arXiv subjects

Yanan Li

Publications and source records attributed to Yanan Li.

At least 37 records · Page 2Linked to original sources

arXiv2025

Roadside Monocular 3D Detection Prompted by 2D Detection

Roadside monocular 3D detection requires detecting objects of predefined classes in an RGB frame and predicting their 3D attributes, such as bird's-eye-view (BEV) locations. It has broad applications in traffic control, vehicle-vehicle communication, and vehicle-infrastructure cooperative perception. To address this task, we introduce Promptable 3D Detector (Pro3D), a novel detector design that leverages 2D detections as prompts. We build our Pro3D upon two key insights. First, compared to a typical 3D detector, a 2D detector is ``easier'' to train due to fewer loss terms and performs significantly better at localizing objects w.r.t 2D metrics. Second, once 2D detections precisely locate objects in the image, a 3D detector can focus on lifting these detections into 3D BEV, especially when fixed camera pose or scene geometry provide an informative prior. To encode and incorporate 2D detections, we explore three methods: (a) concatenating features from both 2D and 3D detectors, (b) attentively fusing 2D and 3D detector features, and (c) encoding properties of predicted 2D bounding boxes \{$x$, $y$, width, height, label\} and attentively fusing them with the 3D detector feature. Interestingly, the third method significantly outperforms the others, underscoring the effectiveness of 2D detections as prompts that offer precise object targets and allow the 3D detector to focus on lifting them into 3D. Pro3D is adaptable for use with a wide range of 2D and 3D detectors with minimal modifications. Comprehensive experiments demonstrate that our Pro3D significantly enhances existing methods, achieving state-of-the-art results on two contemporary benchmarks.

cs.CV↗

arXiv2025

TasselNetV4: A vision foundation model for cross-scene, cross-scale, and cross-species plant counting

Accurate plant counting provides valuable information for agriculture such as crop yield prediction, plant density assessment, and phenotype quantification. Vision-based approaches are currently the mainstream solution. Prior art typically uses a detection or a regression model to count a specific plant. However, plants have biodiversity, and new cultivars are increasingly bred each year. It is almost impossible to exhaust and build all species-dependent counting models. Inspired by class-agnostic counting (CAC) in computer vision, we argue that it is time to rethink the problem formulation of plant counting, from what plants to count to how to count plants. In contrast to most daily objects with spatial and temporal invariance, plants are dynamic, changing with time and space. Their non-rigid structure often leads to worse performance than counting rigid instances like heads and cars such that current CAC and open-world detection models are suboptimal to count plants. In this work, we inherit the vein of the TasselNet plant counting model and introduce a new extension, TasselNetV4, shifting from species-specific counting to cross-species counting. TasselNetV4 marries the local counting idea of TasselNet with the extract-and-match paradigm in CAC. It builds upon a plain vision transformer and incorporates novel multi-branch box-aware local counters used to enhance cross-scale robustness. Two challenging datasets, PAC-105 and PAC-Somalia, are harvested. Extensive experiments against state-of-the-art CAC models show that TasselNetV4 achieves not only superior counting performance but also high efficiency.Our results indicate that TasselNetV4 emerges to be a vision foundation model for cross-scene, cross-scale, and cross-species plant counting.

cs.CV↗

arXiv2025

DSS-Prompt: Dynamic-Static Synergistic Prompting for Few-Shot Class-Incremental Learning

Learning from large-scale pre-trained models with strong generalization ability has shown remarkable success in a wide range of downstream tasks recently, but it is still underexplored in the challenging few-shot class-incremental learning (FSCIL) task. It aims to continually learn new concepts from limited training samples without forgetting the old ones at the same time. In this paper, we introduce DSS-Prompt, a simple yet effective approach that transforms the pre-trained Vision Transformer with minimal modifications in the way of prompts into a strong FSCIL classifier. Concretely, we synergistically utilize two complementary types of prompts in each Transformer block: static prompts to bridge the domain gap between the pre-training and downstream datasets, thus enabling better adaption; and dynamic prompts to capture instance-aware semantics, thus enabling easy transfer from base to novel classes. Specially, to generate dynamic prompts, we leverage a pre-trained multi-modal model to extract input-related diverse semantics, thereby generating complementary input-aware prompts, and then adaptively adjust their importance across different layers. In this way, on top of the prompted visual embeddings, a simple prototype classifier can beat state-of-the-arts without further training on the incremental tasks. We conduct extensive experiments on four benchmarks to validate the effectiveness of our DSS-Prompt and show that it consistently achieves better performance than existing approaches on all datasets and can alleviate the catastrophic forgetting issue as well.

