arXiv subjects

Jie Wu

Publications and source records attributed to Jie Wu.

At least 127 records · Page 7Linked to original sources

arXiv2025

Seedream 3.0 Technical Report

We present Seedream 3.0, a high-performance Chinese-English bilingual image generation foundation model. We develop several technical improvements to address existing challenges in Seedream 2.0, including alignment with complicated prompts, fine-grained typography generation, suboptimal visual aesthetics and fidelity, and limited image resolutions. Specifically, the advancements of Seedream 3.0 stem from improvements across the entire pipeline, from data construction to model deployment. At the data stratum, we double the dataset using a defect-aware training paradigm and a dual-axis collaborative data-sampling framework. Furthermore, we adopt several effective techniques such as mixed-resolution training, cross-modality RoPE, representation alignment loss, and resolution-aware timestep sampling in the pre-training phase. During the post-training stage, we utilize diversified aesthetic captions in SFT, and a VLM-based reward model with scaling, thereby achieving outputs that well align with human preferences. Furthermore, Seedream 3.0 pioneers a novel acceleration paradigm. By employing consistent noise expectation and importance-aware timestep sampling, we achieve a 4 to 8 times speedup while maintaining image quality. Seedream 3.0 demonstrates significant improvements over Seedream 2.0: it enhances overall capabilities, in particular for text-rendering in complicated Chinese characters which is important to professional typography generation. In addition, it provides native high-resolution output (up to 2K), allowing it to generate images with high visual quality.

cs.CV↗

arXiv2025

Seedance 1.0: Exploring the Boundaries of Video Generation Models

Notable breakthroughs in diffusion modeling have propelled rapid improvements in video generation, yet current foundational model still face critical challenges in simultaneously balancing prompt following, motion plausibility, and visual quality. In this report, we introduce Seedance 1.0, a high-performance and inference-efficient video foundation generation model that integrates several core technical improvements: (i) multi-source data curation augmented with precision and meaningful video captioning, enabling comprehensive learning across diverse scenarios; (ii) an efficient architecture design with proposed training paradigm, which allows for natively supporting multi-shot generation and jointly learning of both text-to-video and image-to-video tasks. (iii) carefully-optimized post-training approaches leveraging fine-grained supervised fine-tuning, and video-specific RLHF with multi-dimensional reward mechanisms for comprehensive performance improvements; (iv) excellent model acceleration achieving ~10x inference speedup through multi-stage distillation strategies and system-level optimizations. Seedance 1.0 can generate a 5-second video at 1080p resolution only with 41.4 seconds (NVIDIA-L20). Compared to state-of-the-art video generation models, Seedance 1.0 stands out with high-quality and fast video generation having superior spatiotemporal fluidity with structural stability, precise instruction adherence in complex multi-subject contexts, native multi-shot narrative coherence with consistent subject representation.

cs.CV↗

arXiv2025

New weighted Alexandrov-Fenchel type inequalities and Minkowski inequalities in space forms

In this paper, we establish a broad class of new sharp Alexandrov-Fenchel inequalities involving general convex weight functions for static convex hypersurfaces in hyperbolic space. Additionally, we derive new weighted Minkowski-type inequalities for static convex hypersurfaces in hyperbolic space $\mathbb{H}^n$ and for convex hypersurfaces in the sphere $\mathbb{S}^n$. The tools we shall use are the locally constrained inverse curvature flows in hyperbolic space and in the sphere.

math.DG↗

arXiv2025

PeRL: Permutation-Enhanced Reinforcement Learning for Interleaved Vision-Language Reasoning

