arXiv subjects

Yilin Li

Publications and source records attributed to Yilin Li.

At least 19 recordsLinked to original sources

arXiv2026

AdaptPrint: Response-Adaptive Fingerprinting of Black-Box LLM Services

Black-box LLM services have emerged as a practical deployment paradigm. Nevertheless, their opacity also hinders the systematic assessment of security risks and complicates copyright auditing for model owners. Black-box LLM fingerprinting, which identifies the underlying LLM identity through query-response interactions, offers a promising way to bridge this gap. Existing approaches typically collect responses from target LLM services using a fixed set of queries and perform poorly in the presence of realistic and complex configurations (e.g., system prompt and sampling settings). To overcome these limitations, we propose AdaptPrint, a response-adaptive fingerprinting method for revealing hidden LLM identities in black-box LLM services. AdaptPrint integrates three progressive response consistency probing strategies: Direct Probing, Continuation Probing, and Follow-up Probing. AdaptPrint determines the final LLM identity by performing similarity matching among candidate LLMs. Experimental results show that AdaptPrint significantly outperforms state-of-the-art methods among 27 candidate models, achieving Top-1, Top-3, and Top-5 accuracies of 80.6%, 90.3%, and 92.1%. AdaptPrint also demonstrates strong robustness across different defense strategies and decoding parameters.

cs.CR↗

arXiv2026

Ultrafast configuration changes and anomalous diffusion of an aromatic adsorbate on rare-gas nanoparticles

Nanoparticles (NPs) exhibit tunable catalytic properties and serve as nanoreactors for controlled multimolecular chemistry. The kinetics and reactivity of such systems are critically governed by the surface binding configurations of adsorbates, their stochastic fluctuations, and the adsorbate mobility across the nanosurface. However, resolving these properties with sufficient structural, spatial, and temporal resolution remains a major experimental challenge. Here, we study phthalocyanine adsorbates on rare-gas clusters as a test case. By combining high-resolution two-dimensional electronic spectroscopy and molecular dynamics simulations, we reveal the configurational dynamics of the adsorbates and establish a direct relation between these dynamics and the nanoscale properties of the clusters. Our findings indicate sub-diffusive surface motion and trapping of the adsorbate within single surface facets. Such dynamical behavior seems unexpected considering the weak adsorbate-surface interaction and cluster temperatures close to the sublimation point. These results provide direct insight into the ultrafast binding dynamics of molecular adsorbates on nanoscale objects, which is critical for our understanding of the chemistry of such systems.

physics.chem-ph↗

arXiv2026

Size characterization of neutral rare-gas clusters based on time-resolved polarization anisotropy measurements

The size determination of neutral clusters is experimentally challenging. In particular, weakly-bound rare-gas clusters tend to fragment upon ionization, resulting in systematic errors in cluster size studies. In contrast, characterization of the temporal polarization anisotropy dephasing provides a soft detection scheme for cluster size estimation, which avoids fragmentation of the clusters. Here, we present a systematic experimental study of argon and neon clusters in the size range of 50 to 10.000 atoms using this technique. In order to extract the mean cluster sizes from the data, we present an efficient analytical model of the polarization anisotropy dephasing of an ensemble of doped clusters. The approach shows remarkable sensitivity to small changes in the mean cluster size of just a few tens of atoms and allows us to refine the widely used Hagena scaling law for the estimation of rare-gas cluster sizes.

physics.atm-clus↗

arXiv2026

The Limits of Price Discrimination with a Bayesian Seller

We study the limits of third-degree price discrimination when the production cost is Bayesian and private to the seller, generalizing the seminal work of Bergemann, Brooks and Morris (2015). The rough setup is the following: A monopoly seller sets different prices for buyers in different "segments" of the market so as to maximize seller surplus. Different ways in which the aggregate market is decomposed into segments lead to different welfare outcomes, i.e., (seller surplus, buyer surplus) pairs. When the production cost is Bayesian, the region of achievable welfare outcomes can exhibit complex shapes beyond the clean characterization by Bergemann, Brooks and Morris for the case with a fixed cost. We show that with a Bayesian cost, this region coincides with a proper projection of a polytope defined by a polynomial number of linear constraints, the essential ones of which correspond to flow conservation in a "discounted" flow network. As a result, we give a polynomial-time algorithm that computes optimal market segmentations in terms of any linear combination of the seller surplus and the buyer surplus. En route, we establish the following structural property: Any market can be written as a convex combination of "extremal markets" in a way preserving the seller surplus and the buyer surplus. These extremal markets are piecewise equal-surplus with respect to different possible costs, generalizing a similar notion introduced by Bergemann, Brooks and Morris when the cost is fixed.

