arXiv subjects

Yi Chen

Publications and source records attributed to Yi Chen.

At least 37 records · Page 2Linked to original sources

arXiv2026

GameASG-Bench: Benchmarking Autonomous Software Generation for Game Development

Autonomous software generation (ASG) aims to turn human requirements into executable applications, but delivering these applications does not necessarily establish that their interacting components satisfy the specified behavioral requirements. We introduce GameASG-Bench, a benchmark that makes behavioral testability part of the generation task for game development. Our design declares an evaluation interface specification before generation, fixing legal starting scenarios, player-level actions, stable snapshots, rejection behavior, and invariants while leaving private implementations open. Concretely, we include: (i) static L1 checks that assess source-level compliance; and (ii) browser-executed L2 checks that combine semantic observations with real input and runtime evidence. We implement this protocol as 47 browser-native game-generation tasks spanning 12 primary genres and both 2D and 3D interaction, each with executable checks and an independently verified reference implementation. Our experiments answer four key questions about end-to-end agent performance, tool access and nominal turn budget, reasoning effort, and harness choice. Across nine agent stacks, the highest observed mean L2 check pass rate is 93.2%, yet the highest observed strict task success rate, requiring all L1 and applicable L2 prerequisite and core requirement checks, is only 55.3% (26/47 tasks). For DeepSeek-V4-Flash, full tool access and larger nominal turn budgets yield more strict task successes, while the strict task success rate is not monotonic in reasoning effort. Both tested harnesses achieve 18 strict task successes, but only ten tasks succeed under both. These results expose task-level compliance gaps that high average check pass rates actually obscure.

cs.AI↗

arXiv2026

ROSETTA: Efficient and Accurate Privacy-Preserving LLM Decoding via Hybrid CKKS/TFHE Evaluation

Generative large language models (LLMs) have achieved state-of-the-art performance on many real-world tasks such as code generation and question answering. These models predominantly rely on an autoregressive decoding strategy that generates output tokens sequentially. However, their pervasive deployment raises serious privacy concerns, motivating private inference frameworks based on fully homomorphic encryption (FHE). A major limitation of existing FHE frameworks is their inefficiency in evaluating nonlinear operations, which incur substantial overhead and dominate the decode stage. In this paper, we propose ROSETTA, a hybrid CKKS/TFHE framework that overcomes this limitation. We first observe that nonlinear operations in the decode stage exhibit heterogeneous workload patterns, which can be handled effectively via a hybrid approach. We then realize this with two key contributions: 1) an adaptive segmented lookup-table protocol based on TFHE that enables efficient and accurate evaluation of nonlinear operations; and 2) a scheme-aware operator-selection framework that automatically assigns each nonlinear operator to CKKS or TFHE to minimize end-to-end decoding latency. We demonstrate that ROSETTA achieves up to $4.8\times$ Softmax speedup and $1.5$--$2.1\times$ end-to-end speedup over the SOTA framework CacheMir.

cs.CR↗

arXiv2026

Symmetry-Aware Likelihood-Orbit Aggregation for Selective Left-Right Claim Verification

Frozen vision-language models (VLMs) remain unreliable on fine-grained left-right claims, and raw claim likelihoods need not reliably rank verification errors. After a horizontal-reflection intervention is fixed, how should its induced likelihood measurements be combined into a selective verification signal? We introduce Relation-Orbit, a closed-form contrast with no learned fusion parameters that assigns eight normalized likelihoods to query-supporting and counterfactual roles determined by reflection, inverse relation, and entity exchange. A claim is asserted only when the signed contrast exceeds a threshold selected on held-out data using pointwise Clopper-Pearson upper confidence bounds. On VSR and GQA across four frozen VLMs, Relation-Orbit yields higher mean test coverage at a 10% selective-risk calibration target than an all-eight Orbit-Max baseline in all eight dataset-backbone settings; gains over a nearly abstain-all one-sided intervention score are reported separately. A separate LLaVA-1.5/COCO evaluation, reduced-orbit controls, and a two-sided partition diagnostic further characterize the structural advantage.

