Search arXiv⌕ Search

arXiv subjects

Dehan Kong

Publications and source records attributed to Dehan Kong.

At least 19 recordsLinked to original sources
arXiv2026

WebHarbor: Docking Real Websites for Evolving GUI Agent Environments

GUI web agents require realistic and interactive environments that remain stable during both evaluation and training. Live website evaluation suffers from changing content, access restrictions, and limited coverage of authenticated workflows, while existing synthetic environments for training often lack functional depth and visual fidelity. We propose WebHarbor, a suite of 153 executable and dockerized website environments that faithfully reconstruct the visual interfaces, content, and functional workflows of real-world web services under deterministic control. We build WebHarbor through task-driven human-agent collaboration, where environments evolve as new tasks expand their functional scope, coding agents implement the required changes, and human contributors audit visual fidelity, functional correctness, and task validity. The resulting suite comprises 3,518 tasks spanning multi-step and post-authentication workflows, where existing works often fail to cover. Beyond reliable evaluation, WebHarbor provides high-fidelity synthetic web environments for cost-efficient reinforcement learning (RL). We demonstrate that a Qwen 4B-level GUI agent trained with RL in WebHarbor generalizes to unseen tasks and performs competitively with leading open-source GUI web agents across four benchmarks. The code and data are publicly available at https://github.com/aiming-lab/WebHarbor.

cs.AI↗
arXiv2026

Nash Equilibrium Individualized Treatment Regimes under Network Interference

Network interference creates a fundamental tension in optimal individualized treatment regime learning: maximizing overall population outcomes may come at the expense of some individuals' outcomes. Existing optimal individualized treatment regime methods under interference primarily target population-level objectives without accounting for individual optimality. Motivated by this tension, we introduce Nash equilibrium individualized treatment regimes, or simply Nash regimes, which directly target individual optimality under interference. Under a Nash regime, no individual can improve their conditional expected outcome through unilateral deviation when all other individuals follow the regime. We establish conditions for the existence of Nash regimes. When no exact Nash regime exists, we introduce an $α^*$-Nash regime as an alternative target that minimizes regret relative to individual optimality. We then develop Nash learning, a doubly robust procedure for estimating the proposed regimes. Under a super-population random-graph framework accommodating both sparse and dense networks, we characterize how network sparsity and treatment assignment mechanisms affect estimation performance. Simulations and a real-data application demonstrate that Nash learning improves individual optimality while maintaining competitive population-level performance.

stat.ME↗
arXiv2026

The Culture Within: Towards Mechanistic Understanding of Cultural Reasoning in LLMs

Large Language Models (LLMs) increasingly operate across culturally diverse settings, yet existing evaluations largely measure cultural behavior at the output level. We introduce CultureLens, a reference-grounded benchmark and mechanistic study of how LLMs use cultural information. CultureLens contains 9,261 scenarios across 75 countries and systematically varies country, value, true-norm, and contradictory-norm evidence. Across Llama-3.1-8B, Gemma-3-12B, and Qwen-3.5-27B, stating a contradictory norm lowers aligned answers by 32-47 percentage points, even when the model answers correctly from the country alone. Mechanistically, models default to Anglophone norms when no country is given. The failure under a contradictory norm is not one of missing knowledge, as blocking the norm inside the model restores the correct answer in most cases. Models accept a stated norm according to how it is presented and not whether it is true, which leaves them most exposed to errors in retrieved content. Supervised fine-tuning with contradictory-norm examples raises aligned answers under a contradictory norm to 85-93%, including on unseen countries, at a small cost under a true norm.

