and climate. Whether solar cycle variability is dominated by deterministic chaos or stochastic perturbations remains an open question. Distinguishing between the two mechanisms is crucial for predicting solar cycles. Here we reduce the solar dynamo process responsible for the solar cycle to a one-dimensional iterative map, incorporating recent advance in the observed nonlinearity and stochasticity of the cycle. We demonstrate that deterministic chaos is absent in the nonlinear system, regardless of model parameters, if the generation of the poloidal field follows an increase-then-saturate pattern as the cycle strength increases. The synthesized solar cycles generated by the iterative map exhibit a probability density function (PDF) similar to that of observed normal cycles, supporting the dominant role of stochasticity in solar cycle variability. The parameters governing nonlinearity and stochasticity profoundly influence the PDF. The iterative map provides a quick and effective tool for predicting the range, including uncertainty of the subsequent cycle strength when an ongoing cycle amplitude is known. Due to stochasticity, a solar cycle loses almost all its original information within 1 or 2 cycles. Although the simplicity of the iterative map, the behaviors it exhibits are generic for the nonlinear system. Our results provide guidelines for analyzing solar dynamo models in terms of chaos and stochasticity, highlight the limitation in predicting solar cycle, and motivate further refinement of observational constraints on nonlinear and stochastic processes.

astro-ph.SR↗

arXiv2025

Pre-train, Align, and Disentangle: Empowering Sequential Recommendation with Large Language Models

Sequential Recommendation (SR) aims to leverage the sequential patterns in users' historical interactions to accurately track their preferences. However, the primary reliance of existing SR methods on collaborative data results in challenges such as the cold-start problem and sub-optimal performance. Concurrently, despite the proven effectiveness of large language models (LLMs), their integration into commercial recommender systems is impeded by issues such as high inference latency, incomplete capture of all distribution statistics, and catastrophic forgetting. To address these issues, we introduce a novel Pre-train, Align, and Disentangle (PAD) framework to enhance SR models with LLMs. In particular, we initially pre-train both the SR and LLM models to obtain collaborative and textual embeddings. Subsequently, we propose a characteristic recommendation-anchored alignment loss using multi-kernel maximum mean discrepancy with Gaussian kernels. Lastly, a triple-experts architecture, comprising aligned and modality-specific experts with disentangled embeddings, is fine-tuned in a frequency-aware manner. Experimental results on three public datasets validate the efficacy of PAD, indicating substantial enhancements and compatibility with various SR backbone models, particularly for cold items. The code and datasets are accessible for reproduction at https://github.com/Applied-Machine-Learning-Lab/PAD.

cs.IR↗

arXiv2025

Dark-Matter Induced Scalarization of Black Holes in Extended Scalar-Tensor-Gauss-Bonnet Theories

In the extended scalar-tensor-Gauss-Bonnet theory, spontaneous scalarization in the GB$^-$ regime typically occurs only in rotating black holes, while it is absent in spherically symmetric black holes, a phenomenon known as spin-induced scalarization. However, we find that when the spacetime is permeated by perfect fluid dark matter, spontaneous scalarization can also be induced by dark matter in the GB$^-$ regime. Analytical calculations reveal that this scalarization occurs when the dark matter parameter $b/M$ exceeds a critical value $(b/M)_\text{crit}\simeq1.86287$, a threshold determined by the lower boundary of the unstable region for scalar perturbations as the coupling constant approaches negative infinity. Additionally, we verified these findings through numerical analysis of the time evolution of scalar perturbations, identifying the unstable parameter region. The results show that when coupling constant $-λ/M^2$ is small, spontaneous scalarization only occurs near the extremal black hole limit. As $-λ/M^2$ increases, the scalarization region expands; however, its lower boundary remains above $b/M \simeq 1.86287$, consistent with theoretical predictions.

gr-qc↗

arXiv2025

Second Law of Black Holes Thermodynamics in Non-Quasi-Static Processes

The Bekenstein-Hawking entropy satisfies the generalized second law of black hole thermodynamics for arbitrary thermodynamic evolution within Einstein-Maxwell theory. In contrast, the black hole entropy that satisfies the second law in low-energy effective modified gravity theories related to quantum gravity is derived solely by considering linear perturbations within quasi-static processes. Since astrophysical processes are typically non-quasi-static, deriving a rigorous expression for entropy within the quasi-static framework is not feasible. By treating the effective quantum corrections as first-order perturbations, the black hole entropy in Einstein-Maxwell gravity with generalized quadratic corrections for arbitrary dynamical processes is derived. This entropy is distinct from the Iyer-Wald and Dong-Wald entropies and is applicable to non-quasi-static processes. Furthermore, it is shown that the black hole entropy is consistent with the generalized covariant entropy boundary. This work establishes a framework for examining the generalized second law of black hole thermodynamics in non-static processes within modified gravity theories.

