arXiv subjects

Yiming Zhang

Publications and source records attributed to Yiming Zhang.

At least 127 records · Page 7Linked to original sources

arXiv2025

VLIPP: Towards Physically Plausible Video Generation with Vision and Language Informed Physical Prior

Video diffusion models (VDMs) have advanced significantly in recent years, enabling the generation of highly realistic videos and drawing the attention of the community in their potential as world simulators. However, despite their capabilities, VDMs often fail to produce physically plausible videos due to an inherent lack of understanding of physics, resulting in incorrect dynamics and event sequences. To address this limitation, we propose a novel two-stage image-to-video generation framework that explicitly incorporates physics with vision and language informed physical prior. In the first stage, we employ a Vision Language Model (VLM) as a coarse-grained motion planner, integrating chain-of-thought and physics-aware reasoning to predict a rough motion trajectories/changes that approximate real-world physical dynamics while ensuring the inter-frame consistency. In the second stage, we use the predicted motion trajectories/changes to guide the video generation of a VDM. As the predicted motion trajectories/changes are rough, noise is added during inference to provide freedom to the VDM in generating motion with more fine details. Extensive experimental results demonstrate that our framework can produce physically plausible motion, and comparative evaluations highlight the notable superiority of our approach over existing methods. More video results are available on our Project Page: https://madaoer.github.io/projects/physically_plausible_video_generation.

cs.CV↗

arXiv2025

Safety-Aware Human-Lead Vehicle Platooning by Proactively Reacting to Uncertain Human Behaving

Human-Lead Cooperative Adaptive Cruise Control (HL-CACC) is regarded as a promising vehicle platooning technology in real-world implementation. By utilizing a Human-driven Vehicle (HV) as the platoon leader, HL-CACC reduces the cost and enhances the reliability of perception and decision-making. However, state-of-the-art HL-CACC technology still has a great limitation on driving safety due to the lack of considering the leading human driver's uncertain behavior. In this study, a HL-CACC controller is designed based on Stochastic Model Predictive Control (SMPC). It is enabled to predict the driving intention of the leading Connected Human-Driven Vehicle (CHV). The proposed controller has the following features: i) enhanced perceived safety in oscillating traffic; ii) guaranteed safety against hard brakes; iii) computational efficiency for real-time implementation. The proposed controller is evaluated on a PreScan&Simulink simulation platform. Real vehicle trajectory data is collected for the calibration of the simulation. Results reveal that the proposed controller: i) improves perceived safety by 19.17% in oscillating traffic; ii) enhances actual safety by 7.76% against hard brakes; iii) is confirmed with string stability. The computation time is approximately 3.2 milliseconds when running on a laptop equipped with an Intel i5-13500H CPU. This indicates the proposed controller is ready for real-time implementation.

cs.RO↗

arXiv2025

RankCLIP: Ranking-Consistent Language-Image Pretraining

Self-supervised contrastive learning models, such as CLIP, have set new benchmarks for vision-language models in many downstream tasks. However, their dependency on rigid one-to-one mappings overlooks the complex and often multifaceted relationships between and within texts and images. To this end, we introduce RankCLIP, a novel pre-training method that extends beyond the rigid one-to-one matching framework of CLIP and its variants. By extending the traditional pair-wise loss to list-wise, and leveraging both in-modal and cross-modal ranking consistency, RankCLIP improves the alignment process, enabling it to capture the nuanced many-to-many relationships between and within each modality. Through comprehensive experiments, we demonstrate the effectiveness of RankCLIP in various downstream tasks, notably achieving significant gains in zero-shot classifications over state-of-the-art methods, underscoring the importance of this enhanced learning process.

cs.CV↗

arXiv2025

Beyond Training: Dynamic Token Merging for Zero-Shot Video Understanding

Recent advancements in multimodal large language models (MLLMs) have opened new avenues for video understanding. However, achieving high fidelity in zero-shot video tasks remains challenging. Traditional video processing methods rely heavily on fine-tuning to capture nuanced spatial-temporal details, which incurs significant data and computation costs. In contrast, training-free approaches, though efficient, often lack robustness in preserving context-rich features across complex video content. To this end, we propose DYTO, a novel dynamic token merging framework for zero-shot video understanding that adaptively optimizes token efficiency while preserving crucial scene details. DYTO integrates a hierarchical frame selection and a bipartite token merging strategy to dynamically cluster key frames and selectively compress token sequences, striking a balance between computational efficiency with semantic richness. Extensive experiments across multiple benchmarks demonstrate the effectiveness of DYTO, achieving superior performance compared to both fine-tuned and training-free methods and setting a new state-of-the-art for zero-shot video understanding.