Inspired by the impressive reasoning capabilities demonstrated by reinforcement learning approaches like DeepSeek-R1, recent emerging research has begun exploring the use of reinforcement learning (RL) to enhance vision-language models (VLMs) for multimodal reasoning tasks. However, most existing multimodal reinforcement learning approaches remain limited to spatial reasoning within single-image contexts, yet still struggle to generalize to more complex and real-world scenarios involving multi-image positional reasoning, where understanding the relationships across images is crucial. To address this challenge, we propose a general reinforcement learning approach PeRL tailored for interleaved multimodal tasks, and a multi-stage strategy designed to enhance the exploration-exploitation trade-off, thereby improving learning efficiency and task performance. Specifically, we introduce permutation of image sequences to simulate varied positional relationships to explore more spatial and positional diversity. Furthermore, we design a rollout filtering mechanism for resampling to focus on trajectories that contribute most to learning optimal behaviors to exploit learned policies effectively. We evaluate our model on 5 widely-used multi-image benchmarks and 3 single-image benchmarks. Our experiments confirm that PeRL trained model consistently surpasses R1-related and interleaved VLM baselines by a large margin, achieving state-of-the-art performance on multi-image benchmarks, while preserving comparable performance on single-image tasks.

cs.CV↗

arXiv2025

Towards Seamless Borders: A Method for Mitigating Inconsistencies in Image Inpainting and Outpainting

Image inpainting is the task of reconstructing missing or damaged parts of an image in a way that seamlessly blends with the surrounding content. With the advent of advanced generative models, especially diffusion models and generative adversarial networks, inpainting has achieved remarkable improvements in visual quality and coherence. However, achieving seamless continuity remains a significant challenge. In this work, we propose two novel methods to address discrepancy issues in diffusion-based inpainting models. First, we introduce a modified Variational Autoencoder that corrects color imbalances, ensuring that the final inpainted results are free of color mismatches. Second, we propose a two-step training strategy that improves the blending of generated and existing image content during the diffusion process. Through extensive experiments, we demonstrate that our methods effectively reduce discontinuity and produce high-quality inpainting results that are coherent and visually appealing.

cs.CV↗

arXiv2025

Step-Audio-AQAA: a Fully End-to-End Expressive Large Audio Language Model

Large Audio-Language Models (LALMs) have significantly advanced intelligent human-computer interaction, yet their reliance on text-based outputs limits their ability to generate natural speech responses directly, hindering seamless audio interactions. To address this, we introduce Step-Audio-AQAA, a fully end-to-end LALM designed for Audio Query-Audio Answer (AQAA) tasks. The model integrates a dual-codebook audio tokenizer for linguistic and semantic feature extraction, a 130-billion-parameter backbone LLM and a neural vocoder for high-fidelity speech synthesis. Our post-training approach employs interleaved token-output of text and audio to enhance semantic coherence and combines Direct Preference Optimization (DPO) with model merge to improve performance. Evaluations on the StepEval-Audio-360 benchmark demonstrate that Step-Audio-AQAA excels especially in speech control, outperforming the state-of-art LALMs in key areas. This work contributes a promising solution for end-to-end LALMs and highlights the critical role of token-based vocoder in enhancing overall performance for AQAA tasks.

cs.SD↗

arXiv2025

Chain of Methodologies: Scaling Test Time Computation without Training

Large Language Models (LLMs) often struggle with complex reasoning tasks due to insufficient in-depth insights in their training data, which are typically absent in publicly available documents. This paper introduces the Chain of Methodologies (CoM), an innovative and intuitive prompting framework that enhances structured thinking by integrating human methodological insights, enabling LLMs to tackle complex tasks with extended reasoning. CoM leverages the metacognitive abilities of advanced LLMs, activating systematic reasoning throught user-defined methodologies without explicit fine-tuning. Experiments show that CoM surpasses competitive baselines, demonstrating the potential of training-free prompting methods as robust solutions for complex reasoning tasks and bridging the gap toward human-level reasoning through human-like methodological insights.

cs.CL↗

arXiv2025

Intertwined nematic and d-wave superconductive orders in optimally-doped La1.84Sr0.16CuO4

