arXiv subjects

Jie Chen

Publications and source records attributed to Jie Chen.

At least 145 records · Page 8Linked to original sources

arXiv2025

SOSControl: Enhancing Human Motion Generation through Saliency-Aware Symbolic Orientation and Timing Control

Traditional text-to-motion frameworks often lack precise control, and existing approaches based on joint keyframe locations provide only positional guidance, making it challenging and unintuitive to specify body part orientations and motion timing. To address these limitations, we introduce the Salient Orientation Symbolic (SOS) script, a programmable symbolic framework for specifying body part orientations and motion timing at keyframes. We further propose an automatic SOS extraction pipeline that employs temporally-constrained agglomerative clustering for frame saliency detection and a Saliency-based Masking Scheme (SMS) to generate sparse, interpretable SOS scripts directly from motion data. Moreover, we present the SOSControl framework, which treats the available orientation symbols in the sparse SOS script as salient and prioritizes satisfying these constraints during motion generation. By incorporating SMS-based data augmentation and gradient-based iterative optimization, the framework enhances alignment with user-specified constraints. Additionally, it employs a ControlNet-based ACTOR-PAE Decoder to ensure smooth and natural motion outputs. Extensive experiments demonstrate that the SOS extraction pipeline generates human-interpretable scripts with symbolic annotations at salient keyframes, while the SOSControl framework outperforms existing baselines in motion quality, controllability, and generalizability with respect to motion timing and body part orientation control.

cs.CV↗

arXiv2025

JW-Flare: Accurate Solar Flare Forecasting Method Based on Multimodal Large Language Models

Solar flares, the most powerful explosive phenomena in the solar system, may pose significant hazards to spaceborne satellites and ground-based infrastructure. Despite decades of intensive research, reliable flare prediction remains a challenging task. Large Language Models, as a milestone in artificial intelligence, exhibit exceptional general knowledge and next-token prediction capabilities. Here we introduce JW-Flare, the first Multimodal Large Language Models (MLLMs) explicitly trained for solar flare forecasting through fine-tuning on textual physic parameters of solar active regions and magnetic field images. This method demonstrates state-of-the-art (SOTA) performance for large flares prediction on the test dataset. It effectively identifies all 79 X-class flares from 18,949 test samples, yielding a True Skill Statistic (TSS) of 0.95 and a True Positive Rate (TPR) of 1.00, outperforming traditional predictive models. We further investigate the capability origins of JW-Flare through explainability experiments, revealing that solar physics knowledge acquired during pre-training contributes to flare forecasting performance. Additionally, we evaluate models of different parameter scales, confirming the Scaling_Law of Large Language Models in domain-specific applications, such as solar physics. This study marks a substantial advance in both the scale and accuracy of solar flare forecasting and opens a promising avenue for AI-driven methodologies in broader scientific domains.

astro-ph.SR↗

arXiv2025

Deep Inverse Shading: Consistent Albedo and Surface Detail Recovery via Generative Refinement

Reconstructing human avatars using generative priors is essential for achieving versatile and realistic avatar models. Traditional approaches often rely on volumetric representations guided by generative models, but these methods require extensive volumetric rendering queries, leading to slow training. Alternatively, surface-based representations offer faster optimization through differentiable rasterization, yet they are typically limited by vertex count, restricting mesh resolution and scalability when combined with generative priors. Moreover, integrating generative priors into physically based human avatar modeling remains largely unexplored. To address these challenges, we introduce DIS (Deep Inverse Shading), a unified framework for high-fidelity, relightable avatar reconstruction that incorporates generative priors into a coherent surface representation. DIS centers on a mesh-based model that serves as the target for optimizing both surface and material details. The framework fuses multi-view 2D generative surface normal predictions, rich in detail but often inconsistent, into the central mesh using a normal conversion module. This module converts generative normal outputs into per-triangle surface offsets via differentiable rasterization, enabling the capture of fine geometric details beyond sparse vertex limitations. Additionally, DIS integrates a de-shading module to recover accurate material properties. This module refines albedo predictions by removing baked-in shading and back-propagates reconstruction errors to optimize the geometry. Through joint optimization of geometry and material appearance, DIS achieves physically consistent, high-quality reconstructions suitable for accurate relighting. Our experiments show that DIS delivers SOTA relighting quality, enhanced rendering efficiency, lower memory consumption, and detailed surface reconstruction.

