arXiv subjects

Lin Chen

Publications and source records attributed to Lin Chen.

At least 163 records · Page 9Linked to original sources

arXiv2025

WITNESS: A lightweight and practical approach to fine-grained predictive mutation testing

Existing fine-grained predictive mutation testing studies predominantly rely on deep learning, which faces two critical limitations in practice: (1) Exorbitant computational costs. The deep learning models adopted in these studies demand significant computational resources for training and inference acceleration. This introduces high costs and undermines the cost-reduction goal of predictive mutation testing. (2) Constrained applicability. Although modern mutation testing tools generate mutants both inside and outside methods, current fine-grained predictive mutation testing approaches handle only inside-method mutants. As a result, they cannot predict outside-method mutants, limiting their applicability in real-world scenarios. We propose WITNESS, a new fine-grained predictive mutation testing approach. WITNESS adopts a twofold design: (1) With collected features from both inside-method and outside-method mutants, WITNESS is suitable for all generated mutants. (2) Instead of using computationally expensive deep learning, WITNESS employs lightweight classical machine learning models for training and prediction. This makes it more cost-effective and enabling straightforward explanations of the decision-making processes behind the adopted models. Evaluations on Defects4J projects show that WITNESS consistently achieves state-of-the-art predictive performance across different scenarios. Additionally, WITNESS significantly enhances the efficiency of kill matrix prediction. Post-hoc analysis reveals that features incorporating information from before and after the mutation are the most important among those used in WITNESS. Test case prioritization based on the predicted kill matrix shows that WITNESS delivers results much closer to those obtained by using the actual kill matrix, outperforming baseline approaches.

cs.SE↗

arXiv2025

ECCENTRIC: Edge-Cloud Collaboration Framework for Distributed Inference Using Knowledge Adaptation

The massive growth in the utilization of edge AI has made the applications of machine learning models ubiquitous in different domains. Despite the computation and communication efficiency of these systems, due to limited computation resources on edge devices, relying on more computationally rich systems on the cloud side is inevitable in most cases. Cloud inference systems can achieve the best performance while the computation and communication cost is dramatically increasing by the expansion of a number of edge devices relying on these systems. Hence, there is a trade-off between the computation, communication, and performance of these systems. In this paper, we propose a novel framework, dubbed as Eccentric that learns models with different levels of trade-offs between these conflicting objectives. This framework, based on an adaptation of knowledge from the edge model to the cloud one, reduces the computation and communication costs of the system during inference while achieving the best performance possible. The Eccentric framework can be considered as a new form of compression method suited for edge-cloud inference systems to reduce both computation and communication costs. Empirical studies on classification and object detection tasks corroborate the efficacy of this framework.

cs.DC↗

arXiv2025

Agentar-DeepFinance-100K: A Large-Scale Financial Dataset via Systematic Chain-of-Thought Synthesis Optimization

Recent advancements in large language models (LLMs) have demonstrated remarkable general reasoning capabilities, holding significant potential for applications in the financial domain, a field that requires robust and reliable reasoning. It has been demonstrated that distilling high-quality chain-of-thought (CoT) rationales from advanced general reasoning models offers a promising and efficient path to the financial reasoning model. However, existing CoT synthesis methods suffer from shallow CoT sampling, leaving the question of how to construct a well-designed knowledge space for finance reasoning unexplored. In this paper, we present Agentar-DeepFinance-100K, a large-scale financial reasoning dataset characterized by its systematic CoT synthesis optimization. We first introduce a comprehensive CoT synthesis pipeline featuring Multi-perspective Knowledge Extraction (MKE) and Self-Corrective Rewriting (SCR) to generate exhaustive and deep financial reasoning trajectories. Furthermore, a systematic investigation, termed CoT Cube, is conducted to analyze critical factors that influence CoT effectiveness, such as necessity, length and synthesizer, yielding valuable insights for high-quality financial CoT construction. Experiments demonstrate that models trained on our Agentar-DeepFinance-100K achieve significant improvements on financial benchmarks. We publicly release Agentar-DeepFinance-100K , hoping to advance the research in financial reasoning models.

