arXiv subjects

Yi Li

Publications and source records attributed to Yi Li.

At least 235 records · Page 13Linked to original sources

arXiv2025

Prototype-Based Continual Learning with Label-free Replay Buffer and Cluster Preservation Loss

Continual learning techniques employ simple replay sample selection processes and use them during subsequent tasks. Typically, they rely on labeled data. In this paper, we depart from this by automatically selecting prototypes stored without labels, preserving cluster structures in the latent space across tasks. By eliminating label dependence in the replay buffer and introducing cluster preservation loss, it is demonstrated that the proposed method can maintain essential information from previously encountered tasks while ensuring adaptation to new tasks. "Push-away" and "pull-toward" mechanisms over previously learned prototypes are also introduced for class-incremental and domain-incremental scenarios. These mechanisms ensure the retention of previously learned information as well as adaptation to new classes or domain shifts. The proposed method is evaluated on several benchmarks, including SplitCIFAR100, SplitImageNet32, SplitTinyImageNet, and SplitCaltech256 for class-incremental, as well as R-MNIST and CORe50 for domain-incremental setting using pre-extracted DINOv2 features. Experimental results indicate that the label-free replay-based technique outperforms state-of-the-art continual learning methods and, in some cases, even surpasses offline learning. An unsupervised variant of the proposed technique for the class-incremental setting, avoiding labels use even on incoming data, also demonstrated competitive performance, outperforming particular supervised baselines in some cases. These findings underscore the effectiveness of the proposed framework in retaining prior information and facilitating continual adaptation.

cs.LG↗

arXiv2025

Experimental realization and synchronization of a quantum van der Pol oscillator

Classical self-sustained oscillators, that generate periodic motion without periodic external forcing, are ubiquitous in science and technology. The realization of nonclassical self-oscillators is an important goal of quantum physics. We here present the experimental implementation of a quantum van der Pol oscillator, a paradigmatic autonomous quantum driven-dissipative system with nonlinear damping, using a single trapped atom. We demonstrate the existence of a quantum limit cycle in phase space in the absence of a drive and the occurrence of quantum synchronization when the nonlinear oscillator is externally driven. We additionally show that synchronization can be enhanced with the help of squeezing perpendicular to the direction of the drive and, counterintuitively, linear dissipation. We also observe the bifurcation to a bistable phase-space distribution for large squeezing. Our results pave the way for the exploration of self-sustained quantum oscillators and their application to quantum technology.

quant-ph↗

arXiv2025

Gradient estimates and parabolic frequency under the Laplacian G_2 flow

In this paper, we consider the Laplacian G_2 flow on a closed seven-dimensional manifold M with a closed G_2-structure. We first obtain the gradient estimates of positive solutions of the heat equation under the Laplacian G_2 flow and then we get the Harnack inequality on spacetime. As an application, we prove the monotonicity for positive solutions of the heat equation with bounded Ricci curvature, and get the integral-type Harnack inequality. Besides, we prove the monotonicity of parabolic frequency for positive solutions of the linear heat equation with bounded Bakry-Emery Ricci curvature, and then obtain the backward uniqueness.

math.DG↗

arXiv2025

Observation of thermally activated coherent magnon-magnon coupling in a magnonic hybrid system

We experimentally demonstrate strong magnon-magnon coupling by thermal spin excitations in yttrium iron garnet/permalloy (YIG/Py) hybrid structures using microfocused Brillouin light scattering - an optical technique that enables the detection of zero-wavevector and higher-order wavevector spin waves in a broad frequency range. The thermally activated magnons in the bilayer lead to a hybrid excitation between magnon modes in the conductive Py layer with a wide wavevector range and the first perpendicular standing magnon modes in the insulating YIG layer, facilitated by strong interfacial exchange coupling. To further investigate this coupling, we compare the thermal magnon spectra with the results obtained from electrical excitation and detection methods, which primarily detect the uniform Py mode. The realization of coherent coupling between incoherent (thermal) magnons is important for advancing energy-efficient magnonic devices, particularly in classical as well as quantum spin-wave computing technologies.

cond-mat.mes-hall↗

arXiv2025

Sparse Meets Dense: Unified Generative Recommendations with Cascaded Sparse-Dense Representations

