arXiv subjects

Wei Guo

Publications and source records attributed to Wei Guo.

At least 127 records · Page 7Linked to original sources

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

RoboBrain 2.0 Technical Report

We introduce RoboBrain 2.0, our latest generation of embodied vision-language foundation models, designed to unify perception, reasoning, and planning for complex embodied tasks in physical environments. It comes in two variants: a lightweight 7B model and a full-scale 32B model, featuring a heterogeneous architecture with a vision encoder and a language model. Despite its compact size, RoboBrain 2.0 achieves strong performance across a wide spectrum of embodied reasoning tasks. On both spatial and temporal benchmarks, the 32B variant achieves leading results, surpassing prior open-source and proprietary models. In particular, it supports key real-world embodied AI capabilities, including spatial understanding (e.g., affordance prediction, spatial referring, trajectory forecasting) and temporal decision-making (e.g., closed-loop interaction, multi-agent long-horizon planning, and scene graph updating). This report details the model architecture, data construction, multi-stage training strategies, infrastructure and practical applications. We hope RoboBrain 2.0 advances embodied AI research and serves as a practical step toward building generalist embodied agents. The code, checkpoint and benchmark are available at https://superrobobrain.github.io.

cs.RO↗

arXiv2025

High-order Adaptive Rank Integrators for Multi-scale Linear Kinetic Transport Equations in the Hierarchical Tucker Format

In this paper, we present a new adaptive rank approximation technique for computing solutions to the high-dimensional linear kinetic transport equation. The approach we propose is based on a macro-micro decomposition of the kinetic model in which the angular domain is discretized with a tensor product quadrature rule under the discrete ordinates method. To address the challenges associated with the curse of dimensionality, the proposed low-rank method is cast in the framework of the hierarchical Tucker decomposition. The adaptive rank integrators we propose are built upon high-order discretizations for both time and space. In particular, this work considers implicit-explicit discretizations for time and finite-difference weighted-essentially non-oscillatory discretizations for space. The high-order singular value decomposition is used to perform low-rank truncation of the high-dimensional time-dependent distribution function. The methods are applied to several benchmark problems, where we compare the solution quality and measure compression achieved by the adaptive rank methods against their corresponding full-grid methods. We also demonstrate the benefits of high-order discretizations in the proposed low-rank framework.

math.NA↗

arXiv2025

FoMo4Wheat: Toward reliable crop vision foundation models with globally curated data

Vision-driven field monitoring is central to digital agriculture, yet models built on general-domain pretrained backbones often fail to generalize across tasks, owing to the interaction of fine, variable canopy structures with fluctuating field conditions. We present FoMo4Wheat, one of the first crop-domain vision foundation model pretrained with self-supervision on ImAg4Wheat, the largest and most diverse wheat image dataset to date (2.5 million high-resolution images collected over a decade at 30 global sites, spanning >2,000 genotypes and >500 environmental conditions). This wheat-specific pretraining yields representations that are robust for wheat and transferable to other crops and weeds. Across ten in-field vision tasks at canopy and organ levels, FoMo4Wheat models consistently outperform state-of-the-art models pretrained on general-domain dataset. These results demonstrate the value of crop-specific foundation models for reliable in-field perception and chart a path toward a universal crop foundation model with cross-species and cross-task capabilities. FoMo4Wheat models and the ImAg4Wheat dataset are publicly available online: https://github.com/PheniX-Lab/FoMo4Wheat and https://huggingface.co/PheniX-Lab/FoMo4Wheat. The demonstration website is: https://fomo4wheat.phenix-lab.com/.

cs.CV↗

arXiv2025

An Inexact Low-Rank Source Iteration for Steady-State Radiative Transfer Equation with Diffusion Synthetic Acceleration