cs.CV↗

arXiv2025

Spray flame synthesis of Y2O3-MgO nanoparticles for mid-infrared transparent nanocomposite ceramics

Spray flame synthesis offers a promising method for scalable production of homogeneously mixed Y2O3-MgO nanopowders as next-generation infrared-transparent window material, which has attracted significant attention owing to its excellent optical properties at high temperatures. However, systematic understanding of how flame synthesis parameters influence particle morphology, crystal phase, solid solubility, and subsequent ceramic performance remains insufficiently understood. In this study, we investigated the influence of precursor chemistry on particle crystal phase and examined the solid solubility of MgO in Y2O3 under different flame temperatures, demonstrating that the high-temperature conditions with O2 as dispersion gas allow up to 50 mol% MgO to fully dissolve into Y2O3, far exceeding the equilibrium solubility limit of 7 mol% at the eutectic temperature (2100°C) and near-zero at room temperature. Furthermore, we systematically evaluated how powder characteristics and sintering parameters-including powder deagglomeration methods, vacuum sintering temperature, hot isostatic pressing (HIP) temperature, and initial powder characteristics-affect ceramic microstructures and infrared transmittance. Despite cracking induced by phase transformation and finer particle sizes, ceramics fabricated from oxygen-synthesized monoclinic-dominated powders exhibited superior near-infrared transmittance (56.2% at 1550 nm), attributed to enhanced atomic mixing and effective grain boundary pinning. After optimization, pure cubic phase powders produced intact and crack-free ceramics with outstanding mid-infrared transparency, achieving a maximum transmittance of 84.6% and an average transmittance of 82.3% in 3-5 um range.

cond-mat.mtrl-sci↗

arXiv2025

LoRASuite: Efficient LoRA Adaptation Across Large Language Model Upgrades

As Large Language Models (LLMs) are frequently updated, LoRA weights trained on earlier versions quickly become obsolete. The conventional practice of retraining LoRA weights from scratch on the latest model is costly, time-consuming, and environmentally detrimental, particularly as the diversity of LLMs and downstream tasks expands. This motivates a critical question: "How can we efficiently leverage existing LoRA weights to adapt to newer model versions?" To address this, we propose LoRASuite, a modular approach tailored specifically to various types of LLM updates. First, we compute a transfer matrix utilizing known parameters from both old and new LLMs. Next, we allocate corresponding layers and attention heads based on centered kernel alignment and cosine similarity metrics, respectively. A subsequent small-scale, skillful fine-tuning step ensures numerical stability. Experimental evaluations demonstrate that LoRASuite consistently surpasses small-scale vanilla LoRA methods. Notably, on backbone LLMs such as MiniCPM and Qwen, LoRASuite even exceeds the performance of full-scale LoRA retraining, with average improvements of +1.4 and +6.6 points on math tasks, respectively. Additionally, LoRASuite significantly reduces memory consumption by 5.5 GB and computational time by 78.23%.

cs.LG↗

arXiv2025

Confidence-based Intent Prediction for Teleoperation in Bimanual Robotic Suturing

Robotic-assisted procedures offer enhanced precision, but while fully autonomous systems are limited in task knowledge, difficulties in modeling unstructured environments, and generalisation abilities, fully manual teleoperated systems also face challenges such as delay, stability, and reduced sensory information. To address these, we developed an interactive control strategy that assists the human operator by predicting their motion plan at both high and low levels. At the high level, a surgeme recognition system is employed through a Transformer-based real-time gesture classification model to dynamically adapt to the operator's actions, while at the low level, a Confidence-based Intention Assimilation Controller adjusts robot actions based on user intent and shared control paradigms. The system is built around a robotic suturing task, supported by sensors that capture the kinematics of the robot and task dynamics. Experiments across users with varying skill levels demonstrated the effectiveness of the proposed approach, showing statistically significant improvements in task completion time and user satisfaction compared to traditional teleoperation.