cs.GR↗

arXiv2025

Fast Algorithms for Scheduling Many-body Correlation Functions on Accelerators

Computation of correlation functions is a key operation in Lattice quantum chromodynamics (LQCD) simulations to extract nuclear physics observables. These functions involve many binary batch tensor contractions, each tensor possibly occupying hundreds of MBs of memory. Performing these contractions on GPU accelerators poses the challenge of scheduling them as to optimize tensor reuse and reduce data traffic. In this work we propose two fast novel scheduling algorithms that reorder contractions to increase temporal locality via input/intermediate tensor reuse. Our schedulers take advantage of application-specific features, such as contractions being binary and locality within contraction trees, to optimize the objective of minimizing peak memory. We integrate them into the LQCD analysis software suite Redstar and improve time-to-solution. Our schedulers attain upto 2.1x improvement in peak memory, which is reflected by a reduction of upto 4.2x in evictions, upto 1.8x in data traffic, resulting in upto 1.9x faster correlation function computation time.

cs.DC↗

arXiv2025

Many-body localization and particle multioccupancy in the disordered Bose-Hubbard model

We study the potential influence of the particle multi-occupations on the stability of many-body localization in the disordered Bose-Hubbard model. Within the higher-energy section of the dynamical phase diagram, we find that there is no apparent finite-size boundary drift between the thermal phase and the many-body localized regime. We substantiate this observation by introducing the Van Vleck perturbation theory into the field of many-body localization. The appropriateness of this method rests largely on the peculiar Hilbert-space structure enabled by the particles' Bose statistics. The situation is reversed in the lower-energy section of the dynamical phase diagram, where the significant finite-size boundary drift pushes the putative many-body localized regime up to the greater disorder strengths. We utilize the algebraic projection method to make a connection linking the disordered Bose-Hubbard model in the lower-energy section to an intricate disordered spin chain model. This issue of the finite-size drift could hence be analogous to what happens in the disordered Heisenberg chain. Both trends might be traced back to the particles' intrinsic or emergent single-occupancy constraint like the spin-$1/2$, hard-core boson, or spinless fermion degrees of freedom.

cond-mat.dis-nn↗

arXiv2025

Lessons Learned: A Multi-Agent Framework for Code LLMs to Learn and Improve

Recent studies show that LLMs possess different skills and specialize in different tasks. In fact, we observe that their varied performance occur in several levels of granularity. For example, in the code optimization task, code LLMs excel at different optimization categories and no one dominates others. This observation prompts the question of how one leverages multiple LLM agents to solve a coding problem without knowing their complementary strengths a priori. We argue that a team of agents can learn from each other's successes and failures so as to improve their own performance. Thus, a lesson is the knowledge produced by an agent and passed on to other agents in the collective solution process. We propose a lesson-based collaboration framework, design the lesson solicitation--banking--selection mechanism, and demonstrate that a team of small LLMs with lessons learned can outperform a much larger LLM and other multi-LLM collaboration methods.

cs.AI↗

arXiv2025

Deep Compositional Phase Diffusion for Long Motion Sequence Generation

Recent research on motion generation has shown significant progress in generating semantically aligned motion with singular semantics. However, when employing these models to create composite sequences containing multiple semantically generated motion clips, they often struggle to preserve the continuity of motion dynamics at the transition boundaries between clips, resulting in awkward transitions and abrupt artifacts. To address these challenges, we present Compositional Phase Diffusion, which leverages the Semantic Phase Diffusion Module (SPDM) and Transitional Phase Diffusion Module (TPDM) to progressively incorporate semantic guidance and phase details from adjacent motion clips into the diffusion process. Specifically, SPDM and TPDM operate within the latent motion frequency domain established by the pre-trained Action-Centric Motion Phase Autoencoder (ACT-PAE). This allows them to learn semantically important and transition-aware phase information from variable-length motion clips during training. Experimental results demonstrate the competitive performance of our proposed framework in generating compositional motion sequences that align semantically with the input conditions, while preserving phase transitional continuity between preceding and succeeding motion clips. Additionally, motion inbetweening task is made possible by keeping the phase parameter of the input motion sequences fixed throughout the diffusion process, showcasing the potential for extending the proposed framework to accommodate various application scenarios. Codes are available at https://github.com/asdryau/TransPhase.

cs.MM↗

arXiv2025

Quantification of Electrolyte Degradation in Lithium-ion Batteries with Neutron Imaging Techniques