We propose an inexact low-rank source iteration with diffusion synthetic acceleration (SI-DSA) for solving the multidimensional steady-state radiative transfer equation (RTE) in the second-order formulation. The angular flux is represented in either a low-rank matrix or hierarchical Tucker tensor (HTT) format, enabling substantial reductions in computational resources. Each SI step is solved using a preconditioned low-rank conjugate gradient (CG) method with a diffusion preconditioner. To further improve efficiency, we introduce an adaptive inexact strategy that dynamically relaxes the inner CG tolerance during early SI iterations. The method exploits the tensor-product structure of the discretized operators to perform all matrix-vector operations in low-rank form. Numerical experiments on 2D2V benchmark problems, including diffusion-dominated, transport-dominated, and multiscale problems, demonstrate that the proposed approach achieves errors on the order of $10^{-4}$ to $10^{-5}$ relative to full-rank reference solutions, while reducing the degrees of freedom by up to two orders of magnitude. In the diffusion-dominated case, the low-rank solver achieves speedups exceeding $90\times$ over its full-rank counterpart and remains competitive in solving challenging transport-dominated and multiscale problems while providing substantial storage savings. To our knowledge, this work provides the first low-rank SI-DSA framework for multidimensional steady-state RTE.

math.NA↗

arXiv2025

Higher derivative corrections to Kerr-AdS black hole thermodynamics

Instead of the much more involved covariant counterterm method, we apply the well justified background subtraction method to calculate the first order corrections to Kerr-AdS black hole thermodynamics induced by the higher derivative terms up to the cubic of Riemann tensor, where the computation is further simplified by the decomposition trick for the bulk action. The validity of our results is further substantiated by examining the corrections induced by the Gauss-Bonnet term. Moreover, by comparing our results with those obtained via the ADM and Wald formulas in Lorentzian signature, we can extract some generic information about the first order corrected black hole solution induced by each higher derivative term.

hep-th↗

arXiv2025

JMA: a General Algorithm to Craft Nearly Optimal Targeted Adversarial Example

Most of the approaches proposed so far to craft targeted adversarial examples against Deep Learning classifiers are highly suboptimal and typically rely on increasing the likelihood of the target class, thus implicitly focusing on one-hot encoding settings. In this paper, a more general, theoretically sound, targeted attack is proposed, which resorts to the minimization of a Jacobian-induced Mahalanobis distance term, taking into account the effort (in the input space) required to move the latent space representation of the input sample in a given direction. The minimization is solved by exploiting the Wolfe duality theorem, reducing the problem to the solution of a Non-Negative Least Square (NNLS) problem. The proposed algorithm (referred to as JMA) provides an optimal solution to a linearised version of the adversarial example problem originally introduced by Szegedy et al. The results of the experiments confirm the generality of the proposed attack which is proven to be effective under a wide variety of output encoding schemes. Noticeably, JMA is also effective in a multi-label classification scenario, being capable to induce a targeted modification of up to half the labels in complex multi-label classification scenarios, a capability that is out of reach of all the attacks proposed so far. As a further advantage, JMA requires very few iterations, thus resulting more efficient than existing methods.

cs.LG↗

arXiv2025

FuXi-β: Towards a Lightweight and Fast Large-Scale Generative Recommendation Model

Scaling laws for autoregressive generative recommenders reveal potential for larger, more versatile systems but mean greater latency and training costs. To accelerate training and inference, we investigated the recent generative recommendation models HSTU and FuXi-$α$, identifying two efficiency bottlenecks: the indexing operations in relative temporal attention bias and the computation of the query-key attention map. Additionally, we observed that relative attention bias in self-attention mechanisms can also serve as attention maps. Previous works like Synthesizer have shown that alternative forms of attention maps can achieve similar performance, naturally raising the question of whether some attention maps are redundant. Through empirical experiments, we discovered that using the query-key attention map might degrade the model's performance in recommendation tasks. To address these bottlenecks, we propose a new framework applicable to Transformer-like recommendation models. On one hand, we introduce Functional Relative Attention Bias, which avoids the time-consuming operations of the original relative attention bias, thereby accelerating the process. On the other hand, we remove the query-key attention map from the original self-attention layer and design a new Attention-Free Token Mixer module. Furthermore, by applying this framework to FuXi-$α$, we introduce a new model, FuXi-$β$. Experiments across multiple datasets demonstrate that FuXi-$β$ outperforms previous state-of-the-art models and achieves significant acceleration compared to FuXi-$α$, while also adhering to the scaling law. Notably, FuXi-$β$ shows an improvement of 27% to 47% in the NDCG@10 metric on large-scale industrial datasets compared to FuXi-$α$. Our code is available in a public repository: https://github.com/USTC-StarTeam/FuXi-beta