cs.RO↗

arXiv2025

Solving Instance Detection from an Open-World Perspective

Instance detection (InsDet) aims to localize specific object instances within a novel scene imagery based on given visual references. Technically, it requires proposal detection to identify all possible object instances, followed by instance-level matching to pinpoint the ones of interest. Its open-world nature supports its broad applications from robotics to AR/VR but also presents significant challenges: methods must generalize to unknown testing data distributions because (1) the testing scene imagery is unseen during training, and (2) there are domain gaps between visual references and detected proposals. Existing methods tackle these challenges by synthesizing diverse training examples or utilizing off-the-shelf foundation models (FMs). However, they only partially capitalize the available open-world information. In contrast, we approach InsDet from an Open-World perspective, introducing our method IDOW. We find that, while pretrained FMs yield high recall in instance detection, they are not specifically optimized for instance-level feature matching. Therefore, we adapt pretrained FMs for improved instance-level matching using open-world data. Our approach incorporates metric learning along with novel data augmentations, which sample distractors as negative examples and synthesize novel-view instances to enrich the visual references. Extensive experiments demonstrate that our method significantly outperforms prior works, achieving >10 AP over previous results on two recently released challenging benchmark datasets in both conventional and novel instance detection settings.

cs.CV↗

arXiv2024

Quantum Griffiths singularity in a three-dimensional superconductor to Anderson critical insulator transition

Disorder is ubiquitous in real materials and can have dramatic effects on quantum phase transitions. Originating from the disorder enhanced quantum fluctuation, quantum Griffiths singularity (QGS) has been revealed as a universal phenomenon in quantum criticality of low-dimensional superconductors. However, due to the weak fluctuation effect, QGS is very challenging to detect experimentally in three-dimensional (3D) superconducting systems. Here we report the discovery of QGS associated with the quantum phase transition from 3D superconductor to Anderson critical insulator in a spinel oxide MgTi2O4 (MTO). Under both perpendicular and parallel magnetic field, the dynamical critical exponent diverges when approaching the quantum critical point, demonstrating the existence of 3D QGS. Among 3D superconductors, MTO shows a relatively strong fluctuation effect featured as a wide superconducting transition region. The enhanced fluctuation, which may arise from the mobility edge of Anderson localization, finally leads to the occurrence of 3D quantum phase transition and QGS. Our findings offer a new perspective to understand quantum phase transitions in strongly disordered 3D systems.

cond-mat.supr-con↗

arXiv2024

High-Temperature Anomalous Metal States in Iron-Based Interface Superconductors

The nature of the anomalous metal state has been a major puzzle in condensed matter physics for more than three decades. Here, we report systematic investigation and modulation of the anomalous metal states in high-temperature interface superconductor FeSe films on SrTiO3 substrate. Remarkably, under zero magnetic field, the anomalous metal state persists up to 20 K in pristine FeSe films, an exceptionally high temperature standing out from previous observations. In stark contrast, for the FeSe films with nano-hole arrays, the characteristic temperature of the anomalous metal state is considerably reduced. We demonstrate that the observed anomalous metal states originate from the quantum tunneling of vortices adjusted by the Ohmic dissipation. Our work offers a perspective for understanding the origin and modulation of the anomalous metal states in two-dimensional bosonic systems.

cond-mat.supr-con↗

arXiv2024

The Neglected Tails in Vision-Language Models

Vision-language models (VLMs) excel in zero-shot recognition but their performance varies greatly across different visual concepts. For example, although CLIP achieves impressive accuracy on ImageNet (60-80%), its performance drops below 10% for more than ten concepts like night snake, presumably due to their limited presence in the pretraining data. However, measuring the frequency of concepts in VLMs' large-scale datasets is challenging. We address this by using large language models (LLMs) to count the number of pretraining texts that contain synonyms of these concepts. Our analysis confirms that popular datasets, such as LAION, exhibit a long-tailed concept distribution, yielding biased performance in VLMs. We also find that downstream applications of VLMs, including visual chatbots (e.g., GPT-4V) and text-to-image models (e.g., Stable Diffusion), often fail to recognize or generate images of rare concepts identified by our method. To mitigate the imbalanced performance of zero-shot VLMs, we propose REtrieval-Augmented Learning (REAL). First, instead of prompting VLMs using the original class names, REAL uses their most frequent synonyms found in pretraining texts. This simple change already outperforms costly human-engineered and LLM-enriched prompts over nine benchmark datasets. Second, REAL trains a linear classifier on a small yet balanced set of pretraining data retrieved using concept synonyms. REAL surpasses the previous zero-shot SOTA, using 400x less storage and 10,000x less training time!

