arXiv subjects

Jian Liu

Publications and source records attributed to Jian Liu.

At least 181 records · Page 10Linked to original sources

arXiv2025

Sparse Autoencoder as a Zero-Shot Classifier for Concept Erasing in Text-to-Image Diffusion Models

Text-to-image (T2I) diffusion models have achieved remarkable progress in generating high-quality images but also raise people's concerns about generating harmful or misleading content. While extensive approaches have been proposed to erase unwanted concepts without requiring retraining from scratch, they inadvertently degrade performance on normal generation tasks. In this work, we propose Interpret then Deactivate (ItD), a novel framework to enable precise concept removal in T2I diffusion models while preserving overall performance. ItD first employs a sparse autoencoder (SAE) to interpret each concept as a combination of multiple features. By permanently deactivating the specific features associated with target concepts, we repurpose SAE as a zero-shot classifier that identifies whether the input prompt includes target concepts, allowing selective concept erasure in diffusion models. Moreover, we demonstrate that ItD can be easily extended to erase multiple concepts without requiring further training. Comprehensive experiments across celebrity identities, artistic styles, and explicit content demonstrate ItD's effectiveness in eliminating targeted concepts without interfering with normal concept generation. Additionally, ItD is also robust against adversarial prompts designed to circumvent content filters. Code is available at: https://github.com/NANSirun/Interpret-then-deactivate.

cs.CV↗

arXiv2025

Towards Collaborative Anti-Money Laundering Among Financial Institutions

Money laundering is the process that intends to legalize the income derived from illicit activities, thus facilitating their entry into the monetary flow of the economy without jeopardizing their source. It is crucial to identify such activities accurately and reliably in order to enforce anti-money laundering (AML). Despite considerable efforts to AML, a large number of such activities still go undetected. Rule-based methods were first introduced and are still widely used in current detection systems. With the rise of machine learning, graph-based learning methods have gained prominence in detecting illicit accounts through the analysis of money transfer graphs. Nevertheless, these methods generally assume that the transaction graph is centralized, whereas in practice, money laundering activities usually span multiple financial institutions. Due to regulatory, legal, commercial, and customer privacy concerns, institutions tend not to share data, restricting their utility in practical usage. In this paper, we propose the first algorithm that supports performing AML over multiple institutions while protecting the security and privacy of local data. To evaluate, we construct Alipay-ECB, a real-world dataset comprising digital transactions from Alipay, the world's largest mobile payment platform, alongside transactions from E-Commerce Bank (ECB). The dataset includes over 200 million accounts and 300 million transactions, covering both intra-institution transactions and those between Alipay and ECB. This makes it the largest real-world transaction graph available for analysis. The experimental results demonstrate that our methods can effectively identify cross-institution money laundering subgroups. Additionally, experiments on synthetic datasets also demonstrate that our method is efficient, requiring only a few minutes on datasets with millions of transactions.

cs.SI↗

arXiv2025

Topological Sequence Analysis of Genomes: Delta Complex approaches

Algebraic topology has been widely applied to point cloud data to capture geometric shapes and topological structures. However, its application to genome sequence analysis remains rare. In this work, we propose topological sequence analysis (TSA) techniques by constructing $Δ$-complexes and classifying spaces, leading to persistent homology, and persistent path homology on genome sequences. We also develop $Δ$-complex-based persistent Laplacians to facilitate the topological spectral analysis of genome sequences. Finally, we demonstrate the utility of the proposed TSA approaches in phylogenetic analysis using Ebola virus sequences and whole bacterial genomes. The present TSA methods are more efficient than earlier TSA model, k-mer topology, and thus have a potential to be applied to other time-consuming sequential data analyses, such as those in linguistics, literature, music, media, and social contexts.

