arXiv subjects

Long Liu

Publications and source records attributed to Long Liu.

At least 19 recordsLinked to original sources

arXiv2026

STCC: A Unified Source-Channel Semantic Token Coding Framework for Semantic Communications

Deep Joint Source-Channel Coding (JSCC) has emerged as a promising paradigm for overcoming the ``cliff effect" in wireless communications. However, existing Deep JSCC frameworks operate directly on raw analog data such as image pixels rather than the discrete semantic tokens that foundation models require. Moreover, traditional systems employ fixed, hand-designed constellations that treat all tokens equally, leading to catastrophic random errors under channel noise. In this paper, the Semantic Token Codebook Communication (STCC) is proposed as a unified source-channel semantic token coding framework designed to transmit the discrete semantic tokens of foundation models over noisy channels. The core of STCC is the Semantic Token Codec (STC). It accepts discrete tokens as input, which maintains compatibility with foundation models while employing a residual multiple layer perceptron, i.e., MLP-based encoder that learns geometrically structured constellations optimized with a triple-loss objective. This learned mapping forces the channel topology to align with the semantic embedding space, ensuring that channel noise results in topological errors rather than random corruption. This phenomenon is theoretically and empirically characterized, identifying ``Semantic Drift" in symbolic modalities and ``Structural Distortion" in perceptual modalities, where errors shift predictions to semantically or structurally similar tokens. Extensive experiments demonstrate that STCC significantly outperforms traditional systems in low-SNR regimes, effectively converting channel noise into semantic variations without requiring receiver-side modification.

cs.IT↗

arXiv2026

Is My RPC Response Reliable? Detecting RPC Bugs in Ethereum Blockchain Client under Context

Blockchain clients are fundamental software for running blockchain nodes. They provide users with various RPC (Remote Procedure Call) interfaces to interact with the blockchain. These RPC methods are expected to follow the same specification across different blockchain nodes, providing users with seamless interaction. However, there have been continuous reports on various RPC bugs that can cause unexpected responses or even Denial of Service weakness. Existing studies on blockchain RPC bug detection mainly focus on generating the RPC method calls for testing blockchain clients. However, a wide range of the reported RPC bugs are triggered in various blockchain contexts. To the best of our knowledge, little attention is paid to generating proper contexts that can trigger these context-dependent RPC bugs. In this work, we propose EthCRAFT, a Context-aware RPC Analysis and Fuzzing Tool for client RPC bug detection. EthCRAFT first proposes to explore the state transition program space of blockchain clients and generate various transactions to construct the context. EthCRAFT then designs a context-aware RPC method call generation method to send RPC calls to the blockchain clients. The responses of 5 different client implementations are used as cross-referring oracles to detect the RPC bugs. We evaluate EthCRAFT on real-world RPC bugs collected from the GitHub issues of Ethereum client implementations. Experiment results show that EthCRAFT outperforms existing client RPC detectors by detecting more RPC bugs. Moreover, EthCRAFT has found six new bugs in major Ethereum clients and reported them to the developers. One of the bug fixes has been written into breaking changes in the client's updates. Three of our bug reports have been offered a vulnerability bounty by the Ethereum Foundation.

cs.SE↗

arXiv2025

Room temperature giant magnetoresistance detection of spin hall nano-oscillator dynamics in synthetic antiferromagnetic Spin-Valve

Conventional spin Hall nano-oscillators (SHNOs) face fundamental power limitations due to the low anisotropic magnetoresistance (AMR < 0.3%) of ferromagnetic layers. To address this, we developed a synthetic antiferromagnetic spin-valve (SAF-SV) heterostructure [Ta/NiFe/Ru/NiFe/Cu/NiFe/Hf/Pt] that enables efficient giant magnetoresistance (GMR)-based detection of SHNO dynamics at room temperature. The NiFe/Ru/NiFe SAF reference layer, operating in the spin-flop state, couples with the NiFe free layer through a Cu spacer to achieve a remarkable GMR ratio of 0.568% - exhibiting complete independence of magnetic field/current orientation. Spin-torque ferromagnetic resonance (ST-FMR) verifies that the ferromagnetic resonance linewidth of the free layer can be effectively modulated by dc current through the Pt heavy metal layer, while maintaining decoupled dynamics from the SAF layer. Thermal management via high-thermal-conductivity SiC substrates and AlN capping layers successfully mitigates current-shunting-induced Joule heating. Notably, stable auto-oscillation peaks are observed at 0.82 mA bias current, with oscillation frequency tunable by external magnetic field and potential dual-mode behavior at low fields. This work establishes a new paradigm for room-temperature, high-power spintronic oscillators, offering significant potential for neuromorphic computing and coherent RF communication applications.

