arXiv subjects

Yi Yang

Publications and source records attributed to Yi Yang.

At least 433 records · Page 24Linked to original sources

arXiv2025

ProtChatGPT: Towards Understanding Proteins with Large Language Models

Protein research is crucial in various fundamental disciplines, but understanding their intricate structure-function relationships remains challenging. Recent Large Language Models (LLMs) have made significant strides in comprehending task-specific knowledge, suggesting the potential for ChatGPT-like systems specialized in protein to facilitate basic research. In this work, we introduce ProtChatGPT, which aims at learning and understanding protein structures via natural languages. ProtChatGPT enables users to upload proteins, ask questions, and engage in interactive conversations to produce comprehensive answers. The system comprises protein encoders, a Protein-Language Pertaining Transformer (PLP-former), a projection adapter, and an LLM. The protein first undergoes protein encoders and PLP-former to produce protein embeddings, which are then projected by the adapter to conform with the LLM. The LLM finally combines user questions with projected embeddings to generate informative answers. Experiments show that ProtChatGPT can produce promising responses to proteins and their corresponding questions. We hope that ProtChatGPT could form the basis for further exploration and application in protein research. Code and our pre-trained model will be publicly available.

cs.CE↗

arXiv2025

MCRL4OR: Multimodal Contrastive Representation Learning for Off-Road Environmental Perception

Most studies on environmental perception for autonomous vehicles (AVs) focus on urban traffic environments, where the objects/stuff to be perceived are mainly from man-made scenes and scalable datasets with dense annotations can be used to train supervised learning models. By contrast, it is hard to densely annotate a large-scale off-road driving dataset manually due to the inherently unstructured nature of off-road environments. In this paper, we propose a Multimodal Contrastive Representation Learning approach for Off-Road environmental perception, namely MCRL4OR. This approach aims to jointly learn three encoders for processing visual images, locomotion states, and control actions by aligning the locomotion states with the fused features of visual images and control actions within a contrastive learning framework. The causation behind this alignment strategy is that the inertial locomotion state is the result of taking a certain control action under the current landform/terrain condition perceived by visual sensors. In experiments, we pre-train the MCRL4OR with a large-scale off-road driving dataset and adopt the learned multimodal representations for various downstream perception tasks in off-road driving scenarios. The superior performance in downstream tasks demonstrates the advantages of the pre-trained multimodal representations. The codes can be found in \url{https://github.com/1uciusy/MCRL4OR}.

cs.RO↗

arXiv2025

3D Object Manipulation in a Single Image using Generative Models

Object manipulation in images aims to not only edit the object's presentation but also gift objects with motion. Previous methods encountered challenges in concurrently handling static editing and dynamic generation, while also struggling to achieve fidelity in object appearance and scene lighting. In this work, we introduce \textbf{OMG3D}, a novel framework that integrates the precise geometric control with the generative power of diffusion models, thus achieving significant enhancements in visual performance. Our framework first converts 2D objects into 3D, enabling user-directed modifications and lifelike motions at the geometric level. To address texture realism, we propose CustomRefiner, a texture refinement module that pre-train a customized diffusion model, aligning the details and style of coarse renderings of 3D rough model with the original image, further refine the texture. Additionally, we introduce IllumiCombiner, a lighting processing module that estimates and corrects background lighting to match human visual perception, resulting in more realistic shadow effects. Extensive experiments demonstrate the outstanding visual performance of our approach in both static and dynamic scenarios. Remarkably, all these steps can be done using one NVIDIA 3090. Project page is at https://whalesong-zrs.github.io/OMG3D-projectpage/

cs.CV↗

arXiv2025

Higher-Order Topological Insulators via Momentum-Space Nonsymmorphic Symmetries

The topology of the Brillouin zone, foundational in topological physics, is always assumed to be a torus. We theoretically report the construction of Brillouin real projective plane ($\mathrm{RP}^2$) and the appearance of quadrupole insulating phase, which are enabled by momentum-space nonsymmorphic symmetries stemming from $\mathbb{Z}_2$ synthetic gauge fields. We show that the momentum-space nonsymmorphic symmetries quantize bulk polarization and Wannier-sector polarization nonlocally across different momenta, resulting in quantized corner charges and an isotropic binary bulk quadrupole phase diagram, where the phase transition is triggered by a bulk energy gap closing. Under open boundary conditions, the nontrivial bulk quadrupole phase manifests either trivial or nontrivial edge polarization, resulting from the violation of momentum-space nonsymmorphic symmetries under lattice termination. We present a concrete design for the $\mathrm{RP}^2$ quadrupole insulator based on acoustic resonator arrays and discuss its feasibility in optics, mechanics, and electrical circuits. Our results show that deforming the Brillouin manifold creates opportunities for realizing high-order band topology.

