arXiv subjects

Yafeng Yin

Publications and source records attributed to Yafeng Yin.

At least 19 recordsLinked to original sources

arXiv2026

SignLlama: Enhancing Gloss-free Sign Language Translation by Prioritizing Visual Features for LLMs

Large Language Models (LLMs) have achieved remarkable success across a wide range of tasks. However, fine-tuning LLMs for Gloss-Free Sign Language Translation (GFSLT) remains a challenge. In this paper, we investigate how to effectively adapt LLMs to the GFSLT task. We show that there are two key issues that need to be solved: (1) the inherent distributional gap between visual feature inputs and text feature inputs makes it difficult for LLMs to interpret visual inputs; and (2) existing approaches typically concatenate visual and textual features in an autoregressive framework, which leads to the model overemphasizing textual inputs and deprioritizing visual cues, as LLMs are pretrained predominantly on text-centric data. To address the first challenge, we propose a simple yet effective method named Filtered Pseudo-Gloss CTC Pretraining, which leverages filtered pseudo-gloss sequences generated from text sequences to supervise the training of the visual backbone. To tackle the second issue, we introduce a Visual-Prioritized Distillation training strategy. Specifically, we define a visual-only prediction path in which text inputs are masked, and the model is required to generate the target sequence relying solely on visual inputs. To guide this path, the outputs from the standard visual-textual prediction are then distilled into the visual-only prediction path, encouraging the model to prioritize visual features. Comprehensive experiments and qualitative analyses demonstrate the effectiveness of the proposed model. The proposed SignLlama achieves very competitive performance on multiple datasets for GFSLT tasks, without using any extra modalities or external sign language datasets for pretraining.

cs.CV↗

arXiv2026

Bridging the Gap Between Semantics and Reconstruction:Unifying Sign Language Translation and Production

Recent advances in sign language (SL) research have shown a trend toward unifying multiple sign language understanding (SLU) subtasks, such as isolated sign language recognition (ISLR), continuous sign language recognition (CSLR), and sign language translation (SLT), within a single framework, leading to substantial progress. Meanwhile, sign language production (SLP), which generates sign sequences from text, has also attracted growing attention. This naturally raises an important question: can sign language understanding and production be unified within a single framework? Compared with unifying SLU subtasks, this problem is substantially more challenging. Existing SLU tasks largely share the same direction of mapping, namely from sign inputs to linguistic outputs, whereas SLT and SLP lie in opposite directions of sign-text mapping. A unified framework must therefore address two key challenges: (1) bridging the modality gap between continuous sign motions and discrete text tokens through a shared sign tokenizer that supports both linguistic abstraction and motion reconstruction; and (2) learning a single conditional autoregressive model that can take either sign or text as input and generate the corresponding target sequence in the opposite modality. To this end, we propose Uni-SLTP, a unified framework for SLT and SLP with two key components: (1) a shared sign tokenizer that converts sign sequences into discrete tokens and latent representations, capturing both semantic and reconstructive information; and (2) a unified autoregressive generation model that formulates both tasks as conditional sequence generation. Experiments on widely used public datasets show that Uni-SLTP achieves superior motion accuracy for SLP while maintaining competitive SLT performance.

cs.CL↗

arXiv2026

Sign Language Question Answering: A New Task, Benchmark, and Baseline for Sign Language Understanding

