arXiv subjects

Zhen Li

Publications and source records attributed to Zhen Li.

At least 217 records · Page 12Linked to original sources

arXiv2025

Topo2Seq: Enhanced Topology Reasoning via Topology Sequence Learning

Extracting lane topology from perspective views (PV) is crucial for planning and control in autonomous driving. This approach extracts potential drivable trajectories for self-driving vehicles without relying on high-definition (HD) maps. However, the unordered nature and weak long-range perception of the DETR-like framework can result in misaligned segment endpoints and limited topological prediction capabilities. Inspired by the learning of contextual relationships in language models, the connectivity relations in roads can be characterized as explicit topology sequences. In this paper, we introduce Topo2Seq, a novel approach for enhancing topology reasoning via topology sequences learning. The core concept of Topo2Seq is a randomized order prompt-to-sequence learning between lane segment decoder and topology sequence decoder. The dual-decoder branches simultaneously learn the lane topology sequences extracted from the Directed Acyclic Graph (DAG) and the lane graph containing geometric information. Randomized order prompt-to-sequence learning extracts unordered key points from the lane graph predicted by the lane segment decoder, which are then fed into the prompt design of the topology sequence decoder to reconstruct an ordered and complete lane graph. In this way, the lane segment decoder learns powerful long-range perception and accurate topological reasoning from the topology sequence decoder. Notably, topology sequence decoder is only introduced during training and does not affect the inference efficiency. Experimental evaluations on the OpenLane-V2 dataset demonstrate the state-of-the-art performance of Topo2Seq in topology reasoning.

cs.CV↗

arXiv2025

MARIO: A Mixed Annotation Framework For Polyp Segmentation

Existing polyp segmentation models are limited by high labeling costs and the small size of datasets. Additionally, vast polyp datasets remain underutilized because these models typically rely on a single type of annotation. To address this dilemma, we introduce MARIO, a mixed supervision model designed to accommodate various annotation types, significantly expanding the range of usable data. MARIO learns from underutilized datasets by incorporating five forms of supervision: pixel-level, box-level, polygon-level, scribblelevel, and point-level. Each form of supervision is associated with a tailored loss that effectively leverages the supervision labels while minimizing the noise. This allows MARIO to move beyond the constraints of relying on a single annotation type. Furthermore, MARIO primarily utilizes dataset with weak and cheap annotations, reducing the dependence on large-scale, fully annotated ones. Experimental results across five benchmark datasets demonstrate that MARIO consistently outperforms existing methods, highlighting its efficacy in balancing trade-offs between different forms of supervision and maximizing polyp segmentation performance

cs.CV↗

arXiv2025

Lumina-Video: Efficient and Flexible Video Generation with Multi-scale Next-DiT

Recent advancements have established Diffusion Transformers (DiTs) as a dominant framework in generative modeling. Building on this success, Lumina-Next achieves exceptional performance in the generation of photorealistic images with Next-DiT. However, its potential for video generation remains largely untapped, with significant challenges in modeling the spatiotemporal complexity inherent to video data. To address this, we introduce Lumina-Video, a framework that leverages the strengths of Next-DiT while introducing tailored solutions for video synthesis. Lumina-Video incorporates a Multi-scale Next-DiT architecture, which jointly learns multiple patchifications to enhance both efficiency and flexibility. By incorporating the motion score as an explicit condition, Lumina-Video also enables direct control of generated videos' dynamic degree. Combined with a progressive training scheme with increasingly higher resolution and FPS, and a multi-source training scheme with mixed natural and synthetic data, Lumina-Video achieves remarkable aesthetic quality and motion smoothness at high training and inference efficiency. We additionally propose Lumina-V2A, a video-to-audio model based on Next-DiT, to create synchronized sounds for generated videos. Codes are released at https://www.github.com/Alpha-VLLM/Lumina-Video.

cs.CV↗

arXiv2025

Advancing Dense Endoscopic Reconstruction with Gaussian Splatting-driven Surface Normal-aware Tracking and Mapping