cs.IR↗

arXiv2025

Gotta Hear Them All: Towards Sound Source Aware Audio Generation

Audio synthesis has broad applications in multimedia. Recent advancements have made it possible to generate relevant audios from inputs describing an audio scene, such as images or texts. However, the immersiveness and expressiveness of the generation are limited. One possible problem is that existing methods solely rely on the global scene and overlook details of local sounding objects (i.e., sound sources). To address this issue, we propose a Sound Source-Aware Audio (SS2A) generator. SS2A is able to locally perceive multimodal sound sources from a scene with visual detection and cross-modality translation. It then contrastively learns a Cross-Modal Sound Source (CMSS) Manifold to semantically disambiguate each source. Finally, we attentively mix their CMSS semantics into a rich audio representation, from which a pretrained audio generator outputs the sound. To model the CMSS manifold, we curate a novel single-sound-source visual-audio dataset VGGS3 from VGGSound. We also design a Sound Source Matching Score to clearly measure localized audio relevance. With the effectiveness of explicit sound source modeling, SS2A achieves state-of-the-art performance in extensive image-to-audio tasks. We also qualitatively demonstrate SS2A's ability to achieve intuitive synthesis control by compositing vision, text, and audio conditions. Furthermore, we show that our sound source modeling can achieve competitive video-to-audio performance with a straightforward temporal aggregation mechanism.

cs.MM↗

arXiv2025

WPTrack: A Wi-Fi and Pressure Insole Fusion System for Single Target Tracking

As the Internet of Things (IoT) continues to evolve, indoor location has become a critical element for enabling smart homes, behavioral monitoring, and elderly care. Existing WiFi-based human tracking solutions typically require specialized equipment or multiple Wi-Fi links, a limitation in most indoor settings where only a single pair of Wi-Fi devices is usually available. However, despite efforts to implement human tracking using one Wi-Fi link, significant challenges remain, such as difficulties in acquiring initial positions and blind spots in DFS estimation of tangent direction. To address these challenges, this paper proposes WPTrack, the first Wi-Fi and Pressure Insoles Fusion System for Single Target Tracking. WPTrack collects Channel State Information (CSI) from a single Wi-Fi link and pressure data from 90 insole sensors. The phase difference and Doppler velocity are computed from the CSI, while the pressure sensor data is used to calculate walking velocity. Then, we propose the CSI-pressure fusion model, integrating CSI and pressure data to accurately determine initial positions and facilitate precise human tracking. The simulation results show that the initial position localization accuracy ranges from 0.02 cm to 42.55 cm. The trajectory tracking results obtained from experimental data collected in a real-world environment closely align with the actual trajectory.

cs.NI↗

arXiv2025

H2Tune: Federated Foundation Model Fine-Tuning with Hybrid Heterogeneity

Different from existing federated fine-tuning (FFT) methods for foundation models, hybrid heterogeneous federated fine-tuning (HHFFT) is an under-explored scenario where clients exhibit double heterogeneity in model architectures and downstream tasks. This hybrid heterogeneity introduces two significant challenges: 1) heterogeneous matrix aggregation, where clients adopt different large-scale foundation models based on their task requirements and resource limitations, leading to dimensional mismatches during LoRA parameter aggregation; and 2) multi-task knowledge interference, where local shared parameters, trained with both task-shared and task-specific knowledge, cannot ensure only task-shared knowledge is transferred between clients. To address these challenges, we propose H2Tune, a federated foundation model fine-tuning with hybrid heterogeneity. Our framework H2Tune consists of three key components: (i) sparsified triple matrix decomposition to align hidden dimensions across clients through constructing rank-consistent middle matrices, with adaptive sparsification based on client resources; (ii) relation-guided matrix layer alignment to handle heterogeneous layer structures and representation capabilities; and (iii) alternating task-knowledge disentanglement mechanism to decouple shared and specific knowledge of local model parameters through alternating optimization. Theoretical analysis proves a convergence rate of O(1/\sqrt{T}). Extensive experiments show our method achieves up to 15.4% accuracy improvement compared to state-of-the-art baselines. Our code is available at https://anonymous.4open.science/r/H2Tune-1407.

cs.LG↗

arXiv2025

Proto-EVFL: Enhanced Vertical Federated Learning via Dual Prototype with Extremely Unaligned Data