gr-qc↗

arXiv2025

DRAMA: A Dynamic Packet Routing Algorithm using Multi-Agent Reinforcement Learning with Emergent Communication

The continuous expansion of network data presents a pressing challenge for conventional routing algorithms. As the demand escalates, these algorithms are struggling to cope. In this context, reinforcement learning (RL) and multi-agent reinforcement learning (MARL) algorithms emerge as promising solutions. However, the urgency and importance of the problem are clear, as existing RL/MARL-based routing approaches lack effective communication in run time among routers, making it challenging for individual routers to adapt to complex and dynamic changing networks. More importantly, they lack the ability to deal with dynamically changing network topology, especially the addition of the router, due to the non-scalability of their neural networks. This paper proposes a novel dynamic routing algorithm, DRAMA, incorporating emergent communication in multi-agent reinforcement learning. Through emergent communication, routers could learn how to communicate effectively to maximize the optimization objectives. Meanwhile, a new Q-network and graph-based emergent communication are introduced to dynamically adapt to the changing network topology without retraining while ensuring robust performance. Experimental results showcase DRAMA's superior performance over the traditional routing algorithm and other RL/MARL-based algorithms, achieving a higher delivery rate and lower latency in diverse network scenarios, including dynamic network load and topology. Moreover, an ablation experiment validates the prospect of emergent communication in facilitating packet routing.

cs.NI↗

arXiv2025

Numerical Study of Wheeler-Dewitt Equation beyond Slow-roll approximation

The Wheeler-DeWitt (WDW) equation is analyzed using two boundary proposals: the Hartle-Hawking no-boundary condition and tunneling condition. By compactifying the scale factor $a$ into $ x = a/(1+a) $, we reformulate the WDW equation to find stable numerical solutions with clearer boundary conditions. The no-boundary wave function peaks at the horizon scale, indicating quantum nucleation of classical spacetime, while the tunneling solution shows exponential decay, reflecting vacuum decay from a classically forbidden state. These dynamics are explored under slow-roll and non-slow-roll regimes of a periodic potential, separately, with non-slow-roll scenarios amplifying quantum effects that delay the classical behavior. The results emphasize the role of boundary conditions in quantum cosmology, offering insights into the universe's origin and the interplay between quantum gravity and observable cosmology.

gr-qc↗

arXiv2025

Long-Sequence Recommendation Models Need Decoupled Embeddings

Lifelong user behavior sequences are crucial for capturing user interests and predicting user responses in modern recommendation systems. A two-stage paradigm is typically adopted to handle these long sequences: a subset of relevant behaviors is first searched from the original long sequences via an attention mechanism in the first stage and then aggregated with the target item to construct a discriminative representation for prediction in the second stage. In this work, we identify and characterize, for the first time, a neglected deficiency in existing long-sequence recommendation models: a single set of embeddings struggles with learning both attention and representation, leading to interference between these two processes. Initial attempts to address this issue with some common methods (e.g., linear projections -- a technique borrowed from language processing) proved ineffective, shedding light on the unique challenges of recommendation models. To overcome this, we propose the Decoupled Attention and Representation Embeddings (DARE) model, where two distinct embedding tables are initialized and learned separately to fully decouple attention and representation. Extensive experiments and analysis demonstrate that DARE provides more accurate searches of correlated behaviors and outperforms baselines with AUC gains up to 0.9% on public datasets and notable improvements on Tencent's advertising platform. Furthermore, decoupling embedding spaces allows us to reduce the attention embedding dimension and accelerate the search procedure by 50% without significant performance impact, enabling more efficient, high-performance online serving. Code in PyTorch for experiments, including model analysis, is available at https://github.com/thuml/DARE.