Simultaneous Localization and Mapping (SLAM) is essential for precise surgical interventions and robotic tasks in minimally invasive procedures. While recent advancements in 3D Gaussian Splatting (3DGS) have improved SLAM with high-quality novel view synthesis and fast rendering, these systems struggle with accurate depth and surface reconstruction due to multi-view inconsistencies. Simply incorporating SLAM and 3DGS leads to mismatches between the reconstructed frames. In this work, we present Endo-2DTAM, a real-time endoscopic SLAM system with 2D Gaussian Splatting (2DGS) to address these challenges. Endo-2DTAM incorporates a surface normal-aware pipeline, which consists of tracking, mapping, and bundle adjustment modules for geometrically accurate reconstruction. Our robust tracking module combines point-to-point and point-to-plane distance metrics, while the mapping module utilizes normal consistency and depth distortion to enhance surface reconstruction quality. We also introduce a pose-consistent strategy for efficient and geometrically coherent keyframe sampling. Extensive experiments on public endoscopic datasets demonstrate that Endo-2DTAM achieves an RMSE of $1.87\pm 0.63$ mm for depth reconstruction of surgical scenes while maintaining computationally efficient tracking, high-quality visual appearance, and real-time rendering. Our code will be released at github.com/lastbasket/Endo-2DTAM.

cs.CV↗

arXiv2025

IMAGINE-E: Image Generation Intelligence Evaluation of State-of-the-art Text-to-Image Models

With the rapid development of diffusion models, text-to-image(T2I) models have made significant progress, showcasing impressive abilities in prompt following and image generation. Recently launched models such as FLUX.1 and Ideogram2.0, along with others like Dall-E3 and Stable Diffusion 3, have demonstrated exceptional performance across various complex tasks, raising questions about whether T2I models are moving towards general-purpose applicability. Beyond traditional image generation, these models exhibit capabilities across a range of fields, including controllable generation, image editing, video, audio, 3D, and motion generation, as well as computer vision tasks like semantic segmentation and depth estimation. However, current evaluation frameworks are insufficient to comprehensively assess these models' performance across expanding domains. To thoroughly evaluate these models, we developed the IMAGINE-E and tested six prominent models: FLUX.1, Ideogram2.0, Midjourney, Dall-E3, Stable Diffusion 3, and Jimeng. Our evaluation is divided into five key domains: structured output generation, realism, and physical consistency, specific domain generation, challenging scenario generation, and multi-style creation tasks. This comprehensive assessment highlights each model's strengths and limitations, particularly the outstanding performance of FLUX.1 and Ideogram2.0 in structured and specific domain tasks, underscoring the expanding applications and potential of T2I models as foundational AI tools. This study provides valuable insights into the current state and future trajectory of T2I models as they evolve towards general-purpose usability. Evaluation scripts will be released at https://github.com/jylei16/Imagine-e.

cs.CV↗

arXiv2025

Decomposition of metric tensor in thermodynamic geometry in terms of relaxation timescales

Geometrical methods are extensively applied to thermodynamics including stochastic thermodynamics. In the case of slow-driving linear response regime, a geometrical framework, known as thermodynamic geometry, is established. The key of this framework is the thermodynamic length characterized by a metric tensor defined on the space of controlling variables. As the metric tensor is given in terms of the equilibrium time-correlation functions of the thermodynamic forces, it contains the information of timescales, which may be useful for analyzing the performance of heat engines. In this paper, we show that the metric tensor for underdamped Langevin dynamics can be decomposed in terms of the relaxation times of a system itself, which govern the timescales of the equilibrium time-correlation functions of the thermodynamic forces. As an application of the decomposition of the metric tensor, we demonstrate that it is possible to achieve Carnot efficiency at finite power by taking the vanishing limit of relaxation times without breaking trade-off relations between efficiency and power of heat engines in terms of thermodynamic geometry.

cond-mat.stat-mech↗

arXiv2025

Fair Text to Medical Image Diffusion Model with Subgroup Distribution Aligned Tuning

The text to medical image (T2MedI) with latent diffusion model has great potential to alleviate the scarcity of medical imaging data and explore the underlying appearance distribution of lesions in a specific patient status description. However, as the text to nature image models, we show that the T2MedI model can also bias to some subgroups to overlook the minority ones in the training set. In this work, we first build a T2MedI model based on the pre-trained Imagen model, which has the fixed contrastive language-image pre-training (CLIP) text encoder, while its decoder has been fine-tuned on medical images from the Radiology Objects in COntext (ROCO) dataset. Its gender bias is analyzed qualitatively and quantitatively. Toward this issue, we propose to fine-tune the T2MedI toward the target application dataset to align their sensitive subgroups distribution probability. Specifically, the alignment loss for fine-tuning is guided by an off-the-shelf sensitivity-subgroup classifier to match the classification probability between the generated images and the expected target dataset. In addition, the image quality is maintained by a CLIP-consistency regularization term following a knowledge distillation scheme. For evaluation, we set the target dataset to be enhanced as the BraST18 dataset, and trained a brain magnetic resonance (MR) slice-based gender classifier from it. With our method, the generated MR image can markedly reduce the inconsistency with the gender proportion in the BraTS18 dataset.