cs.CL↗
arXiv2026

Efficient Fréchet Distributional Decoder Alignment for Latent Generative Models

Fréchet Distributional (FD) objectives provide an effective distribution-level signal for post-training generative models. For latent generative models, however, existing FD-based post-training methods optimize only the latent generator while keeping the decoder frozen: the decoder is trained on reconstruction latents obtained from real images but deployed on generator-produced latents at generation time. When these latent distributions differ, this \textbf{decoder train-generation mismatch} means that the decoder is deployed on inputs drawn from a distribution it was not directly trained on. We introduce \textbf{Fréchet Distributional Decoder Alignment (FDDA)}, an efficient post-training method that freezes the latent generator and directly optimizes the decoder on generation-time latents using an FD objective. A single epoch of FDDA consistently improves gFID and gFDr$^6$ across diverse flow-based and autoregressive latent generative models. Across all 11 evaluated models, FDDA reduces gFDr$^6$ by 34.0-59.6\% from the original pretrained models and provides a further 10.2-57.2\% reduction after generator-side FD post-training. FDDA also reduces total wall-clock time by factors of $3.6\times$-$328.7\times$ relative to generator-side FD post-training across these models. These results show that generator-side post-training alone does not exhaust the benefits of FD optimization, and that further gains can be obtained by adapting the decoder at substantially lower cost.

cs.CV↗
arXiv2026

A Latent Distribution Perspective on Evaluating and Improving Latent Generative Models

In latent generative models, reconstruction quality is often assumed to correlate with generative performance. However, reconstruction FID (rFID) can exhibit weak or even negative correlation with generation FID (gFID). We attribute this misalignment to a latent distribution mismatch: reconstruction evaluates the decoder on encoder-induced latents, whereas generation uses the same decoder on latents produced by the generative model. To characterize this shift, we introduce generation-aware reconstruction (GAR), which constructs a continuous trajectory from standard reconstruction toward generation by perturbing encoder latents with noise and denoising them through the generative model before decoding. GAR probes the decoder behavior along this trajectory, making the transition from encoder to generation-time latent distributions observable and diagnosable. The resulting trajectory-based diagnostic, GAR-FID, exhibits strong empirical correlation with gFID across diverse tokenizers and scales. Importantly, intermediate GAR latents become more generation-aware while preserving correspondence with their source images, thereby retaining paired supervision that is absent for fully generated latents. This correspondence enables decoder adaptation on intermediate GAR latents, consistently improving generative quality across model scales. Overall, latent distribution mismatch provides a useful perspective for evaluating and improving latent generative models.

cs.CV↗
arXiv2026

AudioGAR: Bridging Reconstruction and Generation in Latent Audio Generative Models

Latent audio generative models are typically trained in two stages: an audio codec is learned first, followed by a latent generative model. This decomposition leads to a decoder train-generation mismatch: the codec decoder is trained on encoder-induced latents but deployed on generator-produced latents at inference time. Across diverse datasets and latent generative models, we observe clear reconstruction-generation gaps under both FD and FAD, showing that strong reconstruction quality does not necessarily translate into strong end-to-end generation quality. A natural remedy is to adapt the decoder on generation-produced latents, but generated latents lack correspondence with source audio and therefore cannot directly provide the paired supervision used for decoder fine-tuning. We introduce \textbf{AudioGAR}, which constructs intermediate latents by perturbing encoder latents and denoising them through the frozen latent diffusion model. These latents form a trajectory from reconstruction toward generation, with lower-noise latents retaining source correspondence and supporting paired decoder fine-tuning. We fine-tune only the codec decoder on these latents, while keeping the codec encoder and latent generative model frozen. When applied to AudioX, AudioGAR substantially improves generative performance. It requires only 1.5\% of the original training audio hours and 0.26\% of the original training cost.