cs.CV↗

arXiv2025

Duoduo CLIP: Efficient 3D Understanding with Multi-View Images

We introduce Duoduo CLIP, a model for 3D representation learning that learns shape encodings from multi-view images instead of point clouds. The choice of multi-view images allows us to leverage 2D priors from off-the-shelf CLIP models to facilitate fine-tuning with 3D data. Our approach not only shows better generalization compared to existing point cloud methods, but also reduces GPU requirements and training time. In addition, the model is modified with cross-view attention to leverage information across multiple frames of the object which further boosts performance. Notably, our model is permutation invariant to the order of multi-view images while being pose-free. Compared to the current SOTA point cloud method that requires 480 A100 hours to train 1 billion model parameters we only require 57 A5000 hours and 87 million parameters. Multi-view images also provide more flexibility including being able to encode objects with a variable number of images, and performance scales when more views are used. In contrast, point cloud based methods require an entire scan or model of the object. We showcase this flexibility with benchmarks from images of real-world objects. Our model also achieves better performance in more fine-grained text to shape retrieval, demonstrating better text-and-shape alignment than point cloud based models.

cs.CV↗

arXiv2025

Exploiting Student Parallelism for Efficient GPU Inference of BERT-like Models in Online Services

Due to high accuracy, BERT-like models have been widely adopted by text mining and web searching. However, large BERT-like models suffer from inefficient online inference, facing the following two problems on GPUs: (1) their high accuracy relies on the large model depth, which linearly increases the sequential computation on GPUs; (2) stochastic and dynamic online workloads cause extra costs from batching and paddings. Therefore, we present \sys for the real-world setting of GPU inference on online workloads. At its core, \sys adopts stacking distillation and boosting ensemble, distilling the original deep model into a group of shallow but virtually stacked student models running in parallel. This enables \sys to achieve a lower model depth (e.g., two layers) than the others and the lowest inference latency while maintaining accuracy. In addition, adaptive student pruning realizes dynamic student numbers according to changing online workloads. Especially for occasional workload bursts, it can temporarily decrease the student number with minimal accuracy loss to improve system throughput. We conduct comprehensive experiments to verify the effectiveness, whose results show that \sys outperforms the baselines by $4.1\times\sim 1.6\times$ in latency while maintaining accuracy and achieves up to $22.27\times$ higher throughput for workload bursts.

cs.LG↗

arXiv2025

S2O: Static to Openable Enhancement for Articulated 3D Objects

Despite much progress in large 3D datasets there are currently few interactive 3D object datasets, and their scale is limited due to the manual effort required in their construction. We introduce the static to openable (S2O) task which creates interactive articulated 3D objects from static counterparts through openable part detection, motion prediction, and interior geometry completion. We formulate a unified framework to tackle this task, and curate a challenging dataset of openable 3D objects that serves as a test bed for systematic evaluation. Our experiments benchmark methods from prior work, extended and improved methods, and simple yet effective heuristics for the S2O task. We find that turning static 3D objects into interactively openable counterparts is possible but that all methods struggle to generalize to realistic settings of the task, and we highlight promising future work directions. Our work enables efficient creation of interactive 3D objects for robotic manipulation and embodied AI tasks.

cs.CV↗

arXiv2025

InfiFusion: A Unified Framework for Enhanced Cross-Model Reasoning via LLM Fusion

We introduce InfiFusion, an efficient training pipeline designed to integrate multiple domain-specialized Large Language Models (LLMs) into a single pivot model, effectively harnessing the strengths of each source model. Traditional fusion methods either merge model parameters directly or rely on knowledge distillation with rigid assumptions, limiting their flexibility and efficiency. InfiFusion overcomes these limitations by enhancing Universal Logit Distillation (ULD) with Top-K selection and Logits Standardization. We propose two fusion strategies: Pairwise Fusion (InfiFusion$_p$), where each source model knowledge is distilled individually into the pivot model followed by merging and Unified Fusion (InfiFusion$_u$), where knowledge from all source models is distilled simultaneously into the pivot model. InfiFusion outperforms the state-of-the-art models, such as Qwen-2.5-14B-Instruct and Phi-4, across 11 widely applied benchmarks covering reasoning, coding, mathematics, and instruction-following tasks. Notably, InfiFusion achieves this superior performance while significantly reduces computational costs, completing full training with only 160 H800 GPU hours compared to the millions typically required for traditional LLM training.

cs.CL↗

arXiv2025

InfiR : Crafting Effective Small Language Models and Multimodal Small Language Models in Reasoning

Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs) have made significant advancements in reasoning capabilities. However, they still face challenges such as high computational demands and privacy concerns. This paper focuses on developing efficient Small Language Models (SLMs) and Multimodal Small Language Models (MSLMs) that retain competitive reasoning abilities. We introduce a novel training pipeline that enhances reasoning capabilities and facilitates deployment on edge devices, achieving state-of-the-art performance while minimizing development costs. \InfR~ aims to advance AI systems by improving reasoning, reducing adoption barriers, and addressing privacy concerns through smaller model sizes. Resources are available at https://github. com/Reallm-Labs/InfiR.