Non-destructive characterization of lithium-ion batteries provides critical insights for optimizing performance and lifespan while preserving structural integrity. Optimizing electrolyte design in commercial LIBs requires consideration of composition, electrolyte-to-capacity ratio, spatial distribution, and associated degradation pathways. However, existing non-destructive methods for studying electrolyte infiltration, distribution, and degradation in LIBs lack the spatiotemporal resolution required for precise observation and quantification of the electrolyte. In this study, we employ neutron imaging with sufficient spatial resolution ~150 um and large field of view 20x20 cm2 to quantitatively resolve the electrolyte inventory and distribution within LiFePO4/graphite pouch cells under high-temperature accelerated aging. Quantitative standard curves based on neutron transmission attenuation reveal a clear electrolyte dry-out threshold at 3.18 g Ah-1 and the two stages evolutions of EI during cell aging were quantified. By integrating non-destructive electrochemical diagnostics, accelerated graphite material loss and liquid phase Li+ diffusion degradation is observed during pore-drying. Further analysis, including operando cyclic aging, reveals that the neutron transmission below the saturation reference is due to the enrichment of hydrogen nuclei within the solid-electrolyte interphase. Assumed pore-drying does not occur, the SEI signal of the electrodes can be quantitatively decoupled during ageing. Combined analyses with NI, TOF-SIMS, and SEM reveal that high EI cells exhibit uniform SEI growth and reduced degradation, while low EI cells show uneven SEI formation, accelerating capacity loss. This study unveils a dynamic electrolyte infiltration-consumption-dry-out process in LIBs, offering non-destructive and quantitative insights to guide sustainable and durable battery development.

physics.app-ph↗

arXiv2025

Towards Efficient 3D Gaussian Human Avatar Compression: A Prior-Guided Framework

This paper proposes an efficient 3D avatar coding framework that leverages compact human priors and canonical-to-target transformation to enable high-quality 3D human avatar video compression at ultra-low bit rates. The framework begins by training a canonical Gaussian avatar using articulated splatting in a network-free manner, which serves as the foundation for avatar appearance modeling. Simultaneously, a human-prior template is employed to capture temporal body movements through compact parametric representations. This decomposition of appearance and temporal evolution minimizes redundancy, enabling efficient compression: the canonical avatar is shared across the sequence, requiring compression only once, while the temporal parameters, consisting of just 94 parameters per frame, are transmitted with minimal bit-rate. For each frame, the target human avatar is generated by deforming canonical avatar via Linear Blend Skinning transformation, facilitating temporal coherent video reconstruction and novel view synthesis. Experimental results demonstrate that the proposed method significantly outperforms conventional 2D/3D codecs and existing learnable dynamic 3D Gaussian splatting compression method in terms of rate-distortion performance on mainstream multi-view human video datasets, paving the way for seamless immersive multimedia experiences in meta-verse applications.

eess.IV↗

arXiv2025

Graph Diffusion Transformers are In-Context Molecular Designers

In-context learning allows large models to adapt to new tasks from a few demonstrations, but it has shown limited success in molecular design. Existing databases such as ChEMBL contain molecular properties spanning millions of biological assays, yet labeled data for each property remain scarce. To address this limitation, we introduce demonstration-conditioned diffusion models (DemoDiff), which define task contexts using a small set of molecule-score examples instead of text descriptions. These demonstrations guide a denoising Transformer to generate molecules aligned with target properties. For scalable pretraining, we develop a new molecular tokenizer with Node Pair Encoding that represents molecules at the motif level, requiring 5.5$\times$ fewer nodes. We curate a dataset containing millions of context tasks from multiple sources covering both drugs and materials, and pretrain a 0.7-billion-parameter model on it. Across 33 design tasks in six categories, DemoDiff matches or surpasses language models 100-1000$\times$ larger and achieves an average rank of 3.63 compared to 5.25-10.20 for domain-specific approaches. These results position DemoDiff as a molecular foundation model for in-context molecular design. Our code is available at https://github.com/liugangcode/DemoDiff.

cs.LG↗

arXiv2025

Scientific Algorithm Discovery by Augmenting AlphaEvolve with Deep Research

Large language models hold promise as scientific assistants, yet existing agents either rely solely on algorithm evolution or on deep research in isolation, both of which face critical limitations. Pure algorithm evolution, as in AlphaEvolve, depends only on the internal knowledge of LLMs and quickly plateaus in complex domains, while pure deep research proposes ideas without validation, resulting in unrealistic or unimplementable solutions. We present DeepEvolve, an agent that integrates deep research with algorithm evolution, uniting external knowledge retrieval, cross-file code editing, and systematic debugging under a feedback-driven iterative loop. Each iteration not only proposes new hypotheses but also refines, implements, and tests them, avoiding both shallow improvements and unproductive over-refinements. Across nine benchmarks in chemistry, mathematics, biology, materials, and patents, DeepEvolve consistently improves the initial algorithm, producing executable new algorithms with sustained gains. By bridging the gap between unguided evolution and research without grounding, DeepEvolve provides a reliable framework for advancing scientific algorithm discovery. Our code is available at https://github.com/liugangcode/deepevolve.