cond-mat.mes-hall↗

arXiv2025

ORFuzz: Fuzzing the "Other Side" of LLM Safety -- Testing Over-Refusal

Large Language Models (LLMs) increasingly exhibit over-refusal - erroneously rejecting benign queries due to overly conservative safety measures - a critical functional flaw that undermines their reliability and usability. Current methods for testing this behavior are demonstrably inadequate, suffering from flawed benchmarks and limited test generation capabilities, as highlighted by our empirical user study. To the best of our knowledge, this paper introduces the first evolutionary testing framework, ORFuzz, for the systematic detection and analysis of LLM over-refusals. ORFuzz uniquely integrates three core components: (1) safety category-aware seed selection for comprehensive test coverage, (2) adaptive mutator optimization using reasoning LLMs to generate effective test cases, and (3) OR-Judge, a human-aligned judge model validated to accurately reflect user perception of toxicity and refusal. Our extensive evaluations demonstrate that ORFuzz generates diverse, validated over-refusal instances at a rate (6.98% average) more than double that of leading baselines, effectively uncovering vulnerabilities. Furthermore, ORFuzz's outputs form the basis of ORFuzzSet, a new benchmark of 1,855 highly transferable test cases that achieves a superior 63.56% average over-refusal rate across 10 diverse LLMs, significantly outperforming existing datasets. ORFuzz and ORFuzzSet provide a robust automated testing framework and a valuable community resource, paving the way for developing more reliable and trustworthy LLM-based software systems.

cs.SE↗

arXiv2025

BlueLM-2.5-3B Technical Report

We present BlueLM-2.5-3B, a compact and unified dense Multimodal Large Language Model (MLLM) designed for efficient edge-device deployment, offering strong general-purpose and reasoning capabilities. To the best of our knowledge, this is the first 3B-scale MLLM to support both thinking and non-thinking modes, while also enabling explicit control over thinking token budget. BlueLM-2.5-3B is developed through diversified data curation, key data resampling, hybrid heterogeneous reinforcement learning, and a high-performance training infrastructure. Our model achieves superior multimodal capacity while preserving competitive pure-text performance with only 2.9 billion parameters. We conduct comprehensive evaluations across a broad range of multimodal and text-only benchmarks. In thinking mode, BlueLM-2.5-3B achieves comparable performance to Qwen3-4B on text-only benchmarks, and trails the larger Kimi-VL-A3B-16B by only about 5% on average across multimodal evaluations. In non-thinking mode, it outperforms Qwen2.5-VL-3B on the majority of multimodal benchmarks. Additionally, BlueLM-2.5-3B exhibits exceptional data efficiency. All of the aforementioned performance is achieved with substantially less total training data than Qwen2.5-VL-3B and Qwen3-4B. We hope our work contributes to the advancement of high-performance, on-device MLLMs and provides meaningful insights to the research community.

cs.AI↗

arXiv2025

Dual-Forward Path Teacher Knowledge Distillation: Bridging the Capacity Gap Between Teacher and Student

Knowledge distillation (KD) provides an effective way to improve the performance of a student network under the guidance of pre-trained teachers. However, this approach usually brings in a large capacity gap between teacher and student networks, limiting the distillation gains. Previous methods addressing this problem either discard accurate knowledge representation or fail to dynamically adjust the transferred knowledge, which is less effective in addressing the capacity gap problem and hinders students from achieving comparable performance with the pre-trained teacher. In this work, we extend the ideology of prompt-based learning to address the capacity gap problem, and propose Dual-Forward Path Teacher Knowledge Distillation (DFPT-KD), which replaces the pre-trained teacher with a novel dual-forward path teacher to supervise the learning of student. The key to DFPT-KD is prompt-based tuning, i.e., establishing an additional prompt-based forward path within the pre-trained teacher and optimizing it with the pre-trained teacher frozen to make the transferred knowledge compatible with the representation ability of the student. Extensive experiments demonstrate that DFPT-KD leads to trained students performing better than the vanilla KD. To make the transferred knowledge better compatible with the representation abilities of the student, we further fine-tune the whole prompt-based forward path, yielding a novel distillation approach dubbed DFPT-KD+. By extensive experiments, it is shown that DFPT-KD+ improves upon DFPT-KD and achieves state-of-the-art accuracy performance.