In vertical federated learning (VFL), multiple enterprises address aligned sample scarcity by leveraging massive locally unaligned samples to facilitate collaborative learning. However, unaligned samples across different parties in VFL can be extremely class-imbalanced, leading to insufficient feature representation and limited model prediction space. Specifically, class-imbalanced problems consist of intra-party class imbalance and inter-party class imbalance, which can further cause local model bias and feature contribution inconsistency issues, respectively. To address the above challenges, we propose Proto-EVFL, an enhanced VFL framework via dual prototypes. We first introduce class prototypes for each party to learn relationships between classes in the latent space, allowing the active party to predict unseen classes. We further design a probabilistic dual prototype learning scheme to dynamically select unaligned samples by conditional optimal transport cost with class prior probability. Moreover, a mixed prior guided module guides this selection process by combining local and global class prior probabilities. Finally, we adopt an \textit{adaptive gated feature aggregation strategy} to mitigate feature contribution inconsistency by dynamically weighting and aggregating local features across different parties. We proved that Proto-EVFL, as the first bi-level optimization framework in VFL, has a convergence rate of 1/\sqrt T. Extensive experiments on various datasets validate the superiority of our Proto-EVFL. Even in a zero-shot scenario with one unseen class, it outperforms baselines by at least 6.97%

cs.LG↗

arXiv2025

Experimental and theoretical evidence of universality in superfluid vortex reconnections

The minimum separation between reconnecting vortices in fluids and superfluids obeys a universal scaling law with respect to time. The pre-reconnection and the post-reconnection prefactors of this scaling law are different, a property related to irreversibility and to energy transfer and dissipation mechanisms. In the present work, we determine the temperature dependence of these prefactors in superfluid helium from experiments and a numeric model which fully accounts for the coupled dynamics of the superfluid vortex lines and the thermal normal fluid component. At all temperatures, we observe a pre- and post-reconnection asymmetry similar to that observed in other superfluids and in classical viscous fluids, indicating that vortex reconnections display a universal behaviour independent of the small-scale regularising dynamics. We also numerically show that each vortex reconnection event represents a sudden injection of energy in the normal fluid. Finally we argue that in a turbulent flow, these punctuated energy injections can sustain the normal fluid in a perturbed state, provided that the density of superfluid vortices is large enough.

cond-mat.quant-gas↗

arXiv2025

Thought-Augmented Planning for LLM-Powered Interactive Recommender Agent

Interactive recommendation is a typical information-seeking task that allows users to interactively express their needs through natural language and obtain personalized recommendations. Large language model-powered (LLM-powered) agents have become a new paradigm in interactive recommendations, effectively capturing users' real-time needs and enhancing personalized experiences. However, due to limited planning and generalization capabilities, existing formulations of LLM-powered interactive recommender agents struggle to effectively address diverse and complex user intents, such as intuitive, unrefined, or occasionally ambiguous requests. To tackle this challenge, we propose a novel thought-augmented interactive recommender agent system (TAIRA) that addresses complex user intents through distilled thought patterns. Specifically, TAIRA is designed as an LLM-powered multi-agent system featuring a manager agent that orchestrates recommendation tasks by decomposing user needs and planning subtasks, with its planning capacity strengthened through Thought Pattern Distillation (TPD), a thought-augmentation method that extracts high-level thoughts from the agent's and human experts' experiences. Moreover, we designed a set of user simulation schemes to generate personalized queries of different difficulties and evaluate the recommendations based on specific datasets. Through comprehensive experiments conducted across multiple datasets, TAIRA exhibits significantly enhanced performance compared to existing methods. Notably, TAIRA shows a greater advantage on more challenging tasks while generalizing effectively on novel tasks, further validating its superiority in managing complex user intents within interactive recommendation systems. The code is publicly available at:https://github.com/Alcein/TAIRA.

cs.CL↗

arXiv2025

Point Cloud Environment-Based Channel Knowledge Map Construction

Channel knowledge map (CKM) provides certain levels of channel state information (CSI) for an area of interest, serving as a critical enabler for environment-aware communications by reducing the overhead of frequent CSI acquisition. However, existing CKM construction schemes adopt over-simplified environment information, which significantly compromises their accuracy. To address this issue, this work proposes a joint model- and data-driven approach to construct CKM by leveraging point cloud environmental data along with a few samples of location-tagged channel information. First, we propose a novel point selector to identify subsets of point cloud that contain environmental information relevant to multipath channel gains, by constructing a set of co-focal ellipsoids based on different time of arrival (ToAs). Then, we trained a neural channel gain estimator to learn the mapping between each selected subset and its corresponding channel gain, using a real-world dataset we collected through field measurements, comprising environmental point clouds and corresponding channel data. Finally, experimental results demonstrate that: For CKM construction of power delay profile (PDP), the proposed method achieves a root mean squared error (RMSE) of 2.95 dB, significantly lower than the 7.32 dB achieved by the conventional ray-tracing method; for CKM construction of received power values, i.e., radio map, it achieves an RMSE of 1.04 dB, surpassing the Kriging interpolation method with an RMSE of 1.68 dB.