cs.AI↗

arXiv2025

ZeroShotOpt: Towards Zero-Shot Pretrained Models for Efficient Black-Box Optimization

Global optimization of expensive, derivative-free black-box functions requires extreme sample efficiency. While Bayesian optimization (BO) is the current state-of-the-art, its performance hinges on surrogate and acquisition function hyper-parameters that are often hand-tuned and fail to generalize across problem landscapes. We present ZeroShotOpt, a general-purpose, pretrained model for continuous black-box optimization tasks ranging from 2D to 20D. Our approach leverages offline reinforcement learning on large-scale optimization trajectories collected from 12 BO variants. To scale pretraining, we generate millions of synthetic Gaussian process-based functions with diverse landscapes, enabling the model to learn transferable optimization policies. As a result, ZeroShotOpt achieves robust zero-shot generalization on a wide array of unseen benchmarks, matching or surpassing the sample efficiency of leading global optimizers, including BO, while also offering a reusable foundation for future extensions and improvements. Our open-source code, dataset, and model are available at: https://github.com/jamisonmeindl/zeroshotopt

cs.LG↗

arXiv2025

UI2V-Bench: An Understanding-based Image-to-video Generation Benchmark

Generative diffusion models are developing rapidly and attracting increasing attention due to their wide range of applications. Image-to-Video (I2V) generation has become a major focus in the field of video synthesis. However, existing evaluation benchmarks primarily focus on aspects such as video quality and temporal consistency, while largely overlooking the model's ability to understand the semantics of specific subjects in the input image or to ensure that the generated video aligns with physical laws and human commonsense. To address this gap, we propose UI2V-Bench, a novel benchmark for evaluating I2V models with a focus on semantic understanding and reasoning. It introduces four primary evaluation dimensions: spatial understanding, attribute binding, category understanding, and reasoning. To assess these dimensions, we design two evaluation methods based on Multimodal Large Language Models (MLLMs): an instance-level pipeline for fine-grained semantic understanding, and a feedback-based reasoning pipeline that enables step-by-step causal assessment for more accurate evaluation. UI2V-Bench includes approximately 500 carefully constructed text-image pairs and evaluates a range of both open source and closed-source I2V models across all defined dimensions. We further incorporate human evaluations, which show strong alignment with the proposed MLLM-based metrics. Overall, UI2V-Bench fills a critical gap in I2V evaluation by emphasizing semantic comprehension and reasoning ability, offering a robust framework and dataset to support future research and model development in the field.

cs.CV↗

arXiv2025

DyMoDreamer: World Modeling with Dynamic Modulation

A critical bottleneck in deep reinforcement learning (DRL) is sample inefficiency, as training high-performance agents often demands extensive environmental interactions. Model-based reinforcement learning (MBRL) mitigates this by building world models that simulate environmental dynamics and generate synthetic experience, improving sample efficiency. However, conventional world models process observations holistically, failing to decouple dynamic objects and temporal features from static backgrounds. This approach is computationally inefficient, especially for visual tasks where dynamic objects significantly influence rewards and decision-making performance. To address this, we introduce DyMoDreamer, a novel MBRL algorithm that incorporates a dynamic modulation mechanism to improve the extraction of dynamic features and enrich the temporal information. DyMoDreamer employs differential observations derived from a novel inter-frame differencing mask, explicitly encoding object-level motion cues and temporal dynamics. Dynamic modulation is modeled as stochastic categorical distributions and integrated into a recurrent state-space model (RSSM), enhancing the model's focus on reward-relevant dynamics. Experiments demonstrate that DyMoDreamer sets a new state-of-the-art on the Atari $100$k benchmark with a $156.6$\% mean human-normalized score, establishes a new record of $832$ on the DeepMind Visual Control Suite, and gains a $9.5$\% performance improvement after $1$M steps on the Crafter benchmark. Our code is released at https://github.com/Ultraman-Tiga1/DyMoDreamer.

cs.LG↗

arXiv2025

Sparse2Dense: A Keypoint-driven Generative Framework for Human Video Compression and Vertex Prediction