cs.CL↗

arXiv2025

Assessment of Personalized Learning in Immersive and Intelligent Virtual Classroom on Student Engagement

As trends in education evolve, personalized learning has transformed individuals' engagement with knowledge and skill development. In the digital age, state-of-the-art technologies have been increasingly integrated into classrooms to support intelligent education and foster personalized learning experiences. One promising approach is the use of eye-tracking technology to evaluate student engagement in intelligent virtual classrooms. This paper explores the assessment of personalized learning in the virtual classroom and its impact on student engagement through the eye movement paradigm. The study aims to provide insights into how personalized learning approaches can enhance student participation, motivation, and academic performance in the online learning environment. Through a comprehensive literature review, case study, and data analysis, the paper examines the key elements of personalized learning, the methods of assessment, and the resulting effects on student engagement. The findings suggest that the eye movement paradigm has the potential to assess student engagement and promote better educational outcomes.

cs.CY↗

arXiv2025

IIMedGPT: Promoting Large Language Model Capabilities of Medical Tasks by Efficient Human Preference Alignment

Recent researches of large language models(LLM), which is pre-trained on massive general-purpose corpora, have achieved breakthroughs in responding human queries. However, these methods face challenges including limited data insufficiency to support extensive pre-training and can not align responses with users' instructions. To address these issues, we introduce a medical instruction dataset, CMedINS, containing six medical instructions derived from actual medical tasks, which effectively fine-tunes LLM in conjunction with other data. Subsequently, We launch our medical model, IIMedGPT, employing an efficient preference alignment method, Direct preference Optimization(DPO). The results show that our final model outperforms existing medical models in medical dialogue.Datsets, Code and model checkpoints will be released upon acceptance.

cs.CL↗

arXiv2024

DFME: A New Benchmark for Dynamic Facial Micro-expression Recognition

One of the most important subconscious reactions, micro-expression (ME), is a spontaneous, subtle, and transient facial expression that reveals human beings' genuine emotion. Therefore, automatically recognizing ME (MER) is becoming increasingly crucial in the field of affective computing, providing essential technical support for lie detection, clinical psychological diagnosis, and public safety. However, the ME data scarcity has severely hindered the development of advanced data-driven MER models. Despite the recent efforts by several spontaneous ME databases to alleviate this problem, there is still a lack of sufficient data. Hence, in this paper, we overcome the ME data scarcity problem by collecting and annotating a dynamic spontaneous ME database with the largest current ME data scale called DFME (Dynamic Facial Micro-expressions). Specifically, the DFME database contains 7,526 well-labeled ME videos spanning multiple high frame rates, elicited by 671 participants and annotated by more than 20 professional annotators over three years. Furthermore, we comprehensively verify the created DFME, including using influential spatiotemporal video feature learning models and MER models as baselines, and conduct emotion classification and ME action unit classification experiments. The experimental results demonstrate that the DFME database can facilitate research in automatic MER, and provide a new benchmark for this field. DFME will be published via https://mea-lab-421.github.io.

cs.CV↗

arXiv2024

EFSA: Towards Event-Level Financial Sentiment Analysis

In this paper, we extend financial sentiment analysis~(FSA) to event-level since events usually serve as the subject of the sentiment in financial text. Though extracting events from the financial text may be conducive to accurate sentiment predictions, it has specialized challenges due to the lengthy and discontinuity of events in a financial text. To this end, we reconceptualize the event extraction as a classification task by designing a categorization comprising coarse-grained and fine-grained event categories. Under this setting, we formulate the \textbf{E}vent-Level \textbf{F}inancial \textbf{S}entiment \textbf{A}nalysis~(\textbf{EFSA} for short) task that outputs quintuples consisting of (company, industry, coarse-grained event, fine-grained event, sentiment) from financial text. A large-scale Chinese dataset containing $12,160$ news articles and $13,725$ quintuples is publicized as a brand new testbed for our task. A four-hop Chain-of-Thought LLM-based approach is devised for this task. Systematically investigations are conducted on our dataset, and the empirical results demonstrate the benchmarking scores of existing methods and our proposed method can reach the current state-of-the-art. Our dataset and framework implementation are available at https://anonymous.4open.science/r/EFSA-645E

cs.CL↗

arXiv2024

Reinforced Symbolic Learning with Logical Constraints for Predicting Turbine Blade Fatigue Life