math.AT↗

arXiv2025

Reducing the Cost of Dropout in Flash-Attention by Hiding RNG with GEMM

Dropout, a network operator, when enabled is likely to dramatically impact the performance of Flash-Attention, which in turn increases the end-to-end training time of Large-Language-Models (LLMs). The main contributor to such performance degradation is the Random Number Generation (RNG) phase. The state-of-the-art optimization is to fuse RNG into the Flash-Attention kernel. However, while RNG and Attention do not compete on compute or memory resources, they are bounded on the same lower-level architecture bottlenecks. Fusion can hardly hide RNG latency within the Attention kernel. We propose overlapping RNG with previous GEMM layers in the network to hide RNG latency and improve end-to-end performance. RNG and GEMM have distinct resource requirements and hardware bottlenecks, so they can run together without compromising each other's performance. We propose a fine-grained analytical performance model that analyzes low-level architecture resource utilization to evaluate RNG-GEMM overlapping performance benefits. This model, cross-validated by silicon results, shows 1.26x speedup for overlapping RNG and GEMM layers over a sequential implementation on one Transformer Block (one LLM layer including multi-head attention and feed-forward layers), and 1.22x over state-of-the-art fusion implementation, for Llama3 on GH100 GPUs with FP8 precision. Because the kernel patterns are regular, the findings of the shared bottlenecks, as well as the achievable performance benefits, can be generalized to different model architectures, software implementations and hardware configurations.

cs.AR↗

arXiv2025

Ascent and descent of Gorenstein homological properties

Let $φ\colon R\rightarrow A$ be a ring homomorphism, where $R$ is a commutative noetherian ring and $A$ is a finite $R$-algebra. We provide criteria for detecting the ascent and descent of Gorenstein homological properties. %As an application, one can deduce a result that supports a question of Avramov and Foxby. We observe that the ascent and descent of Gorenstein homological property can detect the Gorenstein properties of rings along $φ$. Finally, we describe when $φ$ induces a triangle equivalence between the stable categories of finitely generated Gorenstein projective modules over $R$ and $A$.

math.AC↗

arXiv2025

Charged-current quasielastic neutrino scattering off nuclei with nucleon-nucleon short-range correlations

In recent years, many studies on neutrino-nucleus scattering have been carried out to investigate nuclear structures and the interactions between neutrinos and nucleons. This paper develops a charged-current quasielastic (CCQE) neutrino-nucleus scattering model to explore the nuclear mean-field dynamics and short-range correlation effects. In this model, the nuclear structure effect is depicted using the scaling function, while the neutrinonucleon interaction is represented by the elementary weak cross section. Results indicate that the double-differential cross section of scattered muon is influenced by the energy and momentum of nucleon in nuclei, and the total cross section depends primarily on the incident neutrino energy. Furthermore, incorporating short-range correlations yields the flux-integrated differential cross sections at high-T region producing larger values, a longer tail, and achieving better experimental consistency. It eventually elucidates the physical relationship between the neutrinonucleus scattering cross section and the variation in incident neutrino energy. The studies in this paper furnishes insights for the research of nucleon dynamics and provides detailed examinations of the neutrino-nucleus scattering mechanism.

nucl-th↗

arXiv2025

Activation Approximations Can Incur Safety Vulnerabilities Even in Aligned LLMs: Comprehensive Analysis and Defense

Large Language Models (LLMs) have showcased remarkable capabilities across various domains. Accompanying the evolving capabilities and expanding deployment scenarios of LLMs, their deployment challenges escalate due to their sheer scale and the advanced yet complex activation designs prevalent in notable model series, such as Llama, Gemma, Mistral. These challenges have become particularly pronounced in resource-constrained deployment scenarios, where mitigating inference bottlenecks is imperative. Among various recent efforts, activation approximation has emerged as a promising avenue for pursuing inference efficiency, sometimes considered indispensable in applications such as private inference. Despite achieving substantial speedups with minimal impact on utility, even appearing sound and practical for real-world deployment, the safety implications of activation approximations remain unclear. In this work, we fill this critical gap in LLM safety by conducting the first systematic safety evaluation of activation approximations. Our safety vetting spans seven state-of-the-art techniques across three popular categories (activation polynomialization, activation sparsification, and activation quantization), revealing consistent safety degradation across ten safety-aligned LLMs. To overcome the hurdle of devising a unified defense accounting for diverse activation approximation methods, we perform an in-depth analysis of their shared error patterns and uncover three key findings. We propose QuadA, a novel safety enhancement method tailored to mitigate the safety compromises introduced by activation approximations. Extensive experiments and ablation studies corroborate QuadA's effectiveness in enhancing the safety capabilities of LLMs after activation approximations.

cs.CR↗

arXiv2025

Designing Pin-pression Gripper and Learning its Dexterous Grasping with Online In-hand Adjustment