econ.TH↗

arXiv2026

Causal Estimation of Share-Induced Engagement with Flywheel Effects

Sustainable user growth in online platforms depends not only on acquiring new users but also on reactivating and engaging existing ones through social sharing features. A well-designed sharing feature can trigger a self-reinforcing ``flywheel effect'': reactivated users become potential sharers whose engagement propagates through the network over multiple rounds, amplifying total engagement. Measuring the causal impact of such sharing features is challenging, as their effects unfold through complex social networks and temporal cascades, violating the no-interference assumption underlying classical A/B testing. We develop a framework for experiments on sharing features that accounts for interference caused by the flywheel effect and targets a global treatment effect on share-induced engagement. Our estimator is motivated by a flow-balance identity and interprets share-induced engagement as a geometric amplification process, yielding a closed-form propagation adjustment that accounts for multi-round diffusion using commonly available attribution logs. Under mild conditions, we establish consistency of the proposed estimator and develop a valid A/A testing procedure for pipeline validation. Simulation studies show that our method substantially reduces bias relative to the difference-in-means estimator and first-order adjustments, while the proposed A/A test maintains nominal Type I error. We also extend the framework to a user-level reactivation metric via a Poisson approximation. Finally, we demonstrate the approach on a real-world large-scale online platform and discuss empirical implications for evaluating sharing feature designs.

stat.ME↗

arXiv2026

Design-based edge-level causal inference with machine learning assisted covariate adjustment

We study design-based causal inference for edge-level outcomes in directed networks under dyadic interference. In this setting, outcomes are defined on directed edges and depend on the joint treatment assignments of pairs of units, inducing a complex dependence structure that invalidates standard estimation and inference procedures developed for node-level data. We construct Horvitz--Thompson estimators for a general class of edge-level causal effects and establish their asymptotic normality under mild regularity conditions. To enable valid inference, we develop variance estimators that exploit identifiable components of network dependence, yielding substantially less conservative bounds than classical approaches. To improve efficiency, we incorporate auxiliary covariates through a sample splitting and cross-fitting procedure. A key technical challenge is that standard two-fold sample splitting fails in the presence of edge-level outcomes due to the dependence induced by shared units. To address this issue, we introduce a three-fold sample splitting and cross-fitting scheme that restores the conditional independence required for unbiased estimation. Under a stability condition, the resulting covariate-adjusted estimator is asymptotically normal and accommodates both linear adjustment and flexible machine learning methods. We further introduce a calibration step that guarantees no asymptotic efficiency loss relative to the unadjusted estimator. Simulation studies and a real-data application confirm the theoretical results and demonstrate substantial efficiency gains.

stat.ME↗

arXiv2026

MHSA: A Lightweight Framework for Mitigating Hallucinations via Steered Attention in LVLMs

Large vision-language models (LVLMs) have achieved remarkable performance across diverse multimodal tasks, yet they continue to suffer from hallucinations, generating content that is inconsistent with the visual input. Prior work DHCP (Detecting Hallucinations by Cross-modal Attention Pattern) has explored hallucination detection from the perspective of cross-modal attention, but does not address hallucination mitigation. In this paper, we propose MHSA (Mitigating Hallucinations via Steered Attention), a lightweight framework that mitigates hallucinations by learning to correct cross-modal attention patterns in LVLMs. MHSA trains a simple three-layer MLP generator to produce corrected attention, guided by supervisory signals from the DHCP discriminator and the LVLM itself. During inference, MHSA mitigates both discriminative and generative hallucinations across various datasets and LVLMs by simply replacing the original cross-modal attention with the corrected one, without modifying any LVLM parameters. By extending cross-modal attention mechanisms from hallucination detection to hallucination mitigation, MHSA offers a novel perspective on hallucination research in LVLMs and helps enhance their reliability.