cs.LG↗

arXiv2025

$d$-Wave Flat Fermi Surface in Altermagnets Enables Maximum Charge-to-Spin Conversion

Altermagnets combine antiferromagnetic order with ferromagnet-like spin splitting, a duality that unlocks ultrafast spin-dependent responses. This unique property creates unprecedented opportunities for spin-current generation, overcoming the intrinsic limitations of conventional spin-transfer and spin-orbit torque approaches in magnetic memory technologies. Here, we establish a fundamental relationship between Fermi surface geometry and time-reversal-odd ($\mathcal{T}$-odd) spin currents in altermagnets through combined model analysis and first-principles calculations. We demonstrate that a $d$-wave altermagnet with a flat Fermi surface can achieve a theoretical upper limit of charge-to-spin conversion efficiency (CSE) of 100%. This mechanism is realized in the newly discovered room-temperature altermagnetic metal KV$_2$O$_2$Se, which exhibits a CSE of $\sim$78% at the charge neutrality point, nearly double that of RuO$_2$, setting a new record for $\mathcal{T}$-odd CSE. Under electron doping, this efficiency further increases to $\sim$98%, approaching the theoretical limit. Our work advances the fundamental understanding of $\mathcal{T}$-odd spin currents via Fermi surface geometry engineering and provides key insights for developing next-generation altermagnet-based memory devices.

cond-mat.mtrl-sci↗

arXiv2025

OG-HFYOLO :Orientation gradient guidance and heterogeneous feature fusion for deformation table cell instance segmentation

Table structure recognition is a key task in document analysis. However, the geometric deformation in deformed tables causes a weak correlation between content information and structure, resulting in downstream tasks not being able to obtain accurate content information. To obtain fine-grained spatial coordinates of cells, we propose the OG-HFYOLO model, which enhances the edge response by Gradient Orientation-aware Extractor, combines a Heterogeneous Kernel Cross Fusion module and a scale-aware loss function to adapt to multi-scale objective features, and introduces mask-driven non-maximal suppression in the post-processing, which replaces the traditional bounding box suppression mechanism. Furthermore, we also propose a data generator, filling the gap in the dataset for fine-grained deformation table cell spatial coordinate localization, and derive a large-scale dataset named Deformation Wired Table (DWTAL). Experiments show that our proposed model demonstrates excellent segmentation accuracy on all mainstream instance segmentation models. The dataset and the source code are open source: https://github.com/justliulong/OGHFYOLO.

cs.CV↗

arXiv2025

ASMA-Tune: Unlocking LLMs' Assembly Code Comprehension via Structural-Semantic Instruction Tuning

Assembly code analysis and comprehension play critical roles in applications like reverse engineering, yet they face substantial challenges due to low information density and a lack of explicit syntactic structures. While traditional masked language modeling (MLM) approaches do not explicitly focus on natural language interaction, emerging decoder-focused large language models (LLMs) demonstrate partial success in binary analysis yet remain underexplored for holistic comprehension. We present Assembly Augmented Tuning, an end-to-end structural-semantic instruction tuning framework that synergizes encoder architecture with decoder-based LLMs through a projector module, where the assembly encoder extracts hardware-level structural features, the projector bridges representations with the semantic space, and the instruction-tuned LLM preserves natural language capabilities. Experimental results demonstrate three key advantages: (1) State-of-the-art performance in assembly comprehension with +39.7% Recall@1 and +17.8% MRR improvements over GPT-4-Turbo, (2) Consistent enhancements across base models (24.6-107.4% Recall@1 and 15.2-106.3% MRR on Qwen2.5-Coder, Deepseek-Coder and CodeLlama variants), and (3) Superior instruction-following capabilities (41.5%-118% improvements) with controlled code generation degradation (-8.9% to -35% across architectures).

cs.SE↗

arXiv2025

Gate-Tunable Spin-to-Charge Conversion in Topological Insulator-Magnetic Insulator Heterostructures at Room Temperature

Over the past decade, topological insulators have received enormous attention for their potential in energy-efficient spin-to-charge conversion, enabled by strong spin-orbit coupling and spin-momentum locked surface states. Despite extensive research, the spin-to-charge conversion efficiency, usually characterized by the spin Hall angle ({\theta}SH), remains low at room temperature. In this work, we employed pulsed laser deposition to synthesize high-quality ternary topological insulators (Bi0.1Sb0.9)2Te3 thin films on magnetic insulator Y3Fe5O12. We find that the value of {\theta}SH reaches ~0.76 at room temperature and increases to ~0.9 as the Fermi level is tuned to cross topological surface states via electrical gating. Our findings provide an innovative approach to tailoring the spin-to-charge conversion in topological insulators and pave the way for their applications in energy-efficient spintronic devices.