Generative models have recently gained attention in recommendation systems by directly predicting item identifiers from user interaction sequences. However, existing methods suffer from significant information loss due to the separation of stages such as quantization and sequence modeling, hindering their ability to achieve the modeling precision and accuracy of sequential dense retrieval techniques. Integrating generative and dense retrieval methods remains a critical challenge. To address this, we introduce the Cascaded Organized Bi-Represented generAtive retrieval (COBRA) framework, which innovatively integrates sparse semantic IDs and dense vectors through a cascading process. Our method alternates between generating these representations by first generating sparse IDs, which serve as conditions to aid in the generation of dense vectors. End-to-end training enables dynamic refinement of dense representations, capturing both semantic insights and collaborative signals from user-item interactions. During inference, COBRA employs a coarse-to-fine strategy, starting with sparse ID generation and refining them into dense vectors via the generative model. We further propose BeamFusion, an innovative approach combining beam search with nearest neighbor scores to enhance inference flexibility and recommendation diversity. Extensive experiments on public datasets and offline tests validate our method's robustness. Online A/B tests on a real-world advertising platform with over 200 million daily users demonstrate substantial improvements in key metrics, highlighting COBRA's practical advantages.

cs.IR↗

arXiv2025

AlphaPIG: The Nicest Way to Prolong Interactive Gestures in Extended Reality

Mid-air gestures serve as a common interaction modality across Extended Reality (XR) applications, enhancing engagement and ownership through intuitive body movements. However, prolonged arm movements induce shoulder fatigue, known as "Gorilla Arm Syndrome", degrading user experience and reducing interaction duration. Although existing ergonomic techniques derived from Fitts' law (such as reducing target distance, increasing target width, and modifying control-display gain) provide some fatigue mitigation, their implementation in XR applications remains challenging due to the complex balance between user engagement and physical exertion. We present AlphaPIG, a meta-technique designed to Prolong Interactive Gestures by leveraging real-time fatigue predictions. AlphaPIG assists designers in extending and improving XR interactions by enabling automated fatigue-based interventions. Through adjustment of intervention timing and intensity decay rate, designers can explore and control the trade-off between fatigue reduction and potential effects such as decreased body ownership. We validated AlphaPIG's effectiveness through a study (N=22) implementing the widely-used Go-Go technique. Results demonstrated that AlphaPIG significantly reduces shoulder fatigue compared to non-adaptive Go-Go, while maintaining comparable perceived body ownership and agency. Based on these findings, we discuss positive and negative perceptions of the intervention. By integrating real-time fatigue prediction with adaptive intervention mechanisms, AlphaPIG constitutes a critical first step towards creating fatigue-aware applications in XR.

cs.HC↗

arXiv2025

Split Adaptation for Pre-trained Vision Transformers

Vision Transformers (ViTs), extensively pre-trained on large-scale datasets, have become essential to foundation models, allowing excellent performance on diverse downstream tasks with minimal adaptation. Consequently, there is growing interest in adapting pre-trained ViTs across various fields, including privacy-sensitive domains where clients are often reluctant to share their data. Existing adaptation methods typically require direct data access, rendering them infeasible under these constraints. A straightforward solution may be sending the pre-trained ViT to clients for local adaptation, which poses issues of model intellectual property protection and incurs heavy client computation overhead. To address these issues, we propose a novel split adaptation (SA) method that enables effective downstream adaptation while protecting data and models. SA, inspired by split learning (SL), segments the pre-trained ViT into a frontend and a backend, with only the frontend shared with the client for data representation extraction. But unlike regular SL, SA replaces frontend parameters with low-bit quantized values, preventing direct exposure of the model. SA allows the client to add bi-level noise to the frontend and the extracted data representations, ensuring data protection. Accordingly, SA incorporates data-level and model-level out-of-distribution enhancements to mitigate noise injection's impact on adaptation performance. Our SA focuses on the challenging few-shot adaptation and adopts patch retrieval augmentation for overfitting alleviation. Extensive experiments on multiple datasets validate SA's superiority over state-of-the-art methods and demonstrate its defense against advanced data reconstruction attacks while preventing model leakage with minimal computation cost on the client side. The source codes can be found at https://github.com/conditionWang/Split_Adaptation.

cs.CV↗

arXiv2025

SpecGen: Automated Generation of Formal Program Specifications via Large Language Models