cs.CV↗

arXiv2026

Beyond GSD-as-Token: Continuous Scale Conditioning for Remote Sensing VLMs

Remote sensing vision-language models (RS-VLMs) face a fundamental mismatch with natural-image counterparts: the same geographic object exhibits radically different visual evidence across ground sampling distances (GSDs) spanning multiple orders of magnitude. Yet existing RS-VLMs often discard GSD or inject it as a discrete text token, forcing a single static parameter set to absorb the entire scale spectrum. We introduce ScaleEarth, a parameter-efficient fine-tuning framework built on Qwen3-VL that treats GSD as a continuous conditioning variable governing the model's computation path. At its core, CS-HLoRA (Continuous Scale-Conditioned Hyper-LoRA) modulates the LoRA low-rank subspace through a GSD-driven gate, enabling the model to dynamically route computation by physical scale. To remove reliance on sensor metadata at deployment, we pair CS-HLoRA with SSE-U, a lightweight heteroscedastic sub-head that predicts GSD and its uncertainty from visual features. To provide matching supervision, we construct GeoScale-VQA, a 1.5M-sample scale-layered RS-VQA corpus whose question-answer generation is conditioned on the same physical scalar that drives CS-HLoRA, forming a closed method-data loop. Trained with QLoRA on an 8B backbone, ScaleEarth achieves state-of-the-art results on remote-sensing benchmarks covering diverse Earth-system tasks, including XLRS-Bench and OmniEarth-Bench.

cs.CV↗

arXiv2026

CaliCausalRank: Calibrated Multi-Objective Ad Ranking with Robust Counterfactual Utility Optimization

Ad ranking systems must simultaneously optimize multiple objectives including click-through rate (CTR), conversion rate (CVR), revenue, and user experience metrics. However, production systems face critical challenges: score scale inconsistency across traffic segments undermines threshold transferability, and position bias in click logs causes offline-online metric discrepancies. We propose CaliCausalRank, a unified framework that integrates training-time scale calibration, constraint-based multi-objective optimization, and robust counterfactual utility estimation. Our approach treats score calibration as a first-class training objective rather than post-hoc processing, employs Lagrangian relaxation for constraint satisfaction, and utilizes variance-reduced counterfactual estimators for reliable offline evaluation. Experiments on the Criteo and Avazu datasets demonstrate that CaliCausalRank achieves 1.1% relative AUC improvement, 31.6% calibration error reduction, and 3.2% utility gain compared to the best baseline (PairRank) while maintaining consistent performance across different traffic segments.

cs.LG↗

arXiv2026

Semantics-Aware Denoising: A PLM-Guided Sample Reweighting Strategy for Robust Recommendation

Implicit feedback, such as user clicks, serves as the primary data source for modern recommender systems. However, click interactions inherently contain substantial noise, including accidental clicks, clickbait-induced interactions, and exploratory browsing behaviors that do not reflect genuine user preferences. Training recommendation models with such noisy positive samples leads to degraded prediction accuracy and unreliable recommendations. In this paper, we propose SAID (Semantics-Aware Implicit Denoising), a simple yet effective framework that leverages semantic consistency between user interests and item content to identify and downweight potentially noisy interactions. Our approach constructs textual user interest profiles from historical behaviors and computes semantic similarity with target item descriptions using pre-trained language model (PLM) based text encoders. The similarity scores are then transformed into sample weights that modulate the training loss, effectively reducing the impact of semantically inconsistent clicks. Unlike existing denoising methods that require complex auxiliary networks or multi-stage training procedures, SAID only modifies the loss function while keeping the backbone recommendation model unchanged. Extensive experiments on two real-world datasets demonstrate that SAID consistently improves recommendation performance, achieving up to 2.2% relative improvement in AUC over strong baselines, with particularly notable robustness under high noise conditions.