cs.CV↗

arXiv2025

Towards Multimodal Metaphor Understanding: A Chinese Dataset and Model for Metaphor Mapping Identification

Metaphors play a crucial role in human communication, yet their comprehension remains a significant challenge for natural language processing (NLP) due to the cognitive complexity involved. According to Conceptual Metaphor Theory (CMT), metaphors map a target domain onto a source domain, and understanding this mapping is essential for grasping the nature of metaphors. While existing NLP research has focused on tasks like metaphor detection and sentiment analysis of metaphorical expressions, there has been limited attention to the intricate process of identifying the mappings between source and target domains. Moreover, non-English multimodal metaphor resources remain largely neglected in the literature, hindering a deeper understanding of the key elements involved in metaphor interpretation. To address this gap, we developed a Chinese multimodal metaphor advertisement dataset (namely CM3D) that includes annotations of specific target and source domains. This dataset aims to foster further research into metaphor comprehension, particularly in non-English languages. Furthermore, we propose a Chain-of-Thought (CoT) Prompting-based Metaphor Mapping Identification Model (CPMMIM), which simulates the human cognitive process for identifying these mappings. Drawing inspiration from CoT reasoning and Bi-Level Optimization (BLO), we treat the task as a hierarchical identification problem, enabling more accurate and interpretable metaphor mapping. Our experimental results demonstrate the effectiveness of CPMMIM, highlighting its potential for advancing metaphor comprehension in NLP. Our dataset and code are both publicly available to encourage further advancements in this field.

cs.CL↗

arXiv2025

Thin Accretion Disk Around Rotating Hairy Black Hole: Radiative Property and Optical Appearance

The gravitational decoupling method systematically generates hairy modifications to the solutions in general relativity due to new gravitational sources. In view of the recent advances in astronomical observations, these hairy solutions are expected to be testable in the near-term observations. In this paper, we study the radiative property and optical appearance of the thin accretion disk around the rotating hairy black holes obtained by gravitational decoupling. We numerically compute the radiative flux, temperature, and differential luminosity of the thin accretion disk, and we also show its bolometric image by the ray-tracing method. By comparing with the results for the Kerr metric, we found that the deviations of the observational properties of the thin accretion disk from those of Kerr metric becomes significant in the rapid rotating case, or in the inner region of the disk. These results guide the observational investigations on the rotating hairy black hole.

gr-qc↗

arXiv2024

RuleExplorer: A Scalable Matrix Visualization for Understanding Tree Ensemble Classifiers

The high performance of tree ensemble classifiers benefits from a large set of rules, which, in turn, makes the models hard to understand. To improve interpretability, existing methods extract a subset of rules for approximation using model reduction techniques. However, by focusing on the reduced rule set, these methods often lose fidelity and ignore anomalous rules that, despite their infrequency, play crucial roles in real-world applications. This paper introduces a scalable visual analysis method to explain tree ensemble classifiers that contain tens of thousands of rules. The key idea is to address the issue of losing fidelity by adaptively organizing the rules as a hierarchy rather than reducing them. To ensure the inclusion of anomalous rules, we develop an anomaly-biased model reduction method to prioritize these rules at each hierarchical level. Synergized with this hierarchical organization of rules, we develop a matrix-based hierarchical visualization to support exploration at different levels of detail. Our quantitative experiments and case studies demonstrate how our method fosters a deeper understanding of both common and anomalous rules, thereby enhancing interpretability without sacrificing comprehensiveness.

cs.LG↗

arXiv2024

Observational Features of Thin Accretion Disk Around Rotating Regular Black Hole

Rotating regular black hole, as a promising extension beyond general relativity, offer a phenomenological model that resolves spacetime singularities. In this study, we investigate the observational features of thin accretion disk around a well-known rotating regular black hole, which introduce an exponential converge factor $e^{-k/r}$ to the black hole mass M, where $k$ is the regular parameter. By studying the effects of the regular parameter on key quantities such as energy, angular momentum, and the innermost stable circular orbits of a test particle, we are able to analyze the radiative flux, temperature, and differential luminosity of the thin accretion disk in this rotation regular black hole spacetime. By using the ray-tracing method, we also obtain the bolometric images of thin accretion disk around this rotating regular black hole, with various black hole spins and inclination angels. Our results show that the regular parameter significantly impacts the observables, enhancing the radiation efficiency of thin accretion disk and contracting the lensing bands of black hole image, compared to Kerr black hole. These effects become more pronounced for rapidly rotating regular black hole and large inclination angels, making them more detectable in astrophysical observations.