Formal program specifications play a crucial role in various stages of software development. However, manually crafting formal program specifications is rather difficult, making the job time-consuming and labor-intensive. It is even more challenging to write specifications that correctly and comprehensively describe the semantics of complex programs. To reduce the burden on software developers, automated specification generation methods have emerged. However, existing methods usually rely on predefined templates or grammar, making them struggle to accurately describe the behavior and functionality of complex real-world programs. To tackle this challenge, we introduce SpecGen, a novel technique for formal program specification generation based on Large Language Models. Our key insight is to overcome the limitations of existing methods by leveraging the code comprehension capability of LLMs. The process of SpecGen consists of two phases. The first phase employs a conversational approach that guides the LLM to generate appropriate specifications for a given program. The second phase, designed for where the LLM fails to generate correct specifications, applies four mutation operators to the model-generated specifications and selects verifiable specifications from the mutated ones through a novel heuristic selection strategy. We evaluate SpecGen on two datasets, including the SV-COMP Java category benchmark and a manually constructed dataset. Experimental results demonstrate that SpecGen succeeds in generating verifiable specifications for 279 out of 385 programs, outperforming the existing purely LLM-based approaches and conventional specification generation tools like Houdini and Daikon. Further investigations on the quality of generated specifications indicate that SpecGen can comprehensively articulate the behaviors of the input program.

cs.SE↗

arXiv2025

Near-optimal Active Regression of Single-Index Models

The active regression problem of the single-index model is to solve $\min_x \lVert f(Ax)-b\rVert_p$, where $A$ is fully accessible and $b$ can only be accessed via entry queries, with the goal of minimizing the number of queries to the entries of $b$. When $f$ is Lipschitz, previous results only obtain constant-factor approximations. This work presents the first algorithm that provides a $(1+\varepsilon)$-approximation solution by querying $\tilde{O}(d^{\frac{p}{2}\vee 1}/\varepsilon^{p\vee 2})$ entries of $b$. This query complexity is also shown to be optimal up to logarithmic factors for $p\in [1,2]$ and the $\varepsilon$-dependence of $1/\varepsilon^p$ is shown to be optimal for $p>2$.

cs.DS↗

arXiv2025

Continual Test-Time Adaptation for Single Image Defocus Deblurring via Causal Siamese Networks

Single image defocus deblurring (SIDD) aims to restore an all-in-focus image from a defocused one. Distribution shifts in defocused images generally lead to performance degradation of existing methods during out-of-distribution inferences. In this work, we gauge the intrinsic reason behind the performance degradation, which is identified as the heterogeneity of lens-specific point spread functions. Empirical evidence supports this finding, motivating us to employ a continual test-time adaptation (CTTA) paradigm for SIDD. However, traditional CTTA methods, which primarily rely on entropy minimization, cannot sufficiently explore task-dependent information for pixel-level regression tasks like SIDD. To address this issue, we propose a novel Siamese networks-based continual test-time adaptation framework, which adapts source models to continuously changing target domains only requiring unlabeled target data in an online manner. To further mitigate semantically erroneous textures introduced by source SIDD models under severe degradation, we revisit the learning paradigm through a structural causal model and propose Causal Siamese networks (CauSiam). Our method leverages large-scale pre-trained vision-language models to derive discriminative universal semantic priors and integrates these priors into Siamese networks, ensuring causal identifiability between blurry inputs and restored images. Extensive experiments demonstrate that CauSiam effectively improves the generalization performance of existing SIDD methods in continuously changing domains.

eess.IV↗

arXiv2025

AI Models Still Lag Behind Traditional Numerical Models in Predicting Sudden-Turning Typhoons