cs.CE↗

arXiv2025

FlowRL: Matching Reward Distributions for LLM Reasoning

We propose FlowRL: matching the full reward distribution via flow balancing instead of maximizing rewards in large language model (LLM) reinforcement learning (RL). Recent advanced reasoning models adopt reward-maximizing methods (\eg, PPO and GRPO), which tend to over-optimize dominant reward signals while neglecting less frequent but valid reasoning paths, thus reducing diversity. In contrast, we transform scalar rewards into a normalized target distribution using a learnable partition function, and then minimize the reverse KL divergence between the policy and the target distribution. We implement this idea as a flow-balanced optimization method that promotes diverse exploration and generalizable reasoning trajectories. We conduct experiments on math and code reasoning tasks: FlowRL achieves a significant average improvement of $10.0\%$ over GRPO and $5.1\%$ over PPO on math benchmarks, and performs consistently better on code reasoning tasks. These results highlight reward distribution-matching as a key step toward efficient exploration and diverse reasoning in LLM reinforcement learning.

cs.LG↗

arXiv2025

Spark Transformer: Reactivating Sparsity in FFN and Attention

The discovery of the lazy neuron phenomenon in trained Transformers, where the vast majority of neurons in their feed-forward networks (FFN) are inactive for each token, has spurred tremendous interests in activation sparsity for enhancing large model efficiency. While notable progress has been made in translating such sparsity to wall-time benefits, modern Transformers have moved away from the ReLU activation function crucial to this phenomenon. Existing efforts on re-introducing activation sparsity often degrade model quality, increase parameter count, complicate or slow down training. Sparse attention, the application of sparse activation to the attention mechanism, often faces similar challenges. This paper introduces the Spark Transformer, a novel architecture that achieves a high level of activation sparsity in both FFN and the attention mechanism while maintaining model quality, parameter count, and standard training procedures. Our method realizes sparsity via top-k masking for explicit control over sparsity level. Crucially, we introduce statistical top-k, a hardware-accelerator-friendly, linear-time approximate algorithm that avoids costly sorting and mitigates significant training slowdown from standard top-$k$ operators. Furthermore, Spark Transformer reallocates existing FFN parameters and attention key embeddings to form a low-cost predictor for identifying activated entries. This design not only mitigates quality loss from enforced sparsity, but also enhances wall-time benefit. Pretrained with the Gemma-2 recipe, Spark Transformer demonstrates competitive performance on standard benchmarks while exhibiting significant sparsity: only 8% of FFN neurons are activated, and each token attends to a maximum of 256 tokens. This sparsity translates to a 2.5x reduction in FLOPs, leading to decoding wall-time speedups of up to 1.79x on CPU and 1.40x on GPU.

cs.LG↗

arXiv2025

Revisiting RFID Missing Tag Identification

We revisit the problem of missing tag identification in RFID networks by making three contributions. Firstly, we quantitatively compare and gauge the existing propositions spanning over a decade on missing tag identification. We show that the expected execution time of the best solution in the literature is $Θ\left(N+\frac{(1-α)^2(1-δ)^2}{ ε^2}\right)$, where $δ$ and $ε$ are parameters quantifying the required identification accuracy, $N$ denotes the number of tags in the system, among which $αN$ tags are missing. Secondly, we analytically establish the expected execution time lower-bound for any missing tag identification algorithm as $Θ\left(\frac{N}{\log N}+\frac{(1-δ)^2(1-α)^2}{ε^2 \log \frac{(1-δ)(1-α)}ε}\right)$, thus giving the theoretical performance limit. Thirdly, we develop a novel missing tag identification algorithm by leveraging a tree structure with the expected execution time of $Θ\left(\frac{\log\log N}{\log N}N+\frac{(1-α)^2(1-δ)^2}{ ε^2}\right)$, reducing the time overhead by a factor of up to $\log N$ over the best algorithm in the literature. The key technicality in our design is a novel data structure termed as collision-partition tree (CPT), built on a subset of bits in tag pseudo-IDs, leading to more balanced tree structure and reducing the time complexity in parsing the entire tree.