Recent advances in sign language (SL) understanding (SLU) have led to remarkable progress in tasks such as continuous SL recognition and SL translation. However, these tasks are designed with predefined objectives, requiring models to learn a fixed mapping from sign videos to glosses or spoken-language sentences. As a result, they provide only a limited assessment of whether a model truly understands the semantic content of SL videos. To address this limitation, \textbf{we first propose a new task, Sign Language Question Answering (SLQA)}, which evaluates SL understanding by requiring models to answer arbitrary natural language questions about SL videos. Unlike previous SLU tasks, SLQA provides a more flexible and comprehensive evaluation framework that assesses multiple reasoning capabilities beyond recognition and translation. To facilitate this task, \textbf{we further construct two SignQA benchmarks} based on PHOENIX14T and CSL-Daily by automatically generating question-answer pairs from existing gloss and sentence annotations using carefully designed templates. The resulting datasets cover five complementary question categories, including position reasoning, structural reasoning, visual search, gloss recognition, and translation understanding. \textbf{Finally, we propose a simple yet effective baseline model} equipped with a Question-Conditioned Modulated Temporal Downsampling module and an in-domain knowledge transfer strategy, enabling effective knowledge transfer from existing SLU tasks while enhancing question-aware temporal feature modeling. Extensive experiments demonstrate that our baseline consistently outperforms representative vision-language models across all question categories, establishing a strong benchmark for future research on SLQA. Datasets are available at:{https://huggingface.co/datasets/hulala/SignQA-2026}.

cs.AI↗

arXiv2026

ScaleErasure: Inference-Time Minimal Intervention for Precise Concept Erasure in Next-Scale Autoregressive Image Generation

Concept erasure aims to prevent image generative models from producing unsafe content while preserving their general generative capability. Meanwhile, next-scale autoregressive (AR) image generation has recently emerged as a new generative paradigm characterized by next-scale prediction, for which concept erasure remains largely unexplored. In this paradigm, semantic information is highly compressed at early scales, leading to severe entanglement between unsafe and unrelated semantics. In this paper, we propose ScaleErasure, an inference-time concept erasure method that performs minimal intervention. ScaleErasure precisely selects and guides predicted logits that are most relevant to the unsafe concept, thereby enabling effective erasure under severe semantic entanglement. Specifically, ScaleErasure performs two additional forward passes conditioned on the unsafe concept and the corresponding safe concept, and leverages their outputs to guide the target logits away from unsafe concepts toward safe concepts. To enable precise and minimal intervention, logits selection and guidance are conducted across three dimensions: scales, tokens, and bit channels. Experiments demonstrate that ScaleErasure outperforms adapted baselines in the next-scale AR paradigm, achieving more precise concept erasure while largely preserving general generative capability. The code is available at https://github.com/coziiizz/ScaleErasure.

cs.CV↗

arXiv2026

Multi-Label Test-Time Adaptation with Bayesian Conditional Priors

Multi-label recognition with frozen Vision-Language Models (VLMs) is brittle under distribution shift: standard zero-shot inference scores labels independently, ignoring co-occurrence structure and producing incoherent label sets where dominant concepts suppress weaker but compatible labels. We introduce Bayesian Conditional Priors (BCP) Estimation, a gradient-free test-time adaptation method that injects label dependency without tuning the backbone. BCP views zero-shot logits as a proxy for marginal posteriors under a fixed image-text likelihood and attributes shift-induced errors mainly to a mismatched label prior. For each test image, it selects a high-confidence anchor label and applies an anchor-conditioned Bayesian refinement. This update is closed-form in logit space and admits a pointwise mutual information (PMI) interpretation, explicitly promoting compatible labels and suppressing incompatible ones. BCP operates without target annotations by estimating anchor-conditioned priors online from the unlabeled test stream via lightweight second-order co-occurrence statistics, adding negligible overhead beyond a single forward pass. Across standard multi-label benchmarks and multiple CLIP backbones, BCP consistently outperforms strong TTA baselines, e.g., improving RN50 average mAP from 57.31 to 69.22 and ViT-B/16 from 62.61 to 71.79.

cs.CV↗

arXiv2026

InfoMerge: Information-aware Token Compression for Efficient Video Large Language Models