We introduce a novel design of parallel-jaw grippers drawing inspiration from pin-pression toys. The proposed pin-pression gripper features a distinctive mechanism in which each finger integrates a 2D array of pins capable of independent extension and retraction. This unique design allows the gripper to instantaneously customize its finger's shape to conform to the object being grasped by dynamically adjusting the extension/retraction of the pins. In addition, the gripper excels in in-hand re-orientation of objects for enhanced grasping stability again via dynamically adjusting the pins. To learn the dynamic grasping skills of pin-pression grippers, we devise a dedicated reinforcement learning algorithm with careful designs of state representation and reward shaping. To achieve a more efficient grasp-while-lift grasping mode, we propose a curriculum learning scheme. Extensive evaluations demonstrate that our design, together with the learned skills, leads to highly flexible and robust grasping with much stronger generality to unseen objects than alternatives. We also highlight encouraging physical results of sim-to-real transfer on a physically manufactured pin-pression gripper, demonstrating the practical significance of our novel gripper design and grasping skill. Demonstration videos for this paper are available at https://github.com/siggraph-pin-pression-gripper/pin-pression-gripper-video.

cs.RO↗

arXiv2025

Openness with respect to levels in triangulated categories

Given a compactly generated triangulated category $\mathcal{T}$ equipped with an action of a graded-commutative Noetherian ring $R$, generalizing results of Letz, we prove a general result concerning the openness with respect to levels of compact objects in $\mathcal{T}$. Applications are given to derived categories of commutative Noetherian rings, derived categories of commutative Noetherian DG rings and singularity categories.

math.AC↗

arXiv2025

Seismic Isolation of Optical Tables Using Piezo Actuators

Seismic isolation is crucial for gravitational wave detectors as it minimizes ground vibrations, enabling the detection of faint gravitational wave signals. An active seismic isolation platform for precision measurement experiments is described. The table features piezo actuation along five degrees of freedom: three translational actuations and two tip-tilt degrees of freedom along the horizontal axes. It is stiff in rotation about the vertical axes. A seismometer is used to sense table motion. Piezo actuators are used to suppress seismic noise with feedback control bandwidth of 0.3 to 3 Hz. Suppression levels ranging from 21 to 36 dB of seismic noise within the frequency range of 0.5 to 1.3 Hz are demonstrated, as measured by a witness seismometer on the table, with the suppression direction along the axis of the longitudinal translation of the suspended mirror on the table. The suppression results in 1 $\mathrm{\mathrm{nm/\sqrt{Hz}}}$ residual horizontal motion at 1 Hz. Limitations such as tilt-to-translation coupling that prevent actuation over the desired range of 0.03 to 3 Hz are discussed.

astro-ph.IM↗

arXiv2025

FreeMesh: Boosting Mesh Generation with Coordinates Merging

The next-coordinate prediction paradigm has emerged as the de facto standard in current auto-regressive mesh generation methods. Despite their effectiveness, there is no efficient measurement for the various tokenizers that serialize meshes into sequences. In this paper, we introduce a new metric Per-Token-Mesh-Entropy (PTME) to evaluate the existing mesh tokenizers theoretically without any training. Building upon PTME, we propose a plug-and-play tokenization technique called coordinate merging. It further improves the compression ratios of existing tokenizers by rearranging and merging the most frequent patterns of coordinates. Through experiments on various tokenization methods like MeshXL, MeshAnything V2, and Edgerunner, we further validate the performance of our method. We hope that the proposed PTME and coordinate merging can enhance the existing mesh tokenizers and guide the further development of native mesh generation.

cs.GR↗

arXiv2025

PSBench: a large-scale benchmark for estimating the accuracy of protein complex structural models

Predicting protein complex structures is essential for protein function analysis, protein design, and drug discovery. While AI methods like AlphaFold can predict accurate structural models for many protein complexes, reliably estimating the quality of these predicted models (estimation of model accuracy, or EMA) for model ranking and selection remains a major challenge. A key barrier to developing effective machine learning-based EMA methods is the lack of large, diverse, and well-annotated datasets for training and evaluation. To address this gap, we introduce PSBench, a benchmark suite comprising four large-scale, labeled datasets generated during the 15th and 16th community-wide Critical Assessment of Protein Structure Prediction (CASP15 and CASP16). PSBench includes over one million structural models covering a wide range of protein sequence lengths, complex stoichiometries, functional classes, and modeling difficulties. Each model is annotated with multiple complementary quality scores at the global, local, and interface levels. PSBench also provides multiple evaluation metrics and baseline EMA methods to facilitate rigorous comparisons. To demonstrate PSBench's utility, we trained and evaluated GATE, a graph transformer-based EMA method, on the CASP15 data. GATE was blindly tested in CASP16 (2024), where it ranked among the top-performing EMA methods. These results highlight PSBench as a valuable resource for advancing EMA research in protein complex modeling. PSBench is publicly available at: https://github.com/BioinfoMachineLearning/PSBench.