The anisotropy of the superconducting state and superconducting fluctuations in the CuO2 plane is directly related to the superconducting mechanism of copper oxide superconductors and is therefore pivotal for understanding high-temperature superconductivity. Here, we integrated the high-precision angle-resolved resistivity (ARR) measurement with a rotatable in-plane magnetic field to systematically study the angular dependence of superconducting fluctuations in optimally doped La1.84Sr0.16CuO4 (LSCO). By independently controlling the directions of the current and the magnetic field, we are able to isolate the magneto-resistivity contributed by the superconducting vortex motion and distinguish excitations from nematic superconductivity and d-wave superconductive order based on their respective C2 and C4 symmetries. Signatures of two intertwined superconductive orders are also evident in the measured angular dependence of the critical current. A T-B phase diagram of different types of superconducting fluctuations is determined. These findings are closely related to other intriguing phenomena, such as pair density wave and charge density wave.

cond-mat.supr-con↗

arXiv2025

Co-designed Quantum Discrete Adiabatic Linear System Solver Via Dynamic Circuits

Existing quantum discrete adiabatic approaches are hindered by circuit depth that increases linearly with the number of evolution steps, a significant challenge for current quantum hardware with limited coherence times. To address this, we propose a co-designed framework that synergistically integrates dynamic circuit capabilities with real-time classical processing. This framework reformulates the quantum adiabatic evolution into discrete, dynamically adjustable segments. The unitary operator for each segment is optimized on-the-fly using classical computation, and circuit multiplexing techniques are leveraged to reduce the overall circuit depth scaling from $O(\text{steps}\times\text{depth}(U))$ to $O(\text{depth}(U))$. We implement and benchmark a quantum discrete adiabatic linear solver based on this framework for linear systems of $W \in \{2,4,8,16\}$ dimensions with condition numbers $κ\in \{10,20,30,40,50\}$. Our solver successfully overcomes previous depth limitations, maintaining over 80% solution fidelity even under realistic noise models. Key algorithmic optimizations contributing to this performance include a first-order approximation of the discrete evolution operator, a tailored dynamic circuit design exploiting real-imaginary component separation, and noise-resilient post-processing techniques.

quant-ph↗

arXiv2025

Some results on the k-strong parity property in a graph

A graph $G$ has the $k$-strong parity property if for any $X\subseteq V(G)$ with $|X|$ even, $G$ contains a spanning subgraph $F$ with $d_F(u)\equiv1$ (mod 2) for each $u\in X$ and $d_F(v)\in\{k,k+2,k+4,\ldots\}$ for each $v\in V(G)\setminus X$, where $k\geq2$ is an even integer. Kano and Matsumura proposed a characterization for a graph with the $k$-strong parity property (M. Kano, H. Matsumura, Odd-even factors of graphs, Graphs Combin. 41 (2025) 55). In this paper, we first give a size condition for a graph to have the $k$-strong parity property. Then we establish a signless Laplacian spectral radius condition to guarantee that a graph has the $k$-strong parity property.

math.CO↗

arXiv2025

A Comprehensive Evaluation of Large Language Models on Temporal Event Forecasting

Recently, Large Language Models (LLMs) have demonstrated great potential in various data mining tasks, such as knowledge question answering, mathematical reasoning, and commonsense reasoning. However, the reasoning capability of LLMs on temporal event forecasting has been under-explored. To systematically investigate their abilities in temporal event forecasting, we conduct a comprehensive evaluation of LLM-based methods for temporal event forecasting. Due to the lack of a high-quality dataset that involves both graph and textual data, we first construct a benchmark dataset, named MidEast-TE-mini. Based on this dataset, we design a series of baseline methods, characterized by various input formats and retrieval augmented generation (RAG) modules. From extensive experiments, we find that directly integrating raw texts into the input of LLMs does not enhance zero-shot extrapolation performance. In contrast, fine-tuning LLMs with raw texts can significantly improve performance. Additionally, LLMs enhanced with retrieval modules can effectively capture temporal relational patterns hidden in historical events. However, issues such as popularity bias and the long-tail problem persist in LLMs, particularly in the retrieval-augmented generation (RAG) method. These findings not only deepen our understanding of LLM-based event forecasting methods but also highlight several promising research directions. We consider that this comprehensive evaluation, along with the identified research opportunities, will significantly contribute to future research on temporal event forecasting through LLMs.