cond-mat.mes-hall↗

arXiv2024

Physics-Guided Detector for SAR Airplanes

The disperse structure distributions (discreteness) and variant scattering characteristics (variability) of SAR airplane targets lead to special challenges of object detection and recognition. The current deep learning-based detectors encounter challenges in distinguishing fine-grained SAR airplanes against complex backgrounds. To address it, we propose a novel physics-guided detector (PGD) learning paradigm for SAR airplanes that comprehensively investigate their discreteness and variability to improve the detection performance. It is a general learning paradigm that can be extended to different existing deep learning-based detectors with "backbone-neck-head" architectures. The main contributions of PGD include the physics-guided self-supervised learning, feature enhancement, and instance perception, denoted as PGSSL, PGFE, and PGIP, respectively. PGSSL aims to construct a self-supervised learning task based on a wide range of SAR airplane targets that encodes the prior knowledge of various discrete structure distributions into the embedded space. Then, PGFE enhances the multi-scale feature representation of a detector, guided by the physics-aware information learned from PGSSL. PGIP is constructed at the detection head to learn the refined and dominant scattering point of each SAR airplane instance, thus alleviating the interference from the complex background. We propose two implementations, denoted as PGD and PGD-Lite, and apply them to various existing detectors with different backbones and detection heads. The experiments demonstrate the flexibility and effectiveness of the proposed PGD, which can improve existing detectors on SAR airplane detection with fine-grained classification task (an improvement of 3.1\% mAP most), and achieve the state-of-the-art performance (90.7\% mAP) on SAR-AIRcraft-1.0 dataset. The project is open-source at \url{https://github.com/XAI4SAR/PGD}.

cs.CV↗

arXiv2024

BlueLM-V-3B: Algorithm and System Co-Design for Multimodal Large Language Models on Mobile Devices

The emergence and growing popularity of multimodal large language models (MLLMs) have significant potential to enhance various aspects of daily life, from improving communication to facilitating learning and problem-solving. Mobile phones, as essential daily companions, represent the most effective and accessible deployment platform for MLLMs, enabling seamless integration into everyday tasks. However, deploying MLLMs on mobile phones presents challenges due to limitations in memory size and computational capability, making it difficult to achieve smooth and real-time processing without extensive optimization. In this paper, we present BlueLM-V-3B, an algorithm and system co-design approach specifically tailored for the efficient deployment of MLLMs on mobile platforms. To be specific, we redesign the dynamic resolution scheme adopted by mainstream MLLMs and implement system optimization for hardware-aware deployment to optimize model inference on mobile phones. BlueLM-V-3B boasts the following key highlights: (1) Small Size: BlueLM-V-3B features a language model with 2.7B parameters and a vision encoder with 400M parameters. (2) Fast Speed: BlueLM-V-3B achieves a generation speed of 24.4 token/s on the MediaTek Dimensity 9300 processor with 4-bit LLM weight quantization. (3) Strong Performance: BlueLM-V-3B has attained the highest average score of 66.1 on the OpenCompass benchmark among models with $\leq$ 4B parameters and surpassed a series of models with much larger parameter sizes (e.g., MiniCPM-V-2.6, InternVL2-8B).

cs.CV↗

arXiv2024

Semantic Model Component Implementation for Model-driven Semantic Communications

The key feature of model-driven semantic communication is the propagation of the model. The semantic model component (SMC) is designed to drive the intelligent model to transmit in the physical channel, allowing the intelligence to flow through the networks. According to the characteristics of neural networks with common and individual model parameters, this paper designs the cross-source-domain and cross-task semantic component model. Considering that the basic model is deployed on the edge node, the large server node updates the edge node by transmitting only the semantic component model to the edge node so that the edge node can handle different sources and different tasks. In addition, this paper also discusses how channel noise affects the performance of the model and proposes methods of injection noise and regularization to improve the noise resistance of the model. Experiments show that SMCs use smaller model parameters to achieve cross-source, cross-task functionality while maintaining performance and improving the model's tolerance to noise. Finally, a component transfer-based unmanned vehicle tracking prototype was implemented to verify the feasibility of model components in practical applications.

cs.AI↗

arXiv2024

Early Joint Learning of Emotion Information Makes MultiModal Model Understand You Better