cond-mat.mes-hall↗

arXiv2025

Adaptive Sampled Softmax with Inverted Multi-Index: Methods, Theory and Applications

The softmax function is a cornerstone of multi-class classification, integral to a wide range of machine learning applications, from large-scale retrieval and ranking models to advanced large language models. However, its computational cost grows linearly with the number of classes, which becomes prohibitively expensive in scenarios with millions or even billions of classes. The sampled softmax, which relies on self-normalized importance sampling, has emerged as a powerful alternative, significantly reducing computational complexity. Yet, its estimator remains unbiased only when the sampling distribution matches the true softmax distribution. To improve both approximation accuracy and sampling efficiency, we propose the MIDX Sampler, a novel adaptive sampling strategy based on an inverted multi-index approach. Concretely, we decompose the softmax probability into several multinomial probabilities, each associated with a specific set of codewords and the last associated with the residual score of queries, thus reducing time complexity to the number of codewords instead of the number of classes. To further boost efficiency, we replace the query-specific residual probability with a simple uniform distribution, simplifying the computation while retaining high performance. Our method is backed by rigorous theoretical analysis, addressing key concerns such as sampling bias, gradient bias, convergence rates, and generalization error bounds. The results demonstrate that a smaller divergence from the ideal softmax distribution leads to faster convergence and improved generalization. Extensive experiments on large-scale language models, sequential recommenders, and extreme multi-class classification tasks confirm that the MIDX-Sampler delivers superior effectiveness and efficiency compared to existing approaches.

cs.LG↗

arXiv2025

3DIS-FLUX: simple and efficient multi-instance generation with DiT rendering

The growing demand for controllable outputs in text-to-image generation has driven significant advancements in multi-instance generation (MIG), enabling users to define both instance layouts and attributes. Currently, the state-of-the-art methods in MIG are primarily adapter-based. However, these methods necessitate retraining a new adapter each time a more advanced model is released, resulting in significant resource consumption. A methodology named Depth-Driven Decoupled Instance Synthesis (3DIS) has been introduced, which decouples MIG into two distinct phases: 1) depth-based scene construction and 2) detail rendering with widely pre-trained depth control models. The 3DIS method requires adapter training solely during the scene construction phase, while enabling various models to perform training-free detail rendering. Initially, 3DIS focused on rendering techniques utilizing U-Net architectures such as SD1.5, SD2, and SDXL, without exploring the potential of recent DiT-based models like FLUX. In this paper, we present 3DIS-FLUX, an extension of the 3DIS framework that integrates the FLUX model for enhanced rendering capabilities. Specifically, we employ the FLUX.1-Depth-dev model for depth map controlled image generation and introduce a detail renderer that manipulates the Attention Mask in FLUX's Joint Attention mechanism based on layout information. This approach allows for the precise rendering of fine-grained attributes of each instance. Our experimental results indicate that 3DIS-FLUX, leveraging the FLUX model, outperforms the original 3DIS method, which utilized SD2 and SDXL, and surpasses current state-of-the-art adapter-based methods in terms of both performance and image quality. Project Page: https://limuloo.github.io/3DIS/.

cs.CV↗

arXiv2025

Efficient algorithms for surface density of states in topological photonic and acoustic systems

Topological photonics and acoustics have recently garnered wide research interests for their topological ability to manipulate the light and sound at surfaces. Conventionally, the supercell technique is the standard approach to calculating these boundary effects, whereas it consumes increasingly large computational resources as the supercell size grows. Additionally, it falls short in differentiating the surface states at opposite boundaries and from bulk states due to the finite size of systems. To overcome the limitations, here we provide two complementary efficient methods for obtaining the ideal topological surface states of a semi-infinite system. The first one is the cyclic reduction method, which is based on iteratively inverting the Hamiltonian for a single unit cell, and the other is the transfer matrix method, which relies on the eigenanalysis of a transfer matrix for a pair of unit cells. Benchmarks show that, compared to the traditional supercell method, the cyclic reduction method can reduce both memory and time consumption by two orders of magnitude; the transfer matrix method can reduce memory by an order of magnitude, take less than half the time, and achieve high accuracy. Our methods are applicable to more complex scenarios, such as coated structures, heterostructures, and sandwiched structures. As examples, the surface-density-of-states spectra of photonic Chern insulators, valley photonic crystals, and acoustic topological insulators are demonstrated. Our computational schemes enable direct comparisons with near-field scanning measurements and expedite the exploration of topological artificial materials and the design of topological devices.