cs.IR↗

arXiv2025

No-Regret Online Autobidding Algorithms in First-price Auctions

Automated bidding to optimize online advertising with various constraints, e.g. ROI constraints and budget constraints, is widely adopted by advertisers. A key challenge lies in designing algorithms for non-truthful mechanisms with ROI constraints. While prior work has addressed truthful auctions or non-truthful auctions with weaker benchmarks, this paper provides a significant improvement: We develop online bidding algorithms for repeated first-price auctions with ROI constraints, benchmarking against the optimal randomized strategy in hindsight. In the full feedback setting, where the maximum competing bid is observed, our algorithm achieves a near-optimal $\widetilde{O}(\sqrt{T})$ regret bound, and in the bandit feedback setting (where the bidder only observes whether the bidder wins each auction), our algorithm attains $\widetilde{O}(T^{3/4})$ regret bound.

cs.GT↗

arXiv2025

Following the TRACE: A Structured Path to Empathetic Response Generation with Multi-Agent Models

Empathetic response generation is a crucial task for creating more human-like and supportive conversational agents. However, existing methods face a core trade-off between the analytical depth of specialized models and the generative fluency of Large Language Models (LLMs). To address this, we propose TRACE, Task-decomposed Reasoning for Affective Communication and Empathy, a novel framework that models empathy as a structured cognitive process by decomposing the task into a pipeline for analysis and synthesis. By building a comprehensive understanding before generation, TRACE unites deep analysis with expressive generation. Experimental results show that our framework significantly outperforms strong baselines in both automatic and LLM-based evaluations, confirming that our structured decomposition is a promising paradigm for creating more capable and interpretable empathetic agents. Our code is available at https://anonymous.4open.science/r/TRACE-18EF/README.md.

cs.CL↗

arXiv2025

Subset Selection for Stratified Sampling in Online Controlled Experiments

Online controlled experiments, also known as A/B testing, are the digital equivalent of randomized controlled trials for estimating the impact of marketing campaigns on website visitors. Stratified sampling is a traditional technique for variance reduction to improve the sensitivity (or statistical power) of controlled experiments; this technique first divides the population into strata (homogeneous subgroups) based on stratification variables and then draws samples from each stratum to avoid sampling bias. To enhance the estimation accuracy of stratified sampling, we focus on the problem of selecting a subset of stratification variables that are effective in variance reduction. We design an efficient algorithm that selects stratification variables one by one by simulating a series of stratified sampling processes. We also estimate the computational complexity of our subset selection algorithm. Computational experiments using synthetic and real-world datasets demonstrate that our method can outperform other variance reduction techniques especially when multiple variables have a certain correlation with the outcome variable. Our subset selection method for stratified sampling can improve the sensitivity of online controlled experiments, thus enabling more reliable marketing decisions.

stat.CO↗

arXiv2025

NeuroDeX: Unlocking Diverse Support in Decompiling Deep Neural Network Executables

On-device deep learning models have extensive real world demands. Deep learning compilers efficiently compile models into executables for deployment on edge devices, but these executables may face the threat of reverse engineering. Previous studies have attempted to decompile DNN executables, but they face challenges in handling compilation optimizations and analyzing quantized compiled models. In this paper, we present NeuroDeX to unlock diverse support in decompiling DNN executables. NeuroDeX leverages the semantic understanding capabilities of LLMs along with dynamic analysis to accurately and efficiently perform operator type recognition, operator attribute recovery and model reconstruction. NeuroDeX can recover DNN executables into high-level models towards compilation optimizations, different architectures and quantized compiled models. We conduct experiments on 96 DNN executables across 12 common DNN models. Extensive experimental results demonstrate that NeuroDeX can decompile non-quantized executables into nearly identical high-level models. NeuroDeX can recover functionally similar high-level models for quantized executables, achieving an average top-1 accuracy of 72%. NeuroDeX offers a more comprehensive and effective solution compared to previous DNN executables decompilers.

cs.LG↗

arXiv2025

When Code Crosses Borders: A Security-Centric Study of LLM-based Code Translation