cs.NI↗

arXiv2025

RainSeer: Fine-Grained Rainfall Reconstruction via Physics-Guided Modeling

Reconstructing high-resolution rainfall fields is essential for flood forecasting, hydrological modeling, and climate analysis. However, existing spatial interpolation methods-whether based on automatic weather station (AWS) measurements or enhanced with satellite/radar observations often over-smooth critical structures, failing to capture sharp transitions and localized extremes. We introduce RainSeer, a structure-aware reconstruction framework that reinterprets radar reflectivity as a physically grounded structural prior-capturing when, where, and how rain develops. This shift, however, introduces two fundamental challenges: (i) translating high-resolution volumetric radar fields into sparse point-wise rainfall observations, and (ii) bridging the physical disconnect between aloft hydro-meteors and ground-level precipitation. RainSeer addresses these through a physics-informed two-stage architecture: a Structure-to-Point Mapper performs spatial alignment by projecting mesoscale radar structures into localized ground-level rainfall, through a bidirectional mapping, and a Geo-Aware Rain Decoder captures the semantic transformation of hydro-meteors through descent, melting, and evaporation via a causal spatiotemporal attention mechanism. We evaluate RainSeer on two public datasets-RAIN-F (Korea, 2017-2019) and MeteoNet (France, 2016-2018)-and observe consistent improvements over state-of-the-art baselines, reducing MAE by over 13.31% and significantly enhancing structural fidelity in reconstructed rainfall fields.

cs.LG↗

arXiv2025

Entanglement distillation on symmetric two-qutrit entangled states of rank five

Entanglement distillation is a key step in quantum information, both theoretically and practically. It has been proven that non-positive-partial transpose (NPT) entangled states of rank at most four is 1-distillable under local operation and classical communications. In this paper we investigate the distillation of a more complex family of NPT entangled states, namely a family of symmetric two-qutrit states $\r$ of rank five with given eigenvectors. We explicitly construct five families of such states by requiring four of the five eigenvalues to be the same. We respectively show that some of them are 1-distillable. It turns out that such states may be not 1-distillable for some interval of eigenvalues. We provide some conditions for eigenvalues that allow $\r$ to be 1-distillable or to be 1-undistillable.

quant-ph↗

arXiv2025

Virtual Quantum Markov Chain of four-qubit systems

We extend the framework of virtual quantum Markov chains (VQMCs) from tripartite systems to the four-qubit setting. Structural criteria such as the kernel-inclusion condition are analyzed, showing that they are necessary but not sufficient for the existence of a valid recovery map. By explicit examples, we demonstrate that the four-qubit W state admits a recovery channel and thus qualifies as a VQMC, while the GHZ state does not. We further provide semidefinite programming (SDP) formulations to test recoverability and quantify sampling overhead, and establish the non-convexity of the VQMC set through mixed-state counterexamples. These results supply both theoretical insights and computational tools for studying correlations and recoverability in multipartite quantum systems.

quant-ph↗

arXiv2025

Approximating Quantum States with Positive Partial Transposes in Multipartite System via Linearized Proximal Alternative Direction Method of Multipliers

Numerical approximation of quantum states via convex combinations of states with positive partial transposes (bi-PPT state) in multipartite systems constitutes a fundamental challenge in quantum information science. We reformulate this problem as a linearly constrained optimization problem. An approximate model is constructed through an auxiliary variable and a suitable penalty parameter, balancing constraint violation and approximation error. To slove the approximate model, we design a Linearized Proximal Alternating Direction Method of Multipliers (LPADMM), proving its convergence under a prescribed inequality condition on regularization parameters. The algorithm achieves an iteration complexity of $O(1/ε^2)$ for attaining $ε$-stationary solutions. Numerical validation on diverse quantum systems, including three-qubit W/GHZ states and five-partite GHZ and multiGHZ states with noises, confirms high-quality bi-PPT approximations and decomposability certification, demonstrating the utility of our method for quantum information applications.

math-ph↗

arXiv2025

Spectral characterizations of entanglement witnesses

We present a systematic investigation of the spectral properties of entanglement witnesses (EWs). Specifically, we analyze the infimum and supremum of the largest eigenvalue, the smallest eigenvalue, the negativity (defined as the absolute value of the sum of negative eigenvalues), and the squared Frobenius norm of a unit-trace (normalized) entanglement witness, along with the conditions under which these values are attained. Our study provides distinct characterizations for decomposable (DEWs) and nondecomposable entanglement witnesses (NDEWs). While these two classes share many spectral similarities, we reveal a fundamental divergence by proving that the infimum of the smallest eigenvalue can be attained by DEWs, yet remains strictly unattainable for all NDEWs. We apply the results to provide necessary conditions for an EW to possess a mirrored EW. Furthermore, we demonstrate the superior detection capability of NDEWs by proving that any non-positive-transpose (NPT) state beyond the two-qubit and qubit-qutrit systems can be detected by an NDEW.