q-bio.BM↗

arXiv2025

ABase: the Multi-Tenant NoSQL Serverless Database for Diverse and Dynamic Workloads in Large-scale Cloud Environments

Multi-tenant architectures enhance the elasticity and resource utilization of NoSQL databases by allowing multiple tenants to co-locate and share resources. However, in large-scale cloud environments, the diverse and dynamic nature of workloads poses significant challenges for multi-tenant NoSQL databases. Based on our practical observations, we have identified three crucial challenges: (1) the impact of caching on performance isolation, as cache hits alter request execution and resource consumption, leading to inaccurate traffic control; (2) the dynamic changes in traffic, with changes in tenant traffic trends causing throttling or resource wastage, and changes in access distribution causing hot key pressure or cache hit ratio drops; and (3) the imbalanced layout of data nodes due to tenants' diverse resource requirements, leading to low resource utilization. To address these challenges, we introduce ABase, a multi-tenant NoSQL serverless database developed at ByteDance. ABase introduces a two-layer caching mechanism with a cache-aware isolation mechanism to ensure accurate resource consumption estimates. Furthermore, ABase employs a predictive autoscaling policy to dynamically adjust resources in response to tenant traffic changes and a multi-resource rescheduling algorithm to balance resource utilization across data nodes. With these innovations, ABase has successfully served ByteDance's large-scale cloud environment, supporting a total workload that has achieved a peak QPS of over 13 billion and total storage exceeding 1 EB.

cs.DB↗

arXiv2025

Boosting Single-domain Generalized Object Detection via Vision-Language Knowledge Interaction

Single-Domain Generalized Object Detection~(S-DGOD) aims to train an object detector on a single source domain while generalizing well to diverse unseen target domains, making it suitable for multimedia applications that involve various domain shifts, such as intelligent video surveillance and VR/AR technologies. With the success of large-scale Vision-Language Models, recent S-DGOD approaches exploit pre-trained vision-language knowledge to guide invariant feature learning across visual domains. However, the utilized knowledge remains at a coarse-grained level~(e.g., the textual description of adverse weather paired with the image) and serves as an implicit regularization for guidance, struggling to learn accurate region- and object-level features in varying domains. In this work, we propose a new cross-modal feature learning method, which can capture generalized and discriminative regional features for S-DGOD tasks. The core of our method is the mechanism of Cross-modal and Region-aware Feature Interaction, which simultaneously learns both inter-modal and intra-modal regional invariance through dynamic interactions between fine-grained textual and visual features. Moreover, we design a simple but effective strategy called Cross-domain Proposal Refining and Mixing, which aligns the position of region proposals across multiple domains and diversifies them, enhancing the localization ability of detectors in unseen scenarios. Our method achieves new state-of-the-art results on S-DGOD benchmark datasets, with improvements of +8.8\%~mPC on Cityscapes-C and +7.9\%~mPC on DWD over baselines, demonstrating its efficacy.

cs.CV↗

arXiv2025

LookAhead: Preventing DeFi Attacks via Unveiling Adversarial Contracts

The exploitation of smart contract vulnerabilities in Decentralized Finance (DeFi) has resulted in financial losses exceeding 3 billion US dollars. Existing defense mechanisms primarily focus on detecting and reacting to adversarial transactions executed by attackers that target victim contracts. However, with the emergence of private transaction pools where transactions are sent directly to miners without first appearing in public mempools, current detection tools face significant challenges in identifying attack activities effectively. Based on the fact that most attack logic rely on deploying intermediate smart contracts as supporting components to the exploitation of victim contracts, novel detection methods have been proposed that focus on identifying these adversarial contracts instead of adversarial transactions. However, previous state-of-the-art approaches in this direction have failed to produce results satisfactory enough for real-world deployment. In this paper, we propose LookAhead, a new framework for detecting DeFi attacks via unveiling adversarial contracts. LookAhead leverages common attack patterns, code semantics and intrinsic characteristics found in adversarial smart contracts to train Machine Learning (ML)-based classifiers that can effectively distinguish adversarial contracts from benign ones and make timely predictions of different types of potential attacks. Experiments on our labeled datasets show that LookAhead achieves an F1-score as high as 0.8966, which represents an improvement of over 44.4% compared to the previous state-of-the-art solution, with a False Positive Rate (FPR) at only 0.16%.