cs.CV↗

arXiv2026

Generate to Explore, Select to Exploit: Aligning LLM-based Headline Generation with Personalized Recommendation

In industrial recommendation feeds, presenting a static headline for an item often fails to satisfy the diverse, multimodal interests of the user population, particularly suppressing the needs of long-tail audiences. While Large Language Models (LLMs) have been integrated into recommendation for content understanding or ranking, directly optimizing them to output a single best headline typically leads to mode collapse---converging to generic patterns that satisfy average tastes but miss specific latent intents. To bridge this gap, we introduce GESE (Generate to Explore, Select to Exploit), a framework operating at the system's presentation layer that decouples personalization into generative exploration and selective exploitation. First, we treat the LLM as a probabilistic explorer, utilizing Group Sequence Policy Optimization (GSPO) with a hierarchical reward mechanism to generate a candidate set that maximizes the semantic coverage of potential user interests. Subsequently, a lightweight, real-time feedback-aware selector acts as the exploiter, identifying the optimal realization from the candidate pool based on instant contextual signals. Extensive deployment on a commercial platform with over 100 million daily active users demonstrates that GESE significantly outperforms state-of-the-art baselines, achieving a 2.57% lift in CTR and 0.87% in dwell time. These results validate that decoupling diversity-oriented generation from precision-oriented selection offers a robust blueprint for aligning generative AI with dynamic user utility.

cs.IR↗

arXiv2026

TractSpLearn: Specialized Shared-Manifold Learning for Individualized Detection of Subtle White Matter Alterations in Mild Traumatic Brain Injury

Traumatic brain injury (TBI) often leads to subtle white matter damage that remains undetected on conventional MRI. Diffusion kurtosis imaging (DKI), an extension of diffusion tensor imaging (DTI), provides complementary information on non-Gaussian water diffusion and is sensitive to complex white-matter microstructure. With the advent of ultra-high-field MRI, the spatial resolution and signal-to-noise ratios (SNR) have been significantly enhanced, enabling more precise visualization of subtle abnormalities. Building on these advances, we developed TractSpLearn, an individualized tract-based learning framework that jointly considers within-group variability and between-group differences. Unlike the original TractLearn framework, which learns a normative manifold exclusively from healthy controls, TractSpLearn incorporates both healthy controls and patients to learn a shared manifold with a healthy-anchored representation and an additional patient-related component. To assess the performance of the proposed method, we compared TractSpLearn with the original TractLearn in three cohorts: (i) healthy controls (HC), (ii) athletes with persistent post-concussive syndromes (PPCS), and (iii) athletes with repeated head injuries (RHI), with abnormalities particularly evident in axial kurtosis (AK) and mean diffusivity (MD). In RHI, TractSpLearn highlighted recurrent abnormalities across patients. In the PPCS cohort, the overall group-level differences were more modest, potentially reflecting both limited statistical power due to the small sample size and partial normalization of white-matter alterations during recovery. Still TractSpLearn identified abnormality evidence in more patients and across more affected tracts than TractLearn.

q-bio.QM↗

arXiv2026

Algorithmic Attention and Content Creation on Social Media Platforms

We study revenue-maximizing attention allocation on an ad-funded social media platform governed by recommendation algorithms. Attention is costly and can be monetized through advertising or allocated to increase creators' exposure, creating a trade-off between monetization and production incentives. In a two-sided model with heterogeneous viewers and creators under private information, the optimal recommendation mix includes content that is ex post suboptimal for some viewers to leverage network externalities. These distortions are targeted: low-ability creators are excluded, while high-ability creators are subsidized through exposure or monetary payments. Two-sided complementarities reshape content variety and quality, with implications for personalization regulation and advertising markets.

econ.TH↗

arXiv2026

Focus on What Matters: Fisher-Guided Adaptive Multimodal Fusion for Vulnerability Detection