cs.SD↗
arXiv2026

CADWorld: Computer-Use Benchmark for Long-Horizon Computer-Aided Design

Computer-use agents are increasingly evaluated in realistic desktop environments, but existing benchmarks provide limited coverage of professional engineering workflows whose outputs are persistent, structured artifacts. Mechanical computer-aided design (CAD) is a particularly demanding setting: an agent must manipulate geometry and constraints over long interaction horizons while producing a native project whose dimensions, construction structure, and downstream engineering state remain valid. We introduce \textbf{CADWorld}, a benchmark for long-horizon computer use in FreeCAD. CADWorld contains 200 tasks spanning 11 mechanical-CAD workflow categories, including sketching, part modeling, assembly, CAM, FEM, measurement, mesh processing, and technical drawing. Agents operate through screenshots and GUI actions, while success is determined by task-specific executable checks over saved FreeCAD artifacts and auxiliary outputs, covering geometric properties, parametric structure, constraints, manufacturing state, and simulation results. Across seven current agents on the full benchmark, the strongest agent achieves 17.5\% success, compared with an 87.0\% expert reference pass. We find that weaker agents often fail before producing a valid artifact, whereas stronger agents increasingly fail on structural, geometric, and construction-process requirements. CADWorld therefore exposes a gap between general GUI competence and reliable execution of persistent, verifiable engineering workflows. Project accessible at https://cad-world.github.io.

cs.AI↗
arXiv2026

A Unified Rate-Distortion Perspective on Vector, Product, and Scalar Quantization

Discrete visual tokenization, predominantly driven by vector, scalar, and product quantization, lacks a unified conceptual framework for understanding quantization tradeoffs. In this paper, we propose a unified rate--distortion perspective on modern discrete visual tokenization. By viewing quantization as lossy compression, we characterize the nominal fixed-length coding rate through token count and codebook size, and quantization error as the distortion. Within this framework, we resolve three central questions. First, we theoretically and empirically show that minimizing distortion, rather than maximizing codebook utilization, is the primary intrinsic objective for reconstruction fidelity, with a direct connection to the STE-induced gradient discrepancy. Second, we establish two critical fairness conditions for intrinsic quantization comparison: controlling latent feature statistics and enforcing identical coding rates. Third, under these conditions, we recover the VQ--PQ--SQ distortion hierarchy in modern visual tokenization and show empirically that modern VQ methods achieve the lowest distortion. This work provides a foundational rate--distortion reframing of modern discrete visual tokenization, resolves ambiguities in quantizer evaluation, and provides a controlled framework for isolating intrinsic quantization effectiveness under fixed-rate constraints.

cs.LG↗
arXiv2026

General Phrase Debiaser: Debiasing Masked Language Models at a Multi-Token Level

The social biases and unwelcome stereotypes revealed by pretrained language models are becoming obstacles to their application. Compared to numerous debiasing methods targeting word level, there has been relatively less attention on biases present at phrase level, limiting the performance of debiasing in discipline domains. In this paper, we propose an automatic multi-token debiasing pipeline called \textbf{General Phrase Debiaser}, which is capable of mitigating phrase-level biases in masked language models. Specifically, our method consists of a \textit{phrase filter stage} that generates stereotypical phrases from Wikipedia pages as well as a \textit{model debias stage} that can debias models at the multi-token level to tackle bias challenges on phrases. The latter searches for prompts that trigger model's bias, and then uses them for debiasing. State-of-the-art results on standard datasets and metrics show that our approach can significantly reduce gender biases on both career and multiple disciplines, across models with varying parameter sizes.

cs.CL↗
arXiv2026

VQ-Transplant: Efficient VQ-Module Integration for Pre-trained Visual Tokenizers

Vector Quantization (VQ) underpins modern discrete visual tokenization. However, training quantization modules for state-of-the-art VQ-based models requires significant computational resources which, in practice, all but prevents the development of novel, cutting-edge VQ techniques under resource constraints. To address this limitation, we propose {\bf VQ-Transplant}, a simple framework that enables plug-and-play integration of new VQ modules into frozen, pre-trained tokenizers by replacing their native VQ modules. Crucially, the proposed transplantation process preserves all encoder-decoder parameters, obviating the need for costly end-to-end retraining when modifying the quantization method. To mitigate decoder-quantization mismatch, we introduce a lightweight decoder adaptation strategy (trained for only 5 epochs on ImageNet-1k) to align feature priors with the new quantization space. In our empirical evaluation, we find that VQ-Transplant allows obtaining near state-of-the-art reconstruction fidelity for industry-level models like VAR while reducing the training cost by 95\%. VQ-Transplant