Video Large Language Models (Video-LLMs) achieve strong performance in video understanding, but their excessive visual tokens bring substantial computational overhead. Existing training-free compression methods improve inference efficiency by reducing visual tokens, yet they often rely on local adjacent-frame similarity for temporal redundancy estimation or allocate token budgets mainly according to segment length. Such designs are sensitive to frame-level noise and fail to capture the non-uniform information distribution of real-world videos. To address these challenges, we propose InfoMerge, a training-free visual token compression method that improves token utilization through robust redundancy estimation and content-aware budget allocation. Specifically, we propose the Temporal Fingerprint Difference: a segment-level second-order temporal redundancy estimation strategy, which models the temporal similarity structure of tokens at the same spatial positions within each segment. We further introduce Content-Aware Budget Allocation (CABA), which dynamically allocates segment-level token budgets based on segment uniqueness and spectral-entropy-based representational richness. By reducing repeated preservation of redundant static regions and allocating more tokens to informative segments, InfoMerge makes better use of the limited token budget while maintaining strong performance. Extensive experiments show that InfoMerge achieves strong efficiency--accuracy trade-offs across multiple benchmarks and backbones, with more pronounced advantages under aggressive compression. On LLaVA-OneVision-7B, InfoMerge retains 98.8\% of the original average performance while reducing 85\% of visual tokens and achieving a 4.24-fold speedup in the prefill stage.

cs.CV↗

arXiv2026

Statistical Inference of Day-to-Day Traffic Dynamics

Day-to-day traffic dynamics are widely used to model flow evolution due to travelers' learning and adjustment behavior, yet empirical analysis of these models often relies on descriptive calibration with limited inferential content. This paper develops a statistical inference framework for day-to-day route choice dynamics based on a stochastic individual-level adjustment model. The framework enables uncertainty quantification and formal inference for behavioral parameters from trajectory data. We establish identifiability and consistency under mild conditions, and extend the framework to accommodate demand variation, user heterogeneity through a hierarchical structure, and anonymized observability caused by privacy constraints on trajectory data. Simulation studies demonstrate good finite-sample performance, calibrated uncertainty, and robustness to model misspecification. Empirical analyses of controlled laboratory experiments and real-world trajectory data from Ann Arbor, Michigan, show that the framework can generate novel behavioral insights across settings: it reveals the inadequacy of a purely inter-day learning model once en-route information is introduced, recovers systematic behavioral differences across participant types, and uncovers meaningful day-to-day learning together with substantial demand variation in real-world commuting behavior.

math.OC↗

arXiv2026

Fenchel-Young Estimators of Perturbed Utility Models

The Perturbed Utility Model (PUM) framework provides a generalization of discrete choice analysis, unifying models like Multinomial Logit (MNL) and Sparsemax through convex optimization. However, standard Maximum Likelihood Estimation (MLE) encounters theoretical and computational limitations when applied to this broader class, particularly regarding non-convexity and instability in sparse regimes. To address these issues, this paper introduces a unified estimation framework for PUMs based on the Fenchel-Young loss. By leveraging the intrinsic convex conjugate structure of the choice probabilities, we demonstrate that the Fenchel-Young estimator guarantees global convexity, providing a stable alternative to MLE that accommodates both dense and sparse choice kernels. Furthermore, we establish the framework's asymptotic consistency and normality under standard regularity conditions. Leveraging the tractability of the Fenchel-Young estimator, we further develop a Parametric Basis Estimation (PBE) procedure that estimate utility parameters jointly with a tree-structured perturbation function within a pre-specified basis family. PBE employs a bi-level optimization architecture that parameterizes the unknown perturbation as a learnable convex combination of basis functions. For any fixed perturbation structure, the inner Fenchel--Young estimation problem is globally convex in the utility parameters, yielding a well-defined solution mapping that can be differentiated under regularity conditions. Empirical validation on the Swissmetro dataset demonstrates that the proposed framework improves predictive performance, as measured by the Brier score and Brier Skill Score, compared to the standard MNL baseline.

math.OC↗

arXiv2025

RegionMarker: A Region-Triggered Semantic Watermarking Framework for Embedding-as-a-Service Copyright Protection