Software vulnerability detection can be formulated as a binary classification problem that determines whether a given code snippet contains security defects. Existing multimodal methods typically fuse Natural Code Sequence (NCS) representations extracted by pretrained models with Code Property Graph (CPG) representations extracted by graph neural networks, under the implicit assumption that introducing an additional modality necessarily yields information gain. Through empirical analysis, we demonstrate the limitations of this assumption: pretrained models already encode substantial structural information implicitly, leading to strong overlap between the two modalities; moreover, graph encoders are generally less effective than pretrained language models in feature extraction. As a result, naive fusion not only struggles to obtain complementary signals but can also dilute effective discriminative cues due to noise propagation. To address these challenges, we propose a task-conditioned complementary fusion strategy that uses Fisher information to quantify task relevance, transforming cross-modal interaction from full-spectrum matching into selective fusion within a task-sensitive subspace. Our theoretical analysis shows that, under an isotropic perturbation assumption, this strategy significantly tightens the upper bound on the output error. Based on this insight, we design the TaCCS-DFA framework, which combines online low-rank Fisher subspace estimation with an adaptive gating mechanism to enable efficient task-oriented fusion. Experiments on the BigVul, Devign, and ReVeal benchmarks demonstrate that TaCCS-DFA delivers up to a 6.3-point gain in F1 score with only a 3.4% increase in inference latency, while maintaining low calibration error.

cs.SE↗

arXiv2026

Nucleon 3D intrinsic spin structure from the weak-neutral axial-vector form factors

Relativistic 3D weak-neutral axial-vector four-current and spin distributions inside a nucleon (or a general spin-$\frac{1}{2}$ hadron) including three weak-neutral axial-vector form factors are investigated for the first time. We clarify that the relativistic 3D axial charge distribution in the Breit frame is completely described by the induced pseudotensor form factor $G_T^Z(Q^2)$ rather than by the axial form factor $G_A^Z(Q^2)$. We demonstrate that the quantity $R_A \equiv \sqrt{ \frac{-6}{G_A^Z(0) }\frac{\text{d} G_A^Z(Q^2) }{\text{d} Q^2} \Big|_{Q^2=0} }$ can not be interpreted as the physically meaningful 3D root-mean-square axial radius of a spin-$\frac{1}{2}$ hadron. The genuine 3D root-mean-square axial radius in fact does not exist for any spin-$\frac{1}{2}$ hadron. We also show that the relativistic 3D weak-neutral spin radius $r_\text{spin} = \sqrt{\langle r_\text{spin}^2 \rangle}$, defined as $\langle r_\text{spin}^2 \rangle \equiv R_A^2 + \frac{ 1 }{ 4M^2 }\left[ 1 + \frac{ 2 G_P^Z(0) }{ G_A^Z(0) } \right]$ based on the relativistic and intrinsic 3D weak-neutral spin distribution in the Breit frame, is a physically meaningful radius that can be unambiguously defined for the nucleon. This provides an additional key motivation for the further determination of the induced pseudoscalar form factor $G_P^Z(Q^2)$, e.g. via lattice QCD or model calculations. Numerically, we find that $R_A \approx 0.6510~\text{fm}$ and $r_\text{spin} \approx 2.1054~\text{fm}$. For future experimental measurements of $G_A^Z(Q^2)$ and $G_T^Z(Q^2)$, we also derive the full tree-level unpolarized differential cross sections for neutrino-proton and antineutrino-proton elastic scattering in the lab frame, in hoping to provide a complementary and new perspective to unveil the nucleon spin structure by using (anti)neutrino-based facilities.

hep-ph↗

arXiv2026

Harnessing the Reasoning Economy: A Survey of Efficient Reasoning for Large Language Models