cs.CL↗

arXiv2025

Quantum Lattice Kinetic Scheme for Solving Two-dimensional and Three-dimensional Incompressible Flows

Lattice Boltzmann method (LBM) is particularly well-suited for implementation on quantum circuits owing to its simple algebraic operations and natural parallelism. However, most quantum LBMs fix $τ$ = 1 to avoid nonlinear collision, which restricts the simulation to a fixed mesh size for a given Reynolds number. To preserve the simplicity of setting $τ$ = 1 while enhancing flexibility, we propose a quantum lattice kinetic scheme (LKS) by introducing a constant parameter $A$ into the equilibrium distribution function (EDF), enabling independent adjustment of the fluid's viscosity. This modification removes the constraint on mesh size, making it possible to simulate flows with arbitrary Reynolds numbers. The Chapman-Enskog analysis confirms the modified EDF still recovers the Navier-Stokes equations without compromising collision accuracy. We evaluate the method on 2D and 3D Taylor-Green vortex and lid-driven cavity flows, demonstrating that quantum LKS attains the same accuracy and convergence order as classical LKS. The first application of quantum LBM to 3D incompressible flows represents a significant step forward in large-scale fluid dynamics simulation.

quant-ph↗

arXiv2025

Mogao: An Omni Foundation Model for Interleaved Multi-Modal Generation

Recent progress in unified models for image understanding and generation has been impressive, yet most approaches remain limited to single-modal generation conditioned on multiple modalities. In this paper, we present Mogao, a unified framework that advances this paradigm by enabling interleaved multi-modal generation through a causal approach. Mogao integrates a set of key technical improvements in architecture design, including a deep-fusion design, dual vision encoders, interleaved rotary position embeddings, and multi-modal classifier-free guidance, which allow it to harness the strengths of both autoregressive models for text generation and diffusion models for high-quality image synthesis. These practical improvements also make Mogao particularly effective to process interleaved sequences of text and images arbitrarily. To further unlock the potential of unified models, we introduce an efficient training strategy on a large-scale, in-house dataset specifically curated for joint text and image generation. Extensive experiments show that Mogao not only achieves state-of-the-art performance in multi-modal understanding and text-to-image generation, but also excels in producing high-quality, coherent interleaved outputs. Its emergent capabilities in zero-shot image editing and compositional generation highlight Mogao as a practical omni-modal foundation model, paving the way for future development and scaling the unified multi-modal systems.

cs.CV↗

arXiv2025

Cohomotopy Sets of $(n-1)$-connected $(2n+2)$-manifolds for small $n$

Let $M$ be a closed orientable $(n-1)$-connected $(2n+2)$-manifold, $n\geq 2$. In this paper we combine the Postnikov tower of spheres and the homotopy decomposition of the reduced suspension space $ΣM$ to investigate the cohomotopy sets $π^\ast(M)$ for $n=2,3,4$, under the assumption that $M$ has $2$-torsion-free homology. All cohomotopy sets $π^i(M)$ of such manifolds $M$ are characterized except $π^4(M)$ for $n=3,4$.