physics.optics↗

arXiv2025

Noise-Tolerant Hybrid Prototypical Learning with Noisy Web Data

We focus on the challenging problem of learning an unbiased classifier from a large number of potentially relevant but noisily labeled web images given only a few clean labeled images. This problem is particularly practical because it reduces the expensive annotation costs by utilizing freely accessible web images with noisy labels. Typically, prototypes are representative images or features used to classify or identify other images. However, in the few clean and many noisy scenarios, the class prototype can be severely biased due to the presence of irrelevant noisy images. The resulting prototypes are less compact and discriminative, as previous methods do not take into account the diverse range of images in the noisy web image collections. On the other hand, the relation modeling between noisy and clean images is not learned for the class prototype generation in an end-to-end manner, which results in a suboptimal class prototype. In this article, we introduce a similarity maximization loss named SimNoiPro. Our SimNoiPro first generates noise-tolerant hybrid prototypes composed of clean and noise-tolerant prototypes and then pulls them closer to each other. Our approach considers the diversity of noisy images by explicit division and overcomes the optimization discrepancy issue. This enables better relation modeling between clean and noisy images and helps extract judicious information from the noisy image set. The evaluation results on two extended few-shot classification benchmarks confirm that our SimNoiPro outperforms prior methods in measuring image relations and cleaning noisy data.

cs.CV↗

arXiv2024

Combating Label Noise With A General Surrogate Model For Sample Selection

Modern deep learning systems are data-hungry. Learning with web data is one of the feasible solutions, but will introduce label noise inevitably, which can hinder the performance of deep neural networks. Sample selection is an effective way to deal with label noise. The key is to separate clean samples based on some criterion. Previous methods pay more attention to the small loss criterion where small-loss samples are regarded as clean ones. Nevertheless, such a strategy relies on the learning dynamics of each data instance. Some noisy samples are still memorized due to frequently occurring corrupted learning patterns. To tackle this problem, a training-free surrogate model is preferred, freeing from the effect of memorization. In this work, we propose to leverage the vision-language surrogate model CLIP to filter noisy samples automatically. CLIP brings external knowledge to facilitate the selection of clean samples with its ability of text-image alignment. Furthermore, a margin adaptive loss is designed to regularize the selection bias introduced by CLIP, providing robustness to label noise. We validate the effectiveness of our proposed method on both real-world and synthetic noisy datasets. Our method achieves significant improvement without CLIP involved during the inference stage.

cs.CV↗

arXiv2024

A Predefined-Time Convergent and Noise-Tolerant Zeroing Neural Network Model for Time Variant Quadratic Programming With Application to Robot Motion Planning

This paper develops a predefined-time convergent and noise-tolerant fractional-order zeroing neural network (PTC-NT-FOZNN) model, innovatively engineered to tackle time-variant quadratic programming (TVQP) challenges. The PTC-NT-FOZNN, stemming from a novel iteration within the variable-gain ZNN spectrum, known as FOZNNs, features diminishing gains over time and marries noise resistance with predefined-time convergence, making it ideal for energy-efficient robotic motion planning tasks. The PTC-NT-FOZNN enhances traditional ZNN models by incorporating a newly developed activation function that promotes optimal convergence irrespective of the model's order. When evaluated against six established ZNNs, the PTC-NT-FOZNN, with parameters $0 < α\leq 1$, demonstrates enhanced positional precision and resilience to additive noises, making it exceptionally suitable for TVQP tasks. Thorough practical assessments, including simulations and experiments using a Flexiv Rizon robotic arm, confirm the PTC-NT-FOZNN's capabilities in achieving precise tracking and high computational efficiency, thereby proving its effectiveness for robust kinematic control applications.