cs.IR↗

arXiv2025

Unleashing Vecset Diffusion Model for Fast Shape Generation

3D shape generation has greatly flourished through the development of so-called "native" 3D diffusion, particularly through the Vecset Diffusion Model (VDM). While recent advancements have shown promising results in generating high-resolution 3D shapes, VDM still struggles with high-speed generation. Challenges exist because of difficulties not only in accelerating diffusion sampling but also VAE decoding in VDM, areas under-explored in previous works. To address these challenges, we present FlashVDM, a systematic framework for accelerating both VAE and DiT in VDM. For DiT, FlashVDM enables flexible diffusion sampling with as few as 5 inference steps and comparable quality, which is made possible by stabilizing consistency distillation with our newly introduced Progressive Flow Distillation. For VAE, we introduce a lightning vecset decoder equipped with Adaptive KV Selection, Hierarchical Volume Decoding, and Efficient Network Design. By exploiting the locality of the vecset and the sparsity of shape surface in the volume, our decoder drastically lowers FLOPs, minimizing the overall decoding overhead. We apply FlashVDM to Hunyuan3D-2 to obtain Hunyuan3D-2 Turbo. Through systematic evaluation, we show that our model significantly outperforms existing fast 3D generation methods, achieving comparable performance to the state-of-the-art while reducing inference time by over 45x for reconstruction and 32x for generation. Code and models are available at https://github.com/Tencent/FlashVDM.

cs.CV↗

arXiv2025

BIMII-Net: Brain-Inspired Multi-Iterative Interactive Network for RGB-T Road Scene Semantic Segmentation

RGB-T road scene semantic segmentation enhances visual scene understanding in complex environments characterized by inadequate illumination or occlusion by fusing information from RGB and thermal images. Nevertheless, existing RGB-T semantic segmentation models typically depend on simple addition or concatenation strategies or ignore the differences between information at different levels. To address these issues, we proposed a novel RGB-T road scene semantic segmentation network called Brain-Inspired Multi-Iteration Interaction Network (BIMII-Net). First, to meet the requirements of accurate texture and local information extraction in road scenarios like autonomous driving, we proposed a deep continuous-coupled neural network (DCCNN) architecture based on a brain-inspired model. Second, to enhance the interaction and expression capabilities among multi-modal information, we designed a cross explicit attention-enhanced fusion module (CEAEF-Module) in the feature fusion stage of BIMII-Net to effectively integrate features at different levels. Finally, we constructed a complementary interactive multi-layer decoder structure, incorporating the shallow-level feature iteration module (SFI-Module), the deep-level feature iteration module (DFI-Module), and the multi-feature enhancement module (MFE-Module) to collaboratively extract texture details and global skeleton information, with multi-module joint supervision further optimizing the segmentation results. Experimental results demonstrate that BIMII-Net achieves state-of-the-art (SOTA) performance in the brain-inspired computing domain and outperforms most existing RGB-T semantic segmentation methods. It also exhibits strong generalization capabilities on multiple RGB-T datasets, proving the effectiveness of brain-inspired computer models in multi-modal image segmentation tasks.

cs.CV↗

arXiv2025

RomanTex: Decoupling 3D-aware Rotary Positional Embedded Multi-Attention Network for Texture Synthesis

Painting textures for existing geometries is a critical yet labor-intensive process in 3D asset generation. Recent advancements in text-to-image (T2I) models have led to significant progress in texture generation. Most existing research approaches this task by first generating images in 2D spaces using image diffusion models, followed by a texture baking process to achieve UV texture. However, these methods often struggle to produce high-quality textures due to inconsistencies among the generated multi-view images, resulting in seams and ghosting artifacts. In contrast, 3D-based texture synthesis methods aim to address these inconsistencies, but they often neglect 2D diffusion model priors, making them challenging to apply to real-world objects To overcome these limitations, we propose RomanTex, a multiview-based texture generation framework that integrates a multi-attention network with an underlying 3D representation, facilitated by our novel 3D-aware Rotary Positional Embedding. Additionally, we incorporate a decoupling characteristic in the multi-attention block to enhance the model's robustness in image-to-texture task, enabling semantically-correct back-view synthesis. Furthermore, we introduce a geometry-related Classifier-Free Guidance (CFG) mechanism to further improve the alignment with both geometries and images. Quantitative and qualitative evaluations, along with comprehensive user studies, demonstrate that our method achieves state-of-the-art results in texture quality and consistency.

cs.CV↗