math.AT↗

arXiv2025

A machine learning model for skillful climate system prediction

Climate system models (CSMs), through integrating cross-sphere interactions among the atmosphere, ocean, land, and cryosphere, have emerged as pivotal tools for deciphering climate dynamics and improving forecasting capabilities. Recent breakthroughs in artificial intelligence (AI)-driven meteorological modeling have demonstrated remarkable success in single-sphere systems and partially spheres coupled systems. However, the development of a fully coupled AI-based climate system model encompassing atmosphere-ocean-land-sea ice interactions has remained an unresolved challenge. This paper introduces FengShun-CSM, an AI-based CSM model that provides 60-day global daily forecasts for 29 critical variables across atmospheric, oceanic, terrestrial, and cryospheric domains. The model significantly outperforms the European Centre for Medium-Range Weather Forecasts (ECMWF) subseasonal-to-seasonal (S2S) model in predicting most variables, particularly precipitation, land surface, and oceanic components. This enhanced capability is primarily attributed to its improved representation of intra-seasonal variability modes, most notably the Madden-Julian Oscillation (MJO). Remarkably, FengShun-CSM exhibits substantial potential in predicting subseasonal extreme events. Such breakthroughs will advance its applications in meteorological disaster mitigation, marine ecosystem conservation, and agricultural productivity enhancement. Furthermore, it validates the feasibility of developing AI-powered CSMs through machine learning technologies, establishing a transformative paradigm for next-generation Earth system modeling.

cs.LG↗

arXiv2025

Homotopy Groups and Puppe Sequence of Digraphs

We introduce homotopy groups of digraphs that admit an intuitive description of grid structures, which is a variation of the GLMY homotopy groups introduced by Grigor'yan, Lin, Muranov and Yau in 2014. This direct approach enables a descriptive interpretation of GLMY theory in applications such as network science. Furthermore, we prove that there exists a long exact sequence of homotopy groups of digraphs associated to any based digraph map, that is, there exists a digraph version of the Puppe sequence.

math.AT↗

arXiv2025

How to systematically develop an effective AI-based bias correction model?

This study introduces ReSA-ConvLSTM, an artificial intelligence (AI) framework for systematic bias correction in numerical weather prediction (NWP). We propose three innovations by integrating dynamic climatological normalization, ConvLSTM with temporal causality constraints, and residual self-attention mechanisms. The model establishes a physics-aware nonlinear mapping between ECMWF forecasts and ERA5 reanalysis data. Using 41 years (1981-2021) of global atmospheric data, the framework reduces systematic biases in 2-m air temperature (T2m), 10-m winds (U10/V10), and sea-level pressure (SLP), achieving up to 20% RMSE reduction over 1-7 day forecasts compared to operational ECMWF outputs. The lightweight architecture (10.6M parameters) enables efficient generalization to multiple variables and downstream applications, reducing retraining time by 85% for cross-variable correction while improving ocean model skill through bias-corrected boundary conditions. The ablation experiments demonstrate that our innovations significantly improve the model's correction performance, suggesting that incorporating variable characteristics into the model helps enhance forecasting skills.

cs.LG↗

arXiv2025

Efficient Distributed Retrieval-Augmented Generation for Enhancing Language Model Performance

Small language models (SLMs) support efficient deployments on resource-constrained edge devices, but their limited capacity compromises inference performance. Retrieval-augmented generation (RAG) is a promising solution to enhance model performance by integrating external databases, without requiring intensive on-device model retraining. However, large-scale public databases and user-specific private contextual documents are typically located on the cloud and the device separately, while existing RAG implementations are primarily centralized. To bridge this gap, we propose DRAGON, a distributed RAG framework to enhance on-device SLMs through both general and personal knowledge without the risk of leaking document privacy. Specifically, DRAGON decomposes multi-document RAG into multiple parallel token generation processes performed independently and locally on the cloud and the device, and employs a newly designed Speculative Aggregation, a dual-side speculative algorithm to avoid frequent output synchronization between the cloud and device. A new scheduling algorithm is further introduced to identify the optimal aggregation side based on real-time network conditions. Evaluations on real-world hardware testbed demonstrate a significant performance improvement of DRAGON-up to 1.9x greater gains over standalone SLM compared to the centralized RAG, substantial reduction in per-token latency, and negligible Time to First Token (TTFT) overhead.

cs.LG↗