For bandwidth-constrained multimedia applications, simultaneously achieving ultra-low bitrate human video compression and accurate vertex prediction remains a critical challenge, as it demands the harmonization of dynamic motion modeling, detailed appearance synthesis, and geometric consistency. To address this challenge, we propose Sparse2Dense, a keypoint-driven generative framework that leverages extremely sparse 3D keypoints as compact transmitted symbols to enable ultra-low bitrate human video compression and precise human vertex prediction. The key innovation is the multi-task learning-based and keypoint-aware deep generative model, which could encode complex human motion via compact 3D keypoints and leverage these sparse keypoints to estimate dense motion for video synthesis with temporal coherence and realistic textures. Additionally, a vertex predictor is integrated to learn human vertex geometry through joint optimization with video generation, ensuring alignment between visual content and geometric structure. Extensive experiments demonstrate that the proposed Sparse2Dense framework achieves competitive compression performance for human video over traditional/generative video codecs, whilst enabling precise human vertex prediction for downstream geometry applications. As such, Sparse2Dense is expected to facilitate bandwidth-efficient human-centric media transmission, such as real-time motion analysis, virtual human animation, and immersive entertainment.

cs.CV↗

arXiv2025

Neighborhood Sampling Does Not Learn the Same Graph Neural Network

Neighborhood sampling is an important ingredient in the training of large-scale graph neural networks. It suppresses the exponential growth of the neighborhood size across network layers and maintains feasible memory consumption and time costs. While it becomes a standard implementation in practice, its systemic behaviors are less understood. We conduct a theoretical analysis by using the tool of neural tangent kernels, which characterize the (analogous) training dynamics of neural networks based on their infinitely wide counterparts -- Gaussian processes (GPs). We study several established neighborhood sampling approaches and the corresponding posterior GP. With limited samples, the posteriors are all different, although they converge to the same one as the sample size increases. Moreover, the posterior covariance, which lower-bounds the mean squared prediction error, is uncomparable, aligning with observations that no sampling approach dominates.

cs.LG↗

arXiv2025

Symmetry- and energy-resolved entanglement dynamics in a disordered Bose-Hubbard model

Using numerical quantum quenches with the integration of both symmetry and energy resolutions, we comprehensively study the dynamics of symmetry-resolved entanglement in a disordered Bose-Hubbard (dBH) model, concentrating on the two types of inhomogeneous initial states to target the lower- and higher-energy sections of its dynamical phase diagram. (i) Motivated by the recent experiment [A. Lukin et al., Science 364, 256 (2019)] which focused on the lower-energy dynamic behaviors of the dBH chain, we first show that, at low energies, for a thermalizing state, although the second law of thermodynamics prohibits the decrease of the total entropy over time, for part of the channel-resolved entropies, a long-term entropic reduction may arise at weak disorder. (ii) A companion channel-resolved analysis at strong disorder further hints that the priorly observed double-log growth of the number entropy might not directly indicate the breakdown of MBL in spin or fermion chains, providing a refreshing perspective on this major controversy in the community. (iii) From time-evolving the line-shape low-energy product state, we subsequently reveal an abrupt formation of a novel "entropy imbalance pattern" across the different symmetry channels. Intriguingly, this imbalance melts in the strong-disorder limit. We conjecture that the melting of the entropic pattern, together with the freezing of a concurrent particle-density wave, embodies a dual trait inherent to MBL. (iv) Specifically, we find a cluster MBL regime, unique to the Bose statistics, emerging from the higher-energy section. This cluster MBL regime realizable even at weak disorder appears not suffer from the finite-size drift and is distinguished by its absence of the hallmark of MBL - the unbounded growth of the entanglement entropy. Our theoretical predictions are by and large testable via the present experimental facilities.

cond-mat.dis-nn↗

arXiv2025

Automated algorithm design for convex optimization problems with linear equality constraints

Synthesis of optimization algorithms typically follows a {\em design-then-analyze\/} approach, which can obscure fundamental performance limits and hinder the systematic development of algorithms that operate near these limits. Recently, a framework grounded in robust control theory has emerged as a powerful tool for automating algorithm synthesis. By integrating design and analysis stages, fundamental performance bounds are revealed and synthesis of algorithms that achieve them is enabled. In this paper, we apply this framework to design algorithms for solving strongly convex optimization problems with linear equality constraints. Our approach yields a single-loop, gradient-based algorithm whose convergence rate is independent of the condition number of the constraint matrix. This improves upon the best known rate within the same algorithm class, which depends on the product of the condition numbers of the objective function and the constraint matrix.

math.OC↗