gr-qc↗

arXiv2024

Diffusion-Enhanced Test-time Adaptation with Text and Image Augmentation

Existing test-time prompt tuning (TPT) methods focus on single-modality data, primarily enhancing images and using confidence ratings to filter out inaccurate images. However, while image generation models can produce visually diverse images, single-modality data enhancement techniques still fail to capture the comprehensive knowledge provided by different modalities. Additionally, we note that the performance of TPT-based methods drops significantly when the number of augmented images is limited, which is not unusual given the computational expense of generative augmentation. To address these issues, we introduce IT3A, a novel test-time adaptation method that utilizes a pre-trained generative model for multi-modal augmentation of each test sample from unknown new domains. By combining augmented data from pre-trained vision and language models, we enhance the ability of the model to adapt to unknown new test data. Additionally, to ensure that key semantics are accurately retained when generating various visual and text enhancements, we employ cosine similarity filtering between the logits of the enhanced images and text with the original test data. This process allows us to filter out some spurious augmentation and inadequate combinations. To leverage the diverse enhancements provided by the generation model across different modals, we have replaced prompt tuning with an adapter for greater flexibility in utilizing text templates. Our experiments on the test datasets with distribution shifts and domain gaps show that in a zero-shot setting, IT3A outperforms state-of-the-art test-time prompt tuning methods with a 5.50% increase in accuracy.

cs.CV↗

arXiv2024

Towards quantum-centric simulations of extended molecules: sample-based quantum diagonalization enhanced with density matrix embedding theory

Computing ground-state properties of molecules is a promising application for quantum computers operating in concert with classical high-performance computing resources. Quantum embedding methods are a family of algorithms particularly suited to these computational platforms: they combine high-level calculations on active regions of a molecule with low-level calculations on the surrounding environment, thereby avoiding expensive high-level full-molecule calculations and allowing to distribute computational cost across multiple and heterogeneous computing units. Here, we present the first density matrix embedding theory (DMET) simulations performed in combination with the sample-based quantum diagonalization (SQD) method. We employ the DMET-SQD formalism to compute the ground-state energy of a ring of 18 hydrogen atoms, and the relative energies of the chair, half-chair, twist-boat, and boat conformers of cyclohexane. The full-molecule 41- and 89-qubit simulations are decomposed into 27- and 32-qubit active-region simulations, that we carry out on the ibm_cleveland device, obtaining results in agreement with reference classical methods. Our DMET-SQD calculations mark a tangible progress in the size of active regions that can be accurately tackled by near-term quantum computers, and are an early demonstration of the potential for quantum-centric simulations to accurately treat the electronic structure of large molecules, with the ultimate goal of tackling systems such as peptides and proteins.

quant-ph↗

arXiv2024

V$^2$-SfMLearner: Learning Monocular Depth and Ego-motion for Multimodal Wireless Capsule Endoscopy

Deep learning can predict depth maps and capsule ego-motion from capsule endoscopy videos, aiding in 3D scene reconstruction and lesion localization. However, the collisions of the capsule endoscopies within the gastrointestinal tract cause vibration perturbations in the training data. Existing solutions focus solely on vision-based processing, neglecting other auxiliary signals like vibrations that could reduce noise and improve performance. Therefore, we propose V$^2$-SfMLearner, a multimodal approach integrating vibration signals into vision-based depth and capsule motion estimation for monocular capsule endoscopy. We construct a multimodal capsule endoscopy dataset containing vibration and visual signals, and our artificial intelligence solution develops an unsupervised method using vision-vibration signals, effectively eliminating vibration perturbations through multimodal learning. Specifically, we carefully design a vibration network branch and a Fourier fusion module, to detect and mitigate vibration noises. The fusion framework is compatible with popular vision-only algorithms. Extensive validation on the multimodal dataset demonstrates superior performance and robustness against vision-only algorithms. Without the need for large external equipment, our V$^2$-SfMLearner has the potential for integration into clinical capsule robots, providing real-time and dependable digestive examination tools. The findings show promise for practical implementation in clinical settings, enhancing the diagnostic capabilities of doctors.