Recent advancements in Large Language Models (LLMs) have significantly enhanced their ability to perform complex reasoning tasks, transitioning from fast and intuitive thinking (System 1) to slow and deep reasoning (System 2). While System 2 reasoning improves task accuracy, it often incurs substantial computational costs due to its slow thinking nature and inefficient or unnecessary reasoning behaviors. In contrast, System 1 reasoning is computationally efficient but leads to suboptimal performance. Consequently, it is critical to balance the trade-off between performance (benefits) and computational costs (budgets), giving rise to the concept of reasoning economy. In this survey, we provide a comprehensive analysis of reasoning economy in both the post-training and test-time inference stages of LLMs, encompassing i) the cause of reasoning inefficiency, ii) behavior analysis of different reasoning patterns, and iii) potential solutions to achieve reasoning economy. By offering actionable insights and highlighting open challenges, we aim to shed light on strategies for improving the reasoning economy of LLMs, thereby serving as a valuable resource for advancing research in this evolving area. We also provide a public repository to continually track developments in this fast-evolving field.

cs.CL↗

arXiv2026

DenseScout: Algorithm-System Co-design for Budgeted Tiny Object Selection on Edge Platforms

Deploying high-resolution tiny-object perception on edge platforms requires not only accurate localization, but also selecting a small set of informative patches under compute, transport, and latency constraints. We study budgeted tiny-object selection, where a frontend ranks patch centers from a lightweight proxy and a downstream detector processes only the selected regions. DenseScout is a 1.01M-parameter deployment-oriented dense-response selector that removes detector-style box regression and directly optimizes ranked patch-center prioritization. Its contribution lies in the task-specific selector formulation, the alignment among output representation, supervision, and decoding, and its joint design with transport-aware execution and QoS-oriented evaluation. Under unified protocols on VisDrone and DOTA, DenseScout provides stronger low-budget recall than the evaluated detector-derived selectors; controlled fixed-K inspection experiments further demonstrate advantages over selection-style proxy baselines. Cross-platform profiling on Jetson Orin NX and RK3588 shows that deployable utility depends jointly on selector quality, memory movement, and heterogeneous runtime realization. These results support treating edge tiny-object perception as a selection-and-deployment co-design problem rather than evaluating model accuracy and runtime in isolation.

cs.CV↗

arXiv2026

Bridging the Gap: A Longitudinal Analysis of Extended Identifiers in the Post-Cookie Era

As third-party cookies fade because of browser restrictions, the online advertising ecosystem is turning to extended identifiers (EIDs) as an alternative. EIDs are persistent user identifiers, such as hashed email addresses, that are employed to link users across domains and devices. This paper presents a 41-month longitudinal study examining EID usage in over 145 million HTTP header bidding requests sent to six major supply-side platforms (SSPs) from 616,539 websites. Our findings show that EIDs are widely used and are becoming increasingly prevalent in the digital advertising ecosystem, reaching 83.76% of studied websites by May 2025. Our analysis of the 18 popular EID providers that account for 99.42% of all transmitted EIDs in our dataset raises concerns about the readiness of EIDs as an alternative to third-party cookie tracking. In terms of accuracy, only one identity provider consistently recognizes and identifies that the visitor is a self-identified bot crawler, and many providers regularly transmit multiple EIDs for the same visitor. We also identify privacy concerns with EIDs, as 12 of the providers create persistent EIDs that can identify the same user across visits, websites, devices, and months. Finally, we found that 16 providers transmit EIDs on EU websites without user consent.