quant-ph↗

arXiv2025

Strong convergence rates of stochastic theta methods for index 1 stochastic differential algebraic equations under non-globally Lipschitz conditions

This work investigates numerical approximations of index 1 stochastic differential algebraic equations (SDAEs) with non-constant singular matrices under non-global Lipschitz conditions. Analyzing the strong convergence rates of numerical solutions in this setting is highly nontrivial, due to both the singularity of the constraint matrix and the superlinear growth of the coefficients. To address these challenges, we develop an approach for establishing mean square convergence rates of numerical methods for SDAEs under global monotonicity conditions. Specifically, we prove that each stochastic theta method with $θ\in [\frac{1}{2},1]$ achieves a mean square convergence rate of order $\frac{1}{2}$. Theoretical findings are further validated through a series of numerical experiments.

math.NA↗

arXiv2025

Towards Adaptive ML Benchmarks: Web-Agent-Driven Construction, Domain Expansion, and Metric Optimization

Recent advances in large language models (LLMs) have enabled the emergence of general-purpose agents for automating end-to-end machine learning (ML) workflows, including data analysis, feature engineering, model training, and competition solving. However, existing benchmarks remain limited in task coverage, domain diversity, difficulty modeling, and evaluation rigor, failing to capture the full capabilities of such agents in realistic settings. We present TAM Bench, a diverse, realistic, and structured benchmark for evaluating LLM-based agents on end-to-end ML tasks. TAM Bench features three key innovations: (1) A browser automation and LLM-based task acquisition system that automatically collects and structures ML challenges from platforms such as Kaggle, AIcrowd, and Biendata, spanning multiple task types and data modalities (e.g., tabular, text, image, graph, audio); (2) A leaderboard-driven difficulty modeling mechanism that estimates task complexity using participant counts and score dispersion, enabling scalable and objective task calibration; (3) A multi-dimensional evaluation framework incorporating performance, format compliance, constraint adherence, and task generalization. Based on 150 curated AutoML tasks, we construct three benchmark subsets of different sizes -- Lite, Medium, and Full -- designed for varying evaluation scenarios. The Lite version, with 18 tasks and balanced coverage across modalities and difficulty levels, serves as a practical testbed for daily benchmarking and comparative studies.

cs.AI↗

arXiv2025

Equality condition for a matrix inequality by partial transpose

The partial transpose map is a linear map widely used quantum information theory. We study the equality condition for a matrix inequality generated by partial transpose, namely $\rank(\sum^K_{j=1} A_j^T \otimes B_j)\le K \cdot \rank(\sum^K_{j=1} A_j \otimes B_j)$, where $A_j$'s and $B_j$'s are respectively the matrices of the same size, and $K$ is the Schmidt rank. We explicitly construct the condition when $A_i$'s are column or row vectors, or $2\times 2$ matrices. For the case where the Schmidt rank equals the dimension of $A_j$, we extend the results from $2\times 2$ matrices to square matrices, and further to rectangular matrices. In detail, we show that $\sum^K_{j=1} A_j \otimes B_j$ is locally equivalent to an elegant block-diagonal form consisting solely of identity and zero matrices. We also study the general case for $K=2$, and it turns out that the key is to characterize the expression of matrices $A_j$'s and $B_j$'s.

quant-ph↗

arXiv2025

Unified Construction of Genuine Multipartite Entanglement Measures Based on Geometric Mean and its Applications

Genuine multipartite entanglement (GME) is an important resource in quantum information processing. We systematically study the measures of GME based on the geometric mean of bi-partition entanglements and present a unified construction of GME measures, which gives rise to the widely used GME measures including GME concurrence, the convex-roof extended negativity of GME, the geometric measure of entanglement of GME. Our GME measures satisfy the desirable conditions such as scalability and smoothness. Moreover, we provide fidelity-based analytical lower bounds for our GME measures. Our bounds are tight and can be estimated experiment friendly without requiring quantum state tomography. Furthermore, we apply our results to study the dynamics of GME. We identify an initial condition that influences the sudden death of genuine quadripartite entanglement under individual non-Markovian processes. The GME of Dirac particles with Hawking radiation in the background of a Schwarzschild black hole is also investigated.