cs.CR↗

arXiv2025

MonoDiff9D: Monocular Category-Level 9D Object Pose Estimation via Diffusion Model

Object pose estimation is a core means for robots to understand and interact with their environment. For this task, monocular category-level methods are attractive as they require only a single RGB camera. However, current methods rely on shape priors or CAD models of the intra-class known objects. We propose a diffusion-based monocular category-level 9D object pose generation method, MonoDiff9D. Our motivation is to leverage the probabilistic nature of diffusion models to alleviate the need for shape priors, CAD models, or depth sensors for intra-class unknown object pose estimation. We first estimate coarse depth via DINOv2 from the monocular image in a zero-shot manner and convert it into a point cloud. We then fuse the global features of the point cloud with the input image and use the fused features along with the encoded time step to condition MonoDiff9D. Finally, we design a transformer-based denoiser to recover the object pose from Gaussian noise. Extensive experiments on two popular benchmark datasets show that MonoDiff9D achieves state-of-the-art monocular category-level 9D object pose estimation accuracy without the need for shape priors or CAD models at any stage. Our code will be made public at https://github.com/CNJianLiu/MonoDiff9D.

cs.CV↗

arXiv2025

An Efficient Integrator Scheme for Sampling the (Quantum) Isobaric-Isothermal Ensemble in (Path Integral) Molecular Dynamics Simulations

Because most chemical or biological experiments are performed under conditions of controlled pressure and temperature, it is important to simulate the isobaric-isothermal ensemble at the atomic level to reveal the microscopic mechanism. By extending our configuration sampling protocol for the canonical ensemble, we propose a unified middle scheme to sample the coordinate (configuration) and volume distribution and thereby are able to accurately simulate either classical or quantum isobaric-isothermal processes. Various barostats and thermostats can be employed in the unified middle scheme for simulating real molecular systems with or without holonomic constraints. In particular, we demonstrate the recommended middle scheme by employing the Martyna-Tuckerman-Tobias-Klein barostat and stochastic cell-rescaling barostat, with the Langevin thermostat, in molecular simulation packages (DL_POLY, Amber, Gromacs, etc.). Benchmark numerical tests show that, without additional numerical effort, the middle scheme is competent in increasing the time interval by a factor of 5~10 to achieve the same accuracy of converged results for most thermodynamic properties in (path integral) molecular dynamics simulations.

cond-mat.stat-mech↗

arXiv2025

Nonadiabatic Field: A Conceptually Novel Approach for Nonadiabatic Quantum Molecular Dynamics

Reliable trajectory-based nonadiabatic quantum dynamics methods at the atomic level are critical for understanding many important processes in real systems. The paper reports latest progress of nonadiabatic field (NaF), a conceptually novel approach for nonadiabatic quantum dynamics with independent trajectories. Substantially different from the mainstreams of Ehrenfest-like dynamics and surface hopping methods, the nuclear force in NaF involves the nonadiabatic force arising from the nonadiabatic coupling between different electronic states, in addition to the adiabatic force contributed by a single adiabatic electronic state. NaF is capable of faithfully describing the interplay between electronic and nuclear motion in a broad regime, which covers where the relevant electronic states keep coupled in a wide range or all the time and where the bifurcation characteristic of nuclear motion is essential. NaF is derived from the exact generalized phase space formulation with coordinate-momentum variables, where constraint phase space (CPS) is employed for discrete electronic-state degrees of freedom. We propose efficient integrators for the equations of motion of NaF in both adiabatic and diabatic representations. Since the formalism in the CPS formulation is not unique, NaF can in principle be implemented with various phase space representations of the time correlation function (TCF) for the time-dependent property. They are applied to a suite of representative gas-phase and condensed-phase benchmark models where numerically exact results are available for comparison. It is shown that NaF is relatively insensitive to the phase space representation of the electronic TCF and will be a potential tool for practical and reliable simulations of the quantum mechanical behavior of both electronic and nuclear dynamics of nonadiabatic transition processes in real systems.

physics.chem-ph↗