cs.NI↗

arXiv2026

A Task Force on Strong Coupling Determinations from Event Shapes

The strong coupling constant $α_s$ is a fundamental parameter of the Standard Model. Its precise determination is essential for accurately predicting, studying, and understanding processes at the Large Hadron Collider and future experiments such as the Future Circular Collider. Event shape and correlator observables measured at electron-positron colliders provide one of the cleanest environments for extracting $α_s$, thanks to their sensitivity to $α_s$ and the availability of high-precision data from the Large Electron-Positron Collider. More broadly, such observables provide an ideal setting to develop and test our understanding of the perturbative and non-perturbative elements of Quantum Chromodynamics, which will underpin the field's precision and discovery frontiers for decades to come. Despite these advances, significant discrepancies persist between different determinations of $α_s$ from event shapes, both in the extracted central values and estimated uncertainties. This document motivates the establishment of a dedicated Task Force to coordinate a community-wide effort addressing these open questions. We report on the first two-day meeting held at CERN in November 2025, summarizing the scientific discussion and documenting the experimental analyses identified as priorities during the meeting, as well as the concrete list of tasks to be carried out by the theory community in preparation for future meetings.

hep-ph↗

arXiv2026

Reinforcement Learning Enhanced LLM Agents for Complex Vehicle Routing Problems

Vehicle Routing Problems (VRPs) are fundamental combinatorial optimization problems with widespread applications in various scenarios. The advanced optimization solvers can effectively solve such problems. However, modeling complex VRP variants for solvers often requires substantial domain expertise, which limits the accessibility of advanced optimization technologies. In this paper, we propose Reinforcement Learning Enhanced LLMAgents(RLEA), a multi-agent framework designed to automate the modeling of complex VRPs. RLEA introduces a lightweight neural Planner trained with Soft Q-learning to efficiently orchestrate the actions of LLM-based agents. In addition, we equip the system with an evolutionary memory module and retrieval-augmented generation, enabling the agent to leverage both accumulated experience and external solver knowledge during program generation and refinement for solving VRPs. We evaluated 48 distinct VRP variants across various solvers. The experimental results demonstrate that RLEA outperforms the previous state-of-the-ar method, achieving a 16.67% higher success rate while significantly reducing runtime errors. These results validate that integrating reinforcement learning with LLM-based reasoning is highly effective for automated optimization modeling. The appendix is available at: https://doi.org/10.5281/zenodo.19134435.

cs.AI↗

arXiv2026

ElementCheck: Complexity-Aware Long-Form Text Factuality Evaluation via Sentence Elements

Existing long-form factuality evaluation relies on the decompose-retrieve-verify pipeline. However, the pipeline suffers from noise from claim decomposition and fixed verification granularity, resulting in unreliable results. We propose ElementCheck, a complexity-aware framework that verifies long-form outputs via sentence elements. Instead of uniformly decomposing sentences into atomic sub-claims, ElementCheck extracts entity pairs that are explicitly linked through verifiable connections in the original sentence as elements, and organizes these into an element graph. The graph topology provides a structural signal for estimating sentence complexity, enabling direct verification for simple sentences and targeted element-level refinement and verification for complex ones. To support fine-grained evaluation, we construct a new benchmark \textbf{FastFact-Sent} by mapping isolated claims from FastFact-Bench back to their source sentences. Experiments on FastFact-Sent and two domain-specific benchmarks show ElementCheck consistently improves factuality verification across five backbone models while maintaining a favorable accuracy-cost trade-off. Further analyses demonstrate that complexity-aware verification reduces unnecessary re-verification and maintains stability across different backbones. The code is available at \href{https://github.com/gudehhh666/elementcheck.git}{Here}.

cs.CL↗

arXiv2026

Cut-ViT: Task-Specific Model Pruning via Gram Anchoring Subspace Consistency

Pruning visual foundation models has attracted considerable attention. However, existing methods focus on rigid point-to-point token alignment on a single dataset for pruning, suffering from two limitations: i) robustness degradation, and ii) task-specificity deficiency. To address these limitations, we propose a task-specific pruning pipeline, named Cut-ViT. Specifically, we first construct gram anchoring matrices from both spatial and semantic perspectives, and perform the subspace decomposition to extract the corresponding subspace bases. Basis-agnostic and residual constraints are then adopted to align the gram subspaces between the native and pruned DINOv3 models along spatial and channel dimensions, enabling subnetworks to inherit robust feature representations of native DINOv3. Furthermore, we design spectral entropy adaptation, which quantifies the information density of feature manifolds along spatial and channel dimensions, thereby adapting the pruning objective to specific downstream tasks. Experiments show that Cut-ViT requires approximately one minute on a single A100 GPU to obtain subnetworks at various sparsity levels, using only 20.9% of the time and 45.5% of the GPU memory compared with previous methods, while achieving SOTA performance on six tasks across nine datasets.