quant-ph↗

arXiv2025

SYNAPSE-G: Bridging Large Language Models and Graph Learning for Rare Event Classification

Scarcity of labeled data, especially for rare events, hinders training effective machine learning models. This paper proposes SYNAPSE-G (Synthetic Augmentation for Positive Sampling via Expansion on Graphs), a novel pipeline leveraging Large Language Models (LLMs) to generate synthetic training data for rare event classification, addressing the cold-start problem. This synthetic data serve as seeds for semi-supervised label propagation on a similarity graph constructed between the seeds and a large unlabeled dataset. This identifies candidate positive examples, subsequently labeled by an oracle (human or LLM). The expanded dataset then trains/fine-tunes a classifier. We theoretically analyze how the quality (validity and diversity) of the synthetic data impacts the precision and recall of our method. Experiments on the imbalanced SST2 and MHS datasets demonstrate SYNAPSE-G's effectiveness in finding positive labels, outperforming baselines including nearest neighbor search.

cs.LG↗

arXiv2025

FOBNN: Fast Oblivious Inference via Binarized Neural Networks

The remarkable performance of deep learning has sparked the rise of Deep Learning as a Service (DLaaS), allowing clients to send their personal data to service providers for model predictions. A persistent challenge in this context is safeguarding the privacy of clients' sensitive data. Oblivious inference allows the execution of neural networks on client inputs without revealing either the inputs or the outcomes to the service providers. In this paper, we propose FOBNN, a Fast Oblivious inference framework via Binarized Neural Networks. In FOBNN, through neural network binarization, we convert linear operations (e.g., convolutional and fully-connected operations) into eXclusive NORs (XNORs) and an Oblivious Bit Count (OBC) problem. For secure multiparty computation techniques, like garbled circuits or bitwise secret sharing, XNOR operations incur no communication cost, making the OBC problem the primary bottleneck for linear operations. To tackle this, we first propose the Bit Length Bounding (BLB) algorithm, which minimizes bit representation to decrease redundant computations. Subsequently, we develop the Layer-wise Bit Accumulation (LBA) algorithm, utilizing pure bit operations layer by layer to further boost performance. We also enhance the binarized neural network structure through link optimization and structure exploration. The former optimizes link connections given a network structure, while the latter explores optimal network structures under same secure computation costs. Our theoretical analysis reveals that the BLB algorithm outperforms the state-of-the-art OBC algorithm by a range of 17% to 55%, while the LBA exhibits an improvement of nearly 100%. Comprehensive proof-of-concept evaluation demonstrates that FOBNN outperforms prior art on popular benchmarks and shows effectiveness in emerging bioinformatics.

cs.CR↗

arXiv2025

Prototype Embedding Optimization for Human-Object Interaction Detection in Livestreaming

Livestreaming often involves interactions between streamers and objects, which is critical for understanding and regulating web content. While human-object interaction (HOI) detection has made some progress in general-purpose video downstream tasks, when applied to recognize the interaction behaviors between a streamer and different objects in livestreaming, it tends to focuses too much on the objects and neglects their interactions with the streamer, which leads to object bias. To solve this issue, we propose a prototype embedding optimization for human-object interaction detection (PeO-HOI). First, the livestreaming is preprocessed using object detection and tracking techniques to extract features of the human-object (HO) pairs. Then, prototype embedding optimization is adopted to mitigate the effect of object bias on HOI. Finally, after modelling the spatio-temporal context between HO pairs, the HOI detection results are obtained by the prediction head. The experimental results show that the detection accuracy of the proposed PeO-HOI method has detection accuracies of 37.19%@full, 51.42%@non-rare, 26.20%@rare on the publicly available dataset VidHOI, 45.13%@full, 62.78%@non-rare and 30.37%@rare on the self-built dataset BJUT-HOI, which effectively improves the HOI detection performance in livestreaming.

cs.CV↗