Embedding-as-a-Service (EaaS) is an effective and convenient deployment solution for addressing various NLP tasks. Nevertheless, recent research has shown that EaaS is vulnerable to model extraction attacks, which could lead to significant economic losses for model providers. For copyright protection, existing methods inject watermark embeddings into text embeddings and use them to detect copyright infringement. However, current watermarking methods often resist only a subset of attacks and fail to provide \textit{comprehensive} protection. To this end, we present the region-triggered semantic watermarking framework called RegionMarker, which defines trigger regions within a low-dimensional space and injects watermarks into text embeddings associated with these regions. By utilizing a secret dimensionality reduction matrix to project onto this subspace and randomly selecting trigger regions, RegionMarker makes it difficult for watermark removal attacks to evade detection. Furthermore, by embedding watermarks across the entire trigger region and using the text embedding as the watermark, RegionMarker is resilient to both paraphrasing and dimension-perturbation attacks. Extensive experiments on various datasets show that RegionMarker is effective in resisting different attack methods, thereby protecting the copyright of EaaS.

cs.CL↗

arXiv2025

Aligning LLM agents with human learning and adjustment behavior: a dual agent approach

Effective modeling of how human travelers learn and adjust their travel behavior from interacting with transportation systems is critical for system assessment and planning. However, this task is also difficult due to the complex cognition and decision-making involved in such behavior. Recent research has begun to leverage Large Language Model (LLM) agents for this task. Building on this, we introduce a novel dual-agent framework that enables continuous learning and alignment between LLM agents and human travelers on learning and adaptation behavior from online data streams. Our approach involves a set of LLM traveler agents, equipped with a memory system and a learnable persona, which serve as simulators for human travelers. To ensure behavioral alignment, we introduce an LLM calibration agent that leverages the reasoning and analytical capabilities of LLMs to train the personas of these traveler agents. Working together, this dual-agent system is designed to track and align the underlying decision-making mechanisms of travelers and produce realistic, adaptive simulations. Using a real-world dataset from a day-to-day route choice experiment, we show our approach significantly outperforms existing LLM-based methods in both individual behavioral alignment and aggregate simulation accuracy. Furthermore, we demonstrate that our method moves beyond simple behavioral mimicry to capture the evolution of underlying learning processes, a deeper alignment that fosters robust generalization. Overall, our framework provides a new approach for creating adaptive and behaviorally realistic agents to simulate travelers' learning and adaptation that can benefit transportation simulation and policy analysis.

cs.AI↗

arXiv2025

Managing ride-sourcing drivers at transportation terminals: a lottery-based queueing approach

Problem definition: Transportation terminals such as airports often experience persistent oversupply of idle ride-sourcing drivers, resulting in long driver waiting times and inducing externalities such as curbside congestion. While platforms now employ virtual queues with control levers like dynamic pricing, information provision, and direct admission control to manage this issue, all existing levers involve significant trade-offs and side effects. This limitation highlights the need for an alternative management approach. Methodology/results: We develop a queueing-theoretic framework to model ride-sourcing operations at terminals and propose a novel lottery-based control mechanism for the virtual queue. This non-monetary strategy works by probabilistically assigning a driver's entry position. By directly influencing their expected waiting time, the mechanism in turn shapes their decision to join the queue. We reformulate the resulting infinite-dimensional, non-smooth optimization into a tractable bi-level program by leveraging the threshold structure of the equilibrium. Theoretically, we prove that the lottery mechanism can achieve higher or equal social welfare than FIFO-queue-based dynamic pricing. Numerical experiments in unconstrained markets show that in profit maximization, our approach only narrowly trails dynamic pricing and significantly outperforms static pricing. Furthermore, it is shown that under commission fee caps, the lottery mechanism can surpass dynamic pricing in profitability. Implications: This study introduces a new, non-monetary lever for managing idle ride-sourcing drivers at transportation terminals. By aligning operational practices with queue-based dynamics, the proposed lottery mechanism offers a robust and implementable alternative to pricing-based approaches, with advantages in both unconstrained and regulated markets.

math.OC↗

arXiv2025

Steering When Necessary: Flexible Steering Large Language Models with Backtracking