cs.RO↗

arXiv2024

Early-time millimeter observations of the nearby Type II SN 2024ggi

The short-lived ionized emission lines in early spectroscopy of the nearby type II supernova SN 2024ggi signify the presence of dense circumstellar matter (CSM) close to its progenitor star. We proposed the Atacama Large Millimeter/submillimeter Array (ALMA) observations by its Director's Discretionary Time program to catch the potential synchrotron radiation associated with the ejecta-CSM interaction. Multi-epoch observations were conducted using ALMA band 6 at +8, +13, and +17 days after the discovery. The data show non-detections at the position of SN 2024ggi with a 3sigma upper limit of less than 0.15 mJy, corresponding to a luminosity of approximately 8*10^24 erg/s/Hz. In this paper, we leverage the non-detections to place constraints on the properties of CSM surrounding SN 2024ggi. We investigate both the Wind and Eruptive models for the radial distribution of CSM, assuming a constant mass-loss rate in the Wind model and a distance-variant mass-loss rate in the Eruptive model. The derived CSM distribution for the Wind model does not align with the early-time spectral features, while the ALMA observations suggest a mass-loss rate of ~ 5*10^-3 Msun/year for the Eruptive model. Conducting multi-epoch millimeter/submillimeter observations shortly after the explosion, with a cadence of a few days, could offer a promising opportunity to capture the observable signature of the Eruptive model.

astro-ph.SR↗

arXiv2024

AINet+: Advancing Superpixel Segmentation via Cascaded Association Implantation

Superpixel segmentation has seen significant progress benefiting from the deep convolutional networks. The typical approach entails initial division of the image into grids, followed by a learning process that assigns each pixel to adjacent grid segments. However, reliance on convolutions with confined receptive fields results in an implicit, rather than explicit, understanding of pixel-grid interactions. This limitation often leads to a deficit of contextual information during the mapping of associations. To counteract this, we introduce the Association Implantation (AI) module, designed to allow networks to explicitly engage with pixel-grid relationships. This module embeds grid features directly into the vicinity of the central pixel and employs convolutional operations on an enlarged window, facilitating an adaptive transfer of knowledge. This approach enables the network to explicitly extract context at the pixel-grid level, which is more aligned with the objectives of superpixel segmentation than mere pixel-wise interactions. By integrating the AI module across various layers, we enable a progressive refinement of pixel-superpixel relationships from coarse to fine. To further enhance the assignment of boundary pixels, we've engineered a boundary-aware loss function. This function aids in the discrimination of boundary-adjacent pixels at the feature level, thereby empowering subsequent modules to precisely identify boundary pixels and enhance overall boundary accuracy. Our method has been rigorously tested on four benchmarks, including BSDS500, NYUv2, ACDC, and ISIC2017, and our model can achieve competitive performance with comparison methods.

cs.CV↗

arXiv2024

Decomposed Prototype Learning for Few-Shot Scene Graph Generation

Today's scene graph generation (SGG) models typically require abundant manual annotations to learn new predicate types. Therefore, it is difficult to apply them to real-world applications with massive uncommon predicate categories whose annotations are hard to collect. In this paper, we focus on Few-Shot SGG (FSSGG), which encourages SGG models to be able to quickly transfer previous knowledge and recognize unseen predicates well with only a few examples. However, current methods for FSSGG are hindered by the high intra-class variance of predicate categories in SGG: On one hand, each predicate category commonly has multiple semantic meanings under different contexts. On the other hand, the visual appearance of relation triplets with the same predicate differs greatly under different subject-object compositions. Such great variance of inputs makes it hard to learn generalizable representation for each predicate category with current few-shot learning (FSL) methods. However, we found that this intra-class variance of predicates is highly related to the composed subjects and objects. To model the intra-class variance of predicates with subject-object context, we propose a novel Decomposed Prototype Learning (DPL) model for FSSGG. Specifically, we first construct a decomposable prototype space to capture diverse semantics and visual patterns of subjects and objects for predicates by decomposing them into multiple prototypes. Afterwards, we integrate these prototypes with different weights to generate query-adaptive predicate representation with more reliable semantics for each query sample. We conduct extensive experiments and compare with various baseline methods to show the effectiveness of our method.

cs.CV↗

arXiv2024

CLIP4STR: A Simple Baseline for Scene Text Recognition with Pre-trained Vision-Language Model

Pre-trained vision-language models~(VLMs) are the de-facto foundation models for various downstream tasks. However, scene text recognition methods still prefer backbones pre-trained on a single modality, namely, the visual modality, despite the potential of VLMs to serve as powerful scene text readers. For example, CLIP can robustly identify regular (horizontal) and irregular (rotated, curved, blurred, or occluded) text in images. With such merits, we transform CLIP into a scene text reader and introduce CLIP4STR, a simple yet effective STR method built upon image and text encoders of CLIP. It has two encoder-decoder branches: a visual branch and a cross-modal branch. The visual branch provides an initial prediction based on the visual feature, and the cross-modal branch refines this prediction by addressing the discrepancy between the visual feature and text semantics. To fully leverage the capabilities of both branches, we design a dual predict-and-refine decoding scheme for inference. We scale CLIP4STR in terms of the model size, pre-training data, and training data, achieving state-of-the-art performance on 13 STR benchmarks. Additionally, a comprehensive empirical study is provided to enhance the understanding of the adaptation of CLIP to STR. Our method establishes a simple yet strong baseline for future STR research with VLMs.