cs.CV↗

arXiv2024

Knowledge Graph Driven Recommendation System Algorithm

In this paper, we propose a novel graph neural network-based recommendation model called KGLN, which leverages Knowledge Graph (KG) information to enhance the accuracy and effectiveness of personalized recommendations. We first use a single-layer neural network to merge individual node features in the graph, and then adjust the aggregation weights of neighboring entities by incorporating influence factors. The model evolves from a single layer to multiple layers through iteration, enabling entities to access extensive multi-order associated entity information. The final step involves integrating features of entities and users to produce a recommendation score. The model performance was evaluated by comparing its effects on various aggregation methods and influence factors. In tests over the MovieLen-1M and Book-Crossing datasets, KGLN shows an Area Under the ROC curve (AUC) improvement of 0.3% to 5.9% and 1.1% to 8.2%, respectively, which is better than existing benchmark methods like LibFM, DeepFM, Wide&Deep, and RippleNet.

cs.IR↗

arXiv2023

Real-Time Progressive Learning: Accumulate Knowledge from Control with Neural-Network-Based Selective Memory

Memory, as the basis of learning, determines the storage, update and forgetting of knowledge and further determines the efficiency of learning. Featured with the mechanism of memory, a radial basis function neural network based learning control scheme named real-time progressive learning (RTPL) is proposed to learn the unknown dynamics of the system with guaranteed stability and closed-loop performance. Instead of the Lyapunov-based weight update law of conventional neural network learning control (NNLC), which mainly concentrates on stability and control performance, RTPL employs the selective memory recursive least squares (SMRLS) algorithm to update the weights of the neural network and achieves the following merits: 1) improved learning speed without filtering, 2) robustness to hyperparameter setting of neural networks, 3) good generalization ability, i.e., reuse of learned knowledge in different tasks, and 4) guaranteed learning performance under parameter perturbation. Moreover, RTPL realizes continuous accumulation of knowledge as a result of its reasonably allocated memory while NNLC may gradually forget knowledge that it has learned. Corresponding theoretical analysis and simulation studies demonstrate the effectiveness of RTPL.

eess.SY↗

arXiv2023

A High-Resolution Dataset for Instance Detection with Multi-View Instance Capture

Instance detection (InsDet) is a long-lasting problem in robotics and computer vision, aiming to detect object instances (predefined by some visual examples) in a cluttered scene. Despite its practical significance, its advancement is overshadowed by Object Detection, which aims to detect objects belonging to some predefined classes. One major reason is that current InsDet datasets are too small in scale by today's standards. For example, the popular InsDet dataset GMU (published in 2016) has only 23 instances, far less than COCO (80 classes), a well-known object detection dataset published in 2014. We are motivated to introduce a new InsDet dataset and protocol. First, we define a realistic setup for InsDet: training data consists of multi-view instance captures, along with diverse scene images allowing synthesizing training images by pasting instance images on them with free box annotations. Second, we release a real-world database, which contains multi-view capture of 100 object instances, and high-resolution (6k x 8k) testing images. Third, we extensively study baseline methods for InsDet on our dataset, analyze their performance and suggest future work. Somewhat surprisingly, using the off-the-shelf class-agnostic segmentation model (Segment Anything Model, SAM) and the self-supervised feature representation DINOv2 performs the best, achieving >10 AP better than end-to-end trained InsDet models that repurpose object detectors (e.g., FasterRCNN and RetinaNet).

cs.CV↗

arXiv2023

Prompting Scientific Names for Zero-Shot Species Recognition

Trained on web-scale image-text pairs, Vision-Language Models (VLMs) such as CLIP can recognize images of common objects in a zero-shot fashion. However, it is underexplored how to use CLIP for zero-shot recognition of highly specialized concepts, e.g., species of birds, plants, and animals, for which their scientific names are written in Latin or Greek. Indeed, CLIP performs poorly for zero-shot species recognition with prompts that use scientific names, e.g., "a photo of Lepus Timidus" (which is a scientific name in Latin). Because these names are usually not included in CLIP's training set. To improve performance, prior works propose to use large-language models (LLMs) to generate descriptions (e.g., of species color and shape) and additionally use them in prompts. We find that they bring only marginal gains. Differently, we are motivated to translate scientific names (e.g., Lepus Timidus) to common English names (e.g., mountain hare) and use such in the prompts. We find that common names are more likely to be included in CLIP's training set, and prompting them achieves 2$\sim$5 times higher accuracy on benchmarking datasets of fine-grained species recognition.