Given the interpretability, accuracy, and stability of numerical weather prediction (NWP) models, current operational weather forecasting relies heavily on the NWP approach. In the past two years, the rapid development of Artificial Intelligence (AI) has provided an alternative solution for medium-range (1-10 days) weather forecasting. Bi et al. (2023) (hereafter Bi23) introduced the first AI-based weather prediction (AIWP) model in China, named Pangu-Weather, which offers fast prediction without compromising accuracy. In their work, Bi23 made notable claims regarding its effectiveness in extreme weather predictions. However, this claim lacks persuasiveness because the extreme nature of the two tropical cyclones (TCs) examples presented in Bi23, namely Typhoon Kong-rey and Typhoon Yutu, stems primarily from their intensities rather than their moving paths. Their claim may mislead into another meaning which is that Pangu-Weather works well in predicting unusual typhoon paths, which was not explicitly analyzed. Here, we reassess Pangu-Weather's ability to predict extreme TC trajectories from 2020-2024. Results reveal that while Pangu-Weather overall outperforms NWP models in predicting tropical cyclone (TC) tracks, it falls short in accurately predicting the rarely observed sudden-turning tracks, such as Typhoon Khanun in 2023. We argue that current AIWP models still lag behind traditional NWP models in predicting such rare extreme events in medium-range forecasts.

physics.ao-ph↗

arXiv2025

Towards Secure Program Partitioning for Smart Contracts with LLM's In-Context Learning

Smart contracts are highly susceptible to manipulation attacks due to the leakage of sensitive information. Addressing manipulation vulnerabilities is particularly challenging because they stem from inherent data confidentiality issues rather than straightforward implementation bugs. To tackle this by preventing sensitive information leakage, we present PartitionGPT, the first LLM-driven approach that combines static analysis with the in-context learning capabilities of large language models (LLMs) to partition smart contracts into privileged and normal codebases, guided by a few annotated sensitive data variables. We evaluated PartitionGPT on 18 annotated smart contracts containing 99 sensitive functions. The results demonstrate that PartitionGPT successfully generates compilable, and verified partitions for 78% of the sensitive functions while reducing approximately 30% code compared to function-level partitioning approach. Furthermore, we evaluated PartitionGPT on nine real-world manipulation attacks that lead to a total loss of 25 million dollars, PartitionGPT effectively prevents eight cases, highlighting its potential for broad applicability and the necessity for secure program partitioning during smart contract development to diminish manipulation vulnerabilities.

cs.SE↗

arXiv2025

Phantom Events: Demystifying the Issues of Log Forgery in Blockchain

With the rapid development of blockchain technology, transaction logs play a central role in various applications, including decentralized exchanges, wallets, cross-chain bridges, and other third-party services. However, these logs, particularly those based on smart contract events, are highly susceptible to manipulation and forgery, creating substantial security risks across the ecosystem. To address this issue, we present the first in-depth security analysis of transaction log forgery in EVM-based blockchains, a phenomenon we term Phantom Events. We systematically model five types of attacks and propose a tool designed to detect event forgery vulnerabilities in smart contracts. Our evaluation demonstrates that our approach outperforms existing tools in identifying potential phantom events. Furthermore, we have successfully identified real-world instances for all five types of attacks across multiple decentralized applications. Finally, we call on community developers to take proactive steps to address these critical security vulnerabilities.

cs.CR↗

arXiv2025

Faster Approximation Algorithms for k-Center via Data Reduction

We study efficient algorithms for the Euclidean $k$-Center problem, focusing on the regime of large $k$. We take the approach of data reduction by considering $α$-coreset, which is a small subset $S$ of the dataset $P$ such that any $β$-approximation on $S$ is an $(α+ β)$-approximation on $P$. We give efficient algorithms to construct coresets whose size is $k \cdot o(n)$, which immediately speeds up existing approximation algorithms. Notably, we obtain a near-linear time $O(1)$-approximation when $k = n^c$ for any $0 < c < 1$. We validate the performance of our coresets on real-world datasets with large $k$, and we observe that the coreset speeds up the well-known Gonzalez algorithm by up to $4$ times, while still achieving similar clustering cost. Technically, one of our coreset results is based on a new efficient construction of consistent hashing with competitive parameters. This general tool may be of independent interest for algorithm design in high dimensional Euclidean spaces.