cs.CV↗

arXiv2024

Adaptive Pruning for Large Language Models with Structural Importance Awareness

The recent advancements in large language models (LLMs) have significantly improved language understanding and generation capabilities. However, it is difficult to deploy LLMs on resource-constrained edge devices due to their high computational and storage resource demands. To address this issue, we propose a novel LLM model pruning method, namely structurally-aware adaptive pruning (SAAP), to significantly reduce the computational and memory costs while maintaining model performance. We first define an adaptive importance fusion metric to evaluate the importance of all coupled structures in LLMs by considering their homoscedastic uncertainty. Then, we rank the importance of all modules to determine the specific layers that should be pruned to meet particular performance requirements. Furthermore, we develop a new group fine-tuning strategy to improve the inference efficiency of LLMs. Finally, we evaluate the proposed SAAP method on multiple LLMs across two common tasks, i.e., zero-shot classification and text generation. Experimental results show that our SAAP method outperforms several state-of-the-art baseline methods, achieving 2.17%, 2.37%, and 2.39% accuracy gains on LLaMA-7B, Vicuna-7B, and LLaMA-13B. Additionally, SAAP improves the token generation speed by 5%, showcasing its practical advantages in resource-constrained scenarios.

cs.CL↗

arXiv2024

Consistency of Compositional Generalization across Multiple Levels

Compositional generalization is the capability of a model to understand novel compositions composed of seen concepts. There are multiple levels of novel compositions including phrase-phrase level, phrase-word level, and word-word level. Existing methods achieve promising compositional generalization, but the consistency of compositional generalization across multiple levels of novel compositions remains unexplored. The consistency refers to that a model should generalize to a phrase-phrase level novel composition, and phrase-word/word-word level novel compositions that can be derived from it simultaneously. In this paper, we propose a meta-learning based framework, for achieving consistent compositional generalization across multiple levels. The basic idea is to progressively learn compositions from simple to complex for consistency. Specifically, we divide the original training set into multiple validation sets based on compositional complexity, and introduce multiple meta-weight-nets to generate sample weights for samples in different validation sets. To fit the validation sets in order of increasing compositional complexity, we optimize the parameters of each meta-weight-net independently and sequentially in a multilevel optimization manner. We build a GQA-CCG dataset to quantitatively evaluate the consistency. Experimental results on visual question answering and temporal video grounding, demonstrate the effectiveness of the proposed framework. We release GQA-CCG at https://github.com/NeverMoreLCH/CCG.

cs.CV↗

arXiv2024

An Efficient Occupancy World Model via Decoupled Dynamic Flow and Image-assisted Training

The field of autonomous driving is experiencing a surge of interest in world models, which aim to predict potential future scenarios based on historical observations. In this paper, we introduce DFIT-OccWorld, an efficient 3D occupancy world model that leverages decoupled dynamic flow and image-assisted training strategy, substantially improving 4D scene forecasting performance. To simplify the training process, we discard the previous two-stage training strategy and innovatively reformulate the occupancy forecasting problem as a decoupled voxels warping process. Our model forecasts future dynamic voxels by warping existing observations using voxel flow, whereas static voxels are easily obtained through pose transformation. Moreover, our method incorporates an image-assisted training paradigm to enhance prediction reliability. Specifically, differentiable volume rendering is adopted to generate rendered depth maps through predicted future volumes, which are adopted in render-based photometric consistency. Experiments demonstrate the effectiveness of our approach, showcasing its state-of-the-art performance on the nuScenes and OpenScene benchmarks for 4D occupancy forecasting, end-to-end motion planning and point cloud forecasting. Concretely, it achieves state-of-the-art performances compared to existing 3D world models while incurring substantially lower computational costs.

cs.CV↗

arXiv2024

Personalized Federated Deep Reinforcement Learning for Heterogeneous Edge Content Caching Networks

Proactive caching is essential for minimizing latency and improving Quality of Experience (QoE) in multi-server edge networks. Federated Deep Reinforcement Learning (FDRL) is a promising approach for developing cache policies tailored to dynamic content requests. However, FDRL faces challenges such as an expanding caching action space due to increased content numbers and difficulty in adapting global information to heterogeneous edge environments. In this paper, we propose a Personalized Federated Deep Reinforcement Learning framework for Caching, called PF-DRL-Ca, with the aim to maximize system utility while satisfying caching capability constraints. To manage the expanding action space, we employ a new DRL algorithm, Multi-head Deep Q-Network (MH-DQN), which reshapes the action output layers of DQN into a multi-head structure where each head generates a sub-dimensional action. We next integrate the proposed MH-DQN into a personalized federated training framework, employing a layer-wise approach for training to derive a personalized model that can adapt to heterogeneous environments while exploiting the global information to accelerate learning convergence. Our extensive experimental results demonstrate the superiority of MH-DQN over traditional DRL algorithms on a single server, as well as the advantages of the personal federated training architecture compared to other frameworks.

cs.NI↗