Large language models (LLMs) have achieved remarkable performance across many generation tasks. Nevertheless, effectively aligning them with desired behaviors remains a significant challenge. Activation steering is an effective and cost-efficient approach that directly modifies the activations of LLMs during the inference stage, aligning their responses with the desired behaviors and avoiding the high cost of fine-tuning. Existing methods typically indiscriminately intervene to all generations or rely solely on the question to determine intervention, which limits the accurate assessment of the intervention strength. To this end, we propose the Flexible Activation Steering with Backtracking (FASB) framework, which dynamically determines both the necessity and strength of intervention by tracking the internal states of the LLMs during generation, considering both the question and the generated content. Since intervening after detecting a deviation from the desired behavior is often too late, we further propose the backtracking mechanism to correct the deviated tokens and steer the LLMs toward the desired behavior. Extensive experiments on the TruthfulQA dataset and six multiple-choice datasets demonstrate that our method outperforms baselines. Our code will be released at https://github.com/gjw185/FASB.

cs.CL↗

arXiv2025

Cross-Modal Prototype Augmentation and Dual-Grained Prompt Learning for Social Media Popularity Prediction

Social Media Popularity Prediction is a complex multimodal task that requires effective integration of images, text, and structured information. However, current approaches suffer from inadequate visual-textual alignment and fail to capture the inherent cross-content correlations and hierarchical patterns in social media data. To overcome these limitations, we establish a multi-class framework , introducing hierarchical prototypes for structural enhancement and contrastive learning for improved vision-text alignment. Furthermore, we propose a feature-enhanced framework integrating dual-grained prompt learning and cross-modal attention mechanisms, achieving precise multimodal representation through fine-grained category modeling. Experimental results demonstrate state-of-the-art performance on benchmark metrics, establishing new reference standards for multimodal social media analysis.

cs.IR↗

arXiv2025

Hierarchical Vision-Language Reasoning for Multimodal Multiple-Choice Question Answering

Multimodal Large Language Models (MLLMs) have demonstrated remarkable multimodal understanding capabilities in Visual Question Answering (VQA) tasks by integrating visual and textual features. However, under the challenging ten-choice question evaluation paradigm, existing methods still exhibit significant limitations when processing PDF documents with complex layouts and lengthy content. Notably, current mainstream models suffer from a strong bias toward English training data, resulting in suboptimal performance for Japanese and other language scenarios. To address these challenges, this paper proposes a novel Japanese PDF document understanding framework that combines multimodal hierarchical reasoning mechanisms with Colqwen-optimized retrieval methods, while innovatively introducing a semantic verification strategy through sub-question decomposition. Experimental results demonstrate that our framework not only significantly enhances the model's deep semantic parsing capability for complex documents, but also exhibits superior robustness in practical application scenarios.

cs.IR↗

arXiv2025

Valuing Time in Silicon: Can Large Language Models Replicate Human Value of Travel Time

As a key advancement in artificial intelligence, large language models (LLMs) are set to transform transportation systems. While LLMs offer the potential to simulate human travelers in future mixed-autonomy transportation systems, their behavioral fidelity in complex scenarios remains largely unconfirmed by existing research. This study addresses this gap by conducting a comprehensive analysis of the value of travel time (VOT) of three popular LLMs. We employ a full factorial experimental design to systematically examine LLMs' sensitivities to various transportation contexts, including the choice setting, travel purpose, and socio-demographic factors. Our results reveal a high degree of behavioral similarity between LLMs and humans. Some LLMs exhibit an aggregate VOT similar to that of humans, and all tested models demonstrate human-like sensitivity to travel purpose, income, and the time-cost trade-off ratios of the alternatives. Furthermore, the behavioral patterns of LLMs are highly consistent across varied contexts. However, while the behavior of every single model is highly robust, we also find some heterogeneity across models regarding the magnitude and direction of sensitivity to travel contexts and income elasticity. Overall, this study provides a foundational benchmark for the future development of LLMs as proxies for human travelers, demonstrating their robust decision-making capabilities while cautioning that misaligned magnitudes of economic trade-offs between humans and LLMs necessitate rigorous validation and additional conditioning of LLMs before their application.