cs.CV↗

arXiv2026

Distance Is Not Enough: Forget-Retain Alignment Gap Predicts LLM Relearning Robustness

Machine unlearning aims to make a model forget specific data, yet unlearned LLMs often fail to stay unlearned: brief fine-tuning can revive removed knowledge. Existing robustness predictors rely on global weight-space displacement, but distance alone can be misleading when random or destructive updates collapse performance. We argue that relearning robustness depends on update structure: robust unlearning should affect forget-critical weights while sparing retain-critical ones. We introduce the Forget-Retain Alignment Gap (FRAG), a training-free predictor that scores an update's forget-retain alignment without running a relearning attack, and separates selective from dense updates more reliably than global distance. Building on the forget-critical, retain-sparing principle, Forget-Retain Pruning (FRP) improves relearning robustness. Our results suggest that weight selectivity better explains robustness than distance alone. Code is available at https://github.com/Yi1-Chen/FRAG.

cs.AI↗

arXiv2026

Emergent trans-moiré orbitals and topology in rhombohedral graphene

The fractional quantum anomalous Hall effect (FQAHE) exhibited in fractional Chern insulators has recently been demonstrated in twisted MoTe2 and rhombohedral graphene/hBN moiré superlattices, promising new routes toward topological quantum computation. Central to realizing this promise is the understanding of the underlying microscopic mechanism. This, however, remains elusive in the case of rhombohedral graphene, with the crux being its two seemingly paradoxical conditions: a pronounced small-twist-angle (θ) moiré interface, yet only when electrons are kept distant from it. Here, by scanning tunnelling microscopic imaging with both conditions fulfilled, we capture dramatic electronic structure reshaping in rhombohedral hexalayer graphene by unforeseen 'trans-moiré orbitals', which emerge on the other, distant side of the moiré interface but nevertheless enforce the moiré periodicity at all measured fillings. We visualize a hierarchy of spatially and energetically distinct trans-moiré orbitals which doped electrons must sequentially occupy--the lowest-energy orbital, expectedly responsible for the FQAHE at small fillings, carries a hollow-cage-like shape. Remarkably, these trans-moiré orbitals vanish at θ {\gtrsim} 1°, and so do QAHE plateaus in similar devices. Simulations reveal an interaction-driven charge-redistribution mechanism which shapes the trans-moiré orbitals and corresponding Chern minibands. With our findings providing the missing microscopic link, the paradoxical conditions find a natural explanation: electrons are not simply kept distant from a small-θ moiré interface; they are forced into topological trans-moiré orbitals, forged precisely under such conditions. Our microscopic diagnostics unlocks a wide range of possible 'synthetic' FQAHE platforms.

cond-mat.mes-hall↗

arXiv2026

Observing Macroscopic Consequences of Electroweak Anomalies with Archival DELPHI Data

In this Letter, we emphasize that asymmetries in charge flux produced in the decays of on-shell Z-bosons provide a macroscopic manifestation of electroweak anomalies in the Standard Model (SM). We propose that these can be cleanly observed using charge correlators, providing a new formulation of forward-backward asymmetry measurements that is particularly well suited for precision studies of hadronic decays. Using archival DELPHI data, we perform a first measurement of the one-point charge correlator of electromagnetic charge flux on hadrons, and cleanly observe the macroscopic imprint of the underlying anomaly. Our analysis illustrates the potential of charge correlators as precision electroweak observables, and motivates a renewed effort to resolve longstanding tensions in hadronic asymmetry measurements.

hep-ph↗