cs.DS↗

arXiv2025

GRAPE: Generalizing Robot Policy via Preference Alignment

Despite the recent advancements of vision-language-action (VLA) models on a variety of robotics tasks, they suffer from critical issues such as poor generalizability to unseen tasks, due to their reliance on behavior cloning exclusively from successful rollouts. Furthermore, they are typically fine-tuned to replicate demonstrations collected by experts under different settings, thus introducing distribution bias and limiting their adaptability to diverse manipulation objectives, such as efficiency, safety, and task completion. To bridge this gap, we introduce GRAPE: Generalizing Robot Policy via Preference Alignment. Specifically, GRAPE aligns VLAs on a trajectory level and implicitly models reward from both successful and failure trials to boost generalizability to diverse tasks. Moreover, GRAPE breaks down complex manipulation tasks to independent stages and automatically guides preference modeling through customized spatiotemporal constraints with keypoints proposed by a large vision-language model. Notably, these constraints are flexible and can be customized to align the model with varying objectives, such as safety, efficiency, or task success. We evaluate GRAPE across a diverse array of tasks in both real-world and simulated environments. Experimental results demonstrate that GRAPE enhances the performance of state-of-the-art VLA models, increasing success rates on in-domain and unseen manipulation tasks by 51.79% and 58.20%, respectively. Additionally, GRAPE can be aligned with various objectives, such as safety and efficiency, reducing collision rates by 37.44% and rollout step-length by 11.15%, respectively. All code, models, and data are available at https://grape-vla.github.io/

cs.RO↗

arXiv2025

Consensus statement on the credibility assessment of ML predictors

The rapid integration of machine learning (ML) predictors into in silico medicine has revolutionized the estimation of quantities of interest (QIs) that are otherwise challenging to measure directly. However, the credibility of these predictors is critical, especially when they inform high-stakes healthcare decisions. This position paper presents a consensus statement developed by experts within the In Silico World Community of Practice. We outline twelve key statements forming the theoretical foundation for evaluating the credibility of ML predictors, emphasizing the necessity of causal knowledge, rigorous error quantification, and robustness to biases. By comparing ML predictors with biophysical models, we highlight unique challenges associated with implicit causal knowledge and propose strategies to ensure reliability and applicability. Our recommendations aim to guide researchers, developers, and regulators in the rigorous assessment and deployment of ML predictors in clinical and biomedical contexts.

q-bio.QM↗

arXiv2025

Fine-Grained 1-Day Vulnerability Detection in Binaries via Patch Code Localization

1-day vulnerabilities in binaries have become a major threat to software security. Patch presence test is one of the effective ways to detect the vulnerability. However, existing patch presence test works do not perform well in practical scenarios due to the interference from the various compilers and optimizations, patch-similar code blocks, and irrelevant functions in stripped binaries. In this paper, we propose a novel approach named PLocator, which leverages stable values from both the patch code and its context, extracted from the control flow graph, to accurately locate the real patch code in the target function, offering a practical solution for real-world vulnerability detection scenarios. To evaluate the effectiveness of PLocator, we collected 73 CVEs and constructed two comprehensive datasets ($Dataset_{-irr}$ and $Dataset_{+irr}$), comprising 1,090 and 27,250 test cases at four compilation optimization levels and two compilers with three different experiments, i.e., Same, XO (cross-optimizations), and XC (cross-compilers). The results demonstrate that PLocator achieves an average TPR of 88.2% and FPR of 12.9% in a short amount of time, outperforming state-of-the-art approaches by 26.7% and 63.5%, respectively, indicating that PLocator is more practical for the 1-day vulnerability detection task.

cs.CR↗

arXiv2025

The spectral dynamics and spatial structures of the conditional Lyapunov vector in slave Kolmogorov flow

We conduct direct numerical simulations to investigate the synchronization of Kolmogorov flows in a periodic box, with a focus on the mechanisms underlying the asymptotic evolution of infinitesimal velocity perturbations, also known as conditional leading Lyapunov vectors. This study advances previous work with a spectral analysis of the perturbation, which clarifies the behaviours of the production and dissipation spectra at different coupling wavenumbers. We show that, in simulations with moderate Reynolds numbers, the conditional leading Lyapunov exponent can be smaller than a lower bound proposed previously based on a viscous estimate. A quantitative analysis of the self-similar evolution of the perturbation energy spectrum is presented, extending the existing qualitative discussion. The prerequisites for obtaining self-similar solutions are established, which include an interesting relationship between the integral length scale of the perturbation velocity and the local Lyapunov exponent. By examining the governing equation for the dissipation rate of the velocity perturbation, we reveal the previously neglected roles of the strain rate and vorticity perturbations and uncover their unique geometrical characteristics.

physics.flu-dyn↗