econ.GN↗

arXiv2025

Advanced Sign Language Video Generation with Compressed and Quantized Multi-Condition Tokenization

Sign Language Video Generation (SLVG) seeks to generate identity-preserving sign language videos from spoken language texts. Existing methods primarily rely on the single coarse condition (\eg, skeleton sequences) as the intermediary to bridge the translation model and the video generation model, which limits both the naturalness and expressiveness of the generated videos. To overcome these limitations, we propose SignViP, a novel SLVG framework that incorporates multiple fine-grained conditions for improved generation fidelity. Rather than directly translating error-prone high-dimensional conditions, SignViP adopts a discrete tokenization paradigm to integrate and represent fine-grained conditions (\ie, fine-grained poses and 3D hands). SignViP contains three core components. (1) Sign Video Diffusion Model is jointly trained with a multi-condition encoder to learn continuous embeddings that encapsulate fine-grained motion and appearance. (2) Finite Scalar Quantization (FSQ) Autoencoder is further trained to compress and quantize these embeddings into discrete tokens for compact representation of the conditions. (3) Multi-Condition Token Translator is trained to translate spoken language text to discrete multi-condition tokens. During inference, Multi-Condition Token Translator first translates the spoken language text into discrete multi-condition tokens. These tokens are then decoded to continuous embeddings by FSQ Autoencoder, which are subsequently injected into Sign Video Diffusion Model to guide video generation. Experimental results show that SignViP achieves state-of-the-art performance across metrics, including video quality, temporal coherence, and semantic fidelity. The code is available at https://github.com/umnooob/signvip/.

cs.CV↗

arXiv2025

Aligning LLM with human travel choices: a persona-based embedding learning approach

The advent of large language models (LLMs) presents new opportunities for travel demand modeling. However, behavioral misalignment between LLMs and humans presents obstacles for the usage of LLMs, and existing alignment methods are frequently inefficient or impractical given the constraints of typical travel demand data. This paper introduces a novel framework for aligning LLMs with human travel choice behavior, tailored to the current travel demand data sources. Our framework uses a persona inference and loading process to condition LLMs with suitable prompts to enhance alignment. The inference step establishes a set of base personas from empirical data, and a learned persona loading function driven by behavioral embeddings guides the loading process. We validate our framework on the Swissmetro mode choice dataset, and the results show that our proposed approach significantly outperformed baseline choice models and LLM-based simulation models in predicting both aggregate mode choice shares and individual choice outcomes. Furthermore, we showcase that our framework can generate insights on population behavior through interpretable parameters. Overall, our research offers a more adaptable, interpretable, and resource-efficient pathway to robust LLM-based travel behavior simulation, paving the way to integrate LLMs into travel demand modeling practice in the future.

cs.AI↗

arXiv2025

Contrastive Prompting Enhances Sentence Embeddings in LLMs through Inference-Time Steering

Extracting sentence embeddings from large language models (LLMs) is a practical direction, as it requires neither additional data nor fine-tuning. Previous studies usually focus on prompt engineering to guide LLMs to encode the core semantic information of the sentence into the embedding of the last token. However, the last token in these methods still encodes an excess of non-essential information, such as stop words, limiting its encoding capacity. To this end, we propose a Contrastive Prompting (CP) method that introduces an extra auxiliary prompt to elicit better sentence embedding. By contrasting with the auxiliary prompt, CP can steer existing prompts to encode the core semantics of the sentence, rather than non-essential information. CP is a plug-and-play inference-time intervention method that can be combined with various prompt-based methods. Extensive experiments on Semantic Textual Similarity (STS) tasks and downstream classification tasks demonstrate that our method can improve the performance of existing prompt-based methods across different LLMs. Our code will be released at https://github.com/zifengcheng/CP.

cs.CL↗