eess.SP↗

arXiv2025

PCF-Grasp: Converting Point Completion to Geometry Feature to Enhance 6-DoF Grasp

The 6-Degree of Freedom (DoF) grasp method based on point clouds has shown significant potential in enabling robots to grasp target objects. However, most existing methods are based on the point clouds (2.5D points) generated from single-view depth images. These point clouds only have one surface side of the object providing incomplete geometry information, which mislead the grasping algorithm to judge the shape of the target object, resulting in low grasping accuracy. Humans can accurately grasp objects from a single view by leveraging their geometry experience to estimate object shapes. Inspired by humans, we propose a novel 6-DoF grasping framework that converts the point completion results as object shape features to train the 6-DoF grasp network. Here, point completion can generate approximate complete points from the 2.5D points similar to the human geometry experience, and converting it as shape features is the way to utilize it to improve grasp efficiency. Furthermore, due to the gap between the network generation and actual execution, we integrate a score filter into our framework to select more executable grasp proposals for the real robot. This enables our method to maintain a high grasp quality in any camera viewpoint. Extensive experiments demonstrate that utilizing complete point features enables the generation of significantly more accurate grasp proposals and the inclusion of a score filter greatly enhances the credibility of real-world robot grasping. Our method achieves a 17.8\% success rate higher than the state-of-the-art method in real-world experiments.

cs.RO↗

arXiv2025

AgriCHN: A Comprehensive Cross-domain Resource for Chinese Agricultural Named Entity Recognition

Agricultural named entity recognition is a specialized task focusing on identifying distinct agricultural entities within vast bodies of text, including crops, diseases, pests, and fertilizers. It plays a crucial role in enhancing information extraction from extensive agricultural text resources. However, the scarcity of high-quality agricultural datasets, particularly in Chinese, has resulted in suboptimal performance when employing mainstream methods for this purpose. Most earlier works only focus on annotating agricultural entities while overlook the profound correlation of agriculture with hydrology and meteorology. To fill this blank, we present AgriCHN, a comprehensive open-source Chinese resource designed to promote the accuracy of automated agricultural entity annotation. The AgriCHN dataset has been meticulously curated from a wealth of agricultural articles, comprising a total of 4,040 sentences and encapsulating 15,799 agricultural entity mentions spanning 27 diverse entity categories. Furthermore, it encompasses entities from hydrology to meteorology, thereby enriching the diversity of entities considered. Data validation reveals that, compared with relevant resources, AgriCHN demonstrates outstanding data quality, attributable to its richer agricultural entity types and more fine-grained entity divisions. A benchmark task has also been constructed using several state-of-the-art neural NER models. Extensive experimental results highlight the significant challenge posed by AgriCHN and its potential for further research.

cs.CL↗

arXiv2025

Joint Transmit and Receive Beamforming for Tri-directional Coil-Based Magnetic Induction Communications

In this paper, we enhance the omnidirectional coverage performance of tri-directional coil-based magnetic induction communication (TC-MIC) and reduce the pathloss with a joint transmit and receive magnetic beamforming method. An iterative optimization algorithm incorporating the transmit current vector and receive weight matrix is developed to minimize the pathloss under constant transmit power constraints. We formulate the mathematical models for the mutual inductance of tri-directional coils, receive power, and pathloss. The optimization problem is decomposed into Rayleigh quotient extremum optimization for transmit currents and Cauchy-Schwarz inequality-constrained optimization for receive weights, with an alternating iterative algorithm to approach the global optimum. Numerical results demonstrate that the proposed algorithm converges within an average of 13.6 iterations, achieving up to 54% pathloss reduction compared with equal power allocation schemes. The joint optimization approach exhibits superior angular robustness, maintaining pathloss fluctuation smaller than 2 dB, and reducing fluctuation of pathloss by approximately 45% compared with single-parameter optimization methods.

eess.SP↗