cs.CV↗

arXiv2024

HeadStudio: Text to Animatable Head Avatars with 3D Gaussian Splatting

Creating digital avatars from textual prompts has long been a desirable yet challenging task. Despite the promising results achieved with 2D diffusion priors, current methods struggle to create high-quality and consistent animated avatars efficiently. Previous animatable head models like FLAME have difficulty in accurately representing detailed texture and geometry. Additionally, high-quality 3D static representations face challenges in semantically driving with dynamic priors. In this paper, we introduce \textbf{HeadStudio}, a novel framework that utilizes 3D Gaussian splatting to generate realistic and animatable avatars from text prompts. Firstly, we associate 3D Gaussians with animatable head prior model, facilitating semantic animation on high-quality 3D representations. To ensure consistent animation, we further enhance the optimization from initialization, distillation, and regularization to jointly learn the shape, texture, and animation. Extensive experiments demonstrate the efficacy of HeadStudio in generating animatable avatars from textual prompts, exhibiting appealing appearances. The avatars are capable of rendering high-quality real-time ($\geq 40$ fps) novel views at a resolution of 1024. Moreover, These avatars can be smoothly driven by real-world speech and video. We hope that HeadStudio can enhance digital avatar creation and gain popularity in the community. Code is at: https://github.com/ZhenglinZhou/HeadStudio.

cs.CV↗

arXiv2024

Integrable hierarchy for homogeneous realization of the toroidal Lie algebra $\mathcal{L}^{\rm tor}_{r+1}(\mathfrak{sl}_\ell)$

Starting from a fairly explicit homogeneous realization of the toroidal Lie algebra $\mathcal{L}^{\tor}_{r+1}(\fsl_\ell)$ via a lattice vertex algebra, we derive an integrable hierarchy of Hirota bilinear equations. Moreover, we represent this hierarchy in the form of Lax equations, and show that it is an extension of a certain reduction of the $\ell$-component KP hierarchy.

nlin.SI↗

arXiv2024

Collaborative Hybrid Propagator for Temporal Misalignment in Audio-Visual Segmentation

Audio-visual video segmentation (AVVS) aims to generate pixel-level maps of sound-producing objects that accurately align with the corresponding audio. However, existing methods often face temporal misalignment, where audio cues and segmentation results are not temporally coordinated. Audio provides two critical pieces of information: i) target object-level details and ii) the timing of when objects start and stop producing sounds. Current methods focus more on object-level information but neglect the boundaries of audio semantic changes, leading to temporal misalignment. To address this issue, we propose a Collaborative Hybrid Propagator Framework~(Co-Prop). This framework includes two main steps: Preliminary Audio Boundary Anchoring and Frame-by-Frame Audio-Insert Propagation. To Anchor the audio boundary, we employ retrieval-assist prompts with Qwen large language models to identify control points of audio semantic changes. These control points split the audio into semantically consistent audio portions. After obtaining the control point lists, we propose the Audio Insertion Propagator to process each audio portion using a frame-by-frame audio insertion propagation and matching approach. We curated a compact dataset comprising diverse source conversion cases and devised a metric to assess alignment rates. Compared to traditional simultaneous processing methods, our approach reduces memory requirements and facilitates frame alignment. Experimental results demonstrate the effectiveness of our approach across three datasets and two backbones. Furthermore, our method can be integrated with existing AVVS approaches, offering plug-and-play functionality to enhance their performance.

cs.CV↗

arXiv2024

The Modified Toda Hierarchy

In this paper, modified Toda (mToda) equation is generalized to form an integrable hierarchy in the framework of Sato theory, which is therefore called mToda hierarchy. Inspired by the fact that Toda hierarchy is 2-component generalization of usual KP hierarchy, mToda hierarchy is constructed from bilinear equations of 2-component first modified KP hierarchy, where we provide the corresponding equivalence with Lax formulations. Then it is demonstrated that there are Miura links between Toda and mToda hierarchies, which means the definition of mToda hierarchy here is reasonable. Finally, Darboux transformations of the Toda and mToda hierarchies are also constructed by using the aforementioned Miura links.

nlin.SI↗