Accurate prediction of turbine blade fatigue life is essential for ensuring the safety and reliability of aircraft engines. A significant challenge in this domain is uncovering the intrinsic relationship between mechanical properties and fatigue life. This paper introduces Reinforced Symbolic Learning (RSL), a method that derives predictive formulas linking these properties to fatigue life. RSL incorporates logical constraints during symbolic optimization, ensuring that the generated formulas are both physically meaningful and interpretable. The optimization process is further enhanced using deep reinforcement learning, which efficiently guides the symbolic regression towards more accurate models. The proposed RSL method was evaluated on two turbine blade materials, GH4169 and TC4, to identify optimal fatigue life prediction models. When compared with six empirical formulas and five machine learning algorithms, RSL not only produces more interpretable formulas but also achieves superior or comparable predictive accuracy. Additionally, finite element simulations were conducted to assess mechanical properties at critical points on the blade, which were then used to predict fatigue life under various operating conditions.

cs.LG↗

arXiv2024

Llama Guard 3 Vision: Safeguarding Human-AI Image Understanding Conversations

We introduce Llama Guard 3 Vision, a multimodal LLM-based safeguard for human-AI conversations that involves image understanding: it can be used to safeguard content for both multimodal LLM inputs (prompt classification) and outputs (response classification). Unlike the previous text-only Llama Guard versions (Inan et al., 2023; Llama Team, 2024b,a), it is specifically designed to support image reasoning use cases and is optimized to detect harmful multimodal (text and image) prompts and text responses to these prompts. Llama Guard 3 Vision is fine-tuned on Llama 3.2-Vision and demonstrates strong performance on the internal benchmarks using the MLCommons taxonomy. We also test its robustness against adversarial attacks. We believe that Llama Guard 3 Vision serves as a good starting point to build more capable and robust content moderation tools for human-AI conversation with multimodal capabilities.

cs.CV↗

arXiv2024

Polymetis:Large Language Modeling for Multiple Material Domains

As the application of large language models in various fields continues to expand, materials science also ushers in opportunities for AI-driven innovation. The traditional way of relying on manual search for materials science-related information is now using artificial intelligence technology as an auxiliary tool to improve the efficiency of materials science research. To accelerate researchers' knowledge acquisition and intelligent decision-making support in materials science research, this paper proposes a large language model Polymetis model for a variety of materials fields, aiming to provide highly professional knowledge answers in the field of materials, covering energy materials, functional materials, alloy materials, physical chemistry, biology, and other material directions. The model uses a dataset of about 2 million material knowledge instructions, and in the process of building the dataset, we developed the Intelligent Extraction Large Model (IELM), which is specially used to extract and form structured knowledge from scientific texts, avoiding a large number of costs that need to be manually annotated, and improving efficiency. We inject this data into the GLM4-9B model for learning to enhance its inference capabilities in a variety of material domains. In addition, we have introduced enhanced prompt strategies to ensure that the answers to the model are more organized and comprehensive, providing efficient and comprehensive intelligent support for the diverse needs of materials science exploration, and promoting the development of material science.

cs.AI↗

arXiv2024

FinDVer: Explainable Claim Verification over Long and Hybrid-Content Financial Documents

We introduce FinDVer, a comprehensive benchmark specifically designed to evaluate the explainable claim verification capabilities of LLMs in the context of understanding and analyzing long, hybrid-content financial documents. FinDVer contains 2,400 expert-annotated examples, divided into three subsets: information extraction, numerical reasoning, and knowledge-intensive reasoning, each addressing common scenarios encountered in real-world financial contexts. We assess a broad spectrum of LLMs under long-context and RAG settings. Our results show that even the current best-performing system, GPT-4o, still lags behind human experts. We further provide in-depth analysis on long-context and RAG setting, Chain-of-Thought reasoning, and model reasoning errors, offering insights to drive future advancements. We believe that FinDVer can serve as a valuable benchmark for evaluating LLMs in claim verification over complex, expert-domain documents.

cs.CL↗

arXiv2024

Observation of a Halo Trimer in an Ultracold Bose-Fermi Mixture

The quantum mechanics of three interacting particles gives rise to interesting universal phenomena, such as the staircase of Efimov trimers predicted in the context of nuclear physics and observed in ultracold gases. Here, we observe a novel type of halo trimer using radiofrequency spectroscopy in an ultracold mixture of $^{23}$Na and $^{40}$K atoms. The trimers consist of two light bosons and one heavy fermion, and have the structure of a Feshbach dimer weakly bound to one additional boson. We find that the trimer peak closely follows the dimer resonance over the entire range of explored interaction strengths across an order of magnitude variation of the dimer energy, as reproduced by our theoretical analysis. The presence of this halo trimer is of direct relevance for many-body physics in ultracold mixtures and the association of ultracold molecules.

cond-mat.quant-gas↗