cs.CV↗

arXiv2023

Selective Memory Recursive Least Squares: Recast Forgetting into Memory in RBF Neural Network Based Real-Time Learning

In radial basis function neural network (RBFNN) based real-time learning tasks, forgetting mechanisms are widely used such that the neural network can keep its sensitivity to new data. However, with forgetting mechanisms, some useful knowledge will get lost simply because they are learned a long time ago, which we refer to as the passive knowledge forgetting phenomenon. To address this problem, this paper proposes a real-time training method named selective memory recursive least squares (SMRLS) in which the classical forgetting mechanisms are recast into a memory mechanism. Different from the forgetting mechanism, which mainly evaluates the importance of samples according to the time when samples are collected, the memory mechanism evaluates the importance of samples through both temporal and spatial distribution of samples. With SMRLS, the input space of the RBFNN is evenly divided into a finite number of partitions and a synthesized objective function is developed using synthesized samples from each partition. In addition to the current approximation error, the neural network also updates its weights according to the recorded data from the partition being visited. Compared with classical training methods including the forgetting factor recursive least squares (FFRLS) and stochastic gradient descent (SGD) methods, SMRLS achieves improved learning speed and generalization capability, which are demonstrated by corresponding simulation results.

eess.SY↗

arXiv2023

Trajectory-aware Principal Manifold Framework for Data Augmentation and Image Generation

Data augmentation for deep learning benefits model training, image transformation, medical imaging analysis and many other fields. Many existing methods generate new samples from a parametric distribution, like the Gaussian, with little attention to generate samples along the data manifold in either the input or feature space. In this paper, we verify that there are theoretical and practical advantages of using the principal manifold hidden in the feature space than the Gaussian distribution. We then propose a novel trajectory-aware principal manifold framework to restore the manifold backbone and generate samples along a specific trajectory. On top of the autoencoder architecture, we further introduce an intrinsic dimension regularization term to make the manifold more compact and enable few-shot image generation. Experimental results show that the novel framework is able to extract more compact manifold representation, improve classification accuracy and generate smooth transformation among few samples.

cs.CV↗

arXiv2023

NTIRE 2023 Quality Assessment of Video Enhancement Challenge

This paper reports on the NTIRE 2023 Quality Assessment of Video Enhancement Challenge, which will be held in conjunction with the New Trends in Image Restoration and Enhancement Workshop (NTIRE) at CVPR 2023. This challenge is to address a major challenge in the field of video processing, namely, video quality assessment (VQA) for enhanced videos. The challenge uses the VQA Dataset for Perceptual Video Enhancement (VDPVE), which has a total of 1211 enhanced videos, including 600 videos with color, brightness, and contrast enhancements, 310 videos with deblurring, and 301 deshaked videos. The challenge has a total of 167 registered participants. 61 participating teams submitted their prediction results during the development phase, with a total of 3168 submissions. A total of 176 submissions were submitted by 37 participating teams during the final testing phase. Finally, 19 participating teams submitted their models and fact sheets, and detailed the methods they used. Some methods have achieved better results than baseline methods, and the winning methods have demonstrated superior prediction performance.

cs.CV↗

arXiv2023

A natural NISQ model of quantum self-attention mechanism

The self-attention mechanism (SAM) has demonstrated remarkable success in various applications. However, training SAM on classical computers becomes computationally challenging as the number of trainable parameters grows. Quantum neural networks (QNNs) have been developed as a novel learning model that promises to provide speedup for pattern recognition using near-term Noisy Intermediate-Scale Quantum (NISQ) computers. In this work, we propose a completely natural way of implementing SAM in QNNs, resulting in the quantum self-attention mechanism (QSAM). The fundamental operations of SAM, such as calculating attention scores and producing attention features, are realized by only designing the data encoding and ansatz architecture appropriately. As these are the fundamental components of QNNs, our QSAM can be executed efficiently on near-term NISQ devices. Our QSAM models achieve better performance in terms of both accuracy and circuit complexity on the text categorization task. Moreover, the robustness of QSAM against various typical types of quantum noise is demonstrated, indicating the model's suitability for NISQ devices. The present QSAM will serve as the fundamental building blocks for developing large models of quantum attention neural networks for quantum advantageous applications.

quant-ph↗