In this paper, we present our solutions for emotion recognition in the sub-challenges of Multimodal Emotion Recognition Challenge (MER2024). To mitigate the modal competition issue between audio and text, we adopt an early fusion strategy based on a large language model, where joint training of audio and text is conducted initially. And the joint Audio-Text modal feature will be late-fused with other unimodal features. In order to solve the problems of data insufficiency and class imbalance, We use multiple turns of multi-model voting for data mining. Moreover, to enhance the quality of audio features, we employ speech source separation to preprocess audios. Our model ranks \textbf{2nd} in both MER2024-SEMI and MER2024-NOISE, validating our method's effectiveness.

cs.MM↗

arXiv2024

A specialisation theorem for Lang-N\'eron groups

We show that, for a polarised smooth projective variety $B \hookrightarrow \mathbb{P}^n_k$ of dimension $\geq 2$ over an infinite field $k$ and an abelian variety $A$ over the function field of $B$, there exists a dense Zariski open set of smooth geometrically connected hyperplane sections $h$ of $B$ such that $A$ has good reduction at $h$ and the specialisation homomorphism of Lang-N\'eron groups at $h$ is injective (up to a finite $p$-group in positive characteristic $p$). This gives a positive answer to a conjecture of the first author, which is used to deduce a negative definiteness result on his refined height pairing. This also sheds a new light on N\'eron's specialisation theorem.

math.AG↗

arXiv2024

Chow trace of 1-motives and the Lang-N\'eron groups

We show that in the case of primary field extensions, the extension of scalars of Deligne $1$-motives admits a left adjoint, called Chow image, and a right adjoint, called Chow trace. This generalizes Chow's results on abelian varieties. Then we study the Chow trace in the framework of Voevodsky's triangulated categories of (\'etale) motives. With respect to the $1$-motivic $t$-structure on the category of Voevodsky's homological $1$-motives, the zero-th direct image of an abelian variety is given by the Chow trace, and the first direct image is the $0$-motive defined by the (geometric) Lang-N\'eron group.

math.AG↗

arXiv2024

Strongly asymmetric magnetization switching and programmable complete Boolean logic enabled by long-range intralayer Dzyaloshinskii-Moriya interaction

Electrical switching of magnetization is central to spintronics. Despite the enormous efforts on the spin torques and the Dzyaloshinskii-Moriya interaction (DMI) effects, some fundamental physics for electrical switching of magnetization is still missing as indicated by a number of remarkable long-standing puzzles. Here, we report the discovery of the long-range intralayer DMI effect widely existing in magnetic heterostructure, which is distinct from the yet-known DMI effects as it describes the chiral coupling of two orthogonal magnetic domains within the same magnetic layer via the mediation of an adjacent heavy metal layer. The long-range intralayer DMI generates a strong perpendicular effective magnetic field (H_DMI^z) on the perpendicular magnetization. Characteristically, H_DMI^z varies with the sign/magnitude of the interfacial DMI constant, the applied in-plane magnetic fields, and the distribution of the perpendicular magnetic anisotropy. The long-range intralayer DMI results in striking consequences including the strongly asymmetric current/field switching of perpendicular magnetization, hysteresis loop shift of perpendicular magnetization in the absence of in-plane direct current, and sharp, complete switching of perpendicular magnetization purely by an in-plane magnetic field. Utilizing the long-range intralayer DMI effect, we demonstrate programable, complete Boolean logic operations (i.e., AND, NAND, NOT, OR, and NOR) within a single spin-orbit torque device. These results will stimulate the investigation of the long-range intralayer DMI effect and its impacts on a variety of spintronic devices.

cond-mat.mtrl-sci↗

arXiv2024

Active Generation Network of Human Skeleton for Action Recognition

Data generation is a data augmentation technique for enhancing the generalization ability for skeleton-based human action recognition. Most existing data generation methods face challenges to ensure the temporal consistency of the dynamic information for action. In addition, the data generated by these methods lack diversity when only a few training samples are available. To solve those problems, We propose a novel active generative network (AGN), which can adaptively learn various action categories by motion style transfer to generate new actions when the data for a particular action is only a single sample or few samples. The AGN consists of an action generation network and an uncertainty metric network. The former, with ST-GCN as the Backbone, can implicitly learn the morphological features of the target action while preserving the category features of the source action. The latter guides generating actions. Specifically, an action recognition model generates prediction vectors for each action, which is then scored using an uncertainty metric. Finally, UMN provides the uncertainty sampling basis for the generated actions.

cs.CV↗