Code translation is crucial for cross-language codebase migration, and large language models (LLMs) have emerged as a promising technique to automate this process. However, the security implications of using LLMs for code translation remain largely unexplored, as existing evaluations primarily focus on syntactic and functional correctness. To bridge this gap, we conduct a security-centric empirical study to investigate the risks of vulnerabilities being introduced or preserved during LLM-based translation. Our study involves a rigorous evaluation of five state-of-the-art LLMs on a curated dataset of 720 security-related code samples across four programming languages (Java, PHP, C, C++) and nine Common Weakness Enumeration (CWE) categories. The results reveal significant security degradation, with 28.6\% to 45\% of translations introducing new vulnerabilities. Web-related flaws, particularly in input validation, proved most challenging for LLMs. Furthermore, we identify and categorize the root causes of these vulnerable translations into a taxonomy of five major error types. Based on our findings, we develop and evaluate a Retrieval-Augmented Generation (RAG)-based mitigation strategy, which successfully reduces the vulnerability introduction rate by 32.8\%. Our study provides the first large-scale evidence of serious security risks in LLM-based code translation and demonstrates the potential of knowledge-enhanced prompting to mitigate them.

cs.CR↗

arXiv2025

VaccineRAG: Boosting Multimodal Large Language Models' Immunity to Harmful RAG Samples

Retrieval Augmented Generation enhances the response accuracy of Large Language Models (LLMs) by integrating retrieval and generation modules with external knowledge, demonstrating particular strength in real-time queries and Visual Question Answering tasks. However, the effectiveness of RAG is frequently hindered by the precision of the retriever: many retrieved samples fed into the generation phase are irrelevant or misleading, posing a critical bottleneck to LLMs' performance. To address this challenge, we introduce VaccineRAG, a novel Chain-of-Thought-based retrieval-augmented generation dataset. On one hand, VaccineRAG employs a benchmark to evaluate models using data with varying positive/negative sample ratios, systematically exposing inherent weaknesses in current LLMs. On the other hand, it enhances models' sample-discrimination capabilities by prompting LLMs to generate explicit Chain-of-Thought (CoT) analysis for each sample before producing final answers. Furthermore, to enhance the model's ability to learn long-sequence complex CoT content, we propose Partial-GRPO. By modeling the outputs of LLMs as multiple components rather than a single whole, our model can make more informed preference selections for complex sequences, thereby enhancing its capacity to learn complex CoT. Comprehensive evaluations and ablation studies on VaccineRAG validate the effectiveness of the proposed scheme. The code and dataset will be publicly released soon.

cs.CL↗

arXiv2025

Collaborative Evolution of Intelligent Agents in Large-Scale Microservice Systems

This paper proposes an intelligent service optimization method based on a multi-agent collaborative evolution mechanism to address governance challenges in large-scale microservice architectures. These challenges include complex service dependencies, dynamic topology structures, and fluctuating workloads. The method models each service as an agent and introduces graph representation learning to construct a service dependency graph. This enables agents to perceive and embed structural changes within the system. Each agent learns its policy based on a Markov Decision Process. A centralized training and decentralized execution framework is used to integrate local autonomy with global coordination. To enhance overall system performance and adaptability, a game-driven policy optimization mechanism is designed. Through a selection-mutation process, agent strategy distributions are dynamically adjusted. This supports adaptive collaboration and behavioral evolution among services. Under this mechanism, the system can quickly respond and achieve stable policy convergence when facing scenarios such as sudden workload spikes, topology reconfigurations, or resource conflicts. To evaluate the effectiveness of the proposed method, experiments are conducted on a representative microservice simulation platform. Comparative analyses are performed against several advanced approaches, focusing on coordination efficiency, adaptability, and policy convergence performance. Experimental results show that the proposed method outperforms others in several key metrics. It significantly improves governance efficiency and operational stability in large-scale microservice systems. The method demonstrates strong practical value and engineering feasibility.

cs.DC↗

arXiv2025

Harnessing Rule-Based Reinforcement Learning for Enhanced Grammatical Error Correction

Grammatical error correction is a significant task in NLP. Traditional methods based on encoder-decoder models have achieved certain success, but the application of LLMs in this field is still underexplored. Current research predominantly relies on supervised fine-tuning to train LLMs to directly generate the corrected sentence, which limits the model's powerful reasoning ability. To address this limitation, we propose a novel framework based on Rule-Based RL. Through experiments on the Chinese datasets, our Rule-Based RL framework achieves \textbf{state-of-the-art }performance, with a notable increase in \textbf{recall}. This result clearly highlights the advantages of using RL to steer LLMs, offering a more controllable and reliable paradigm for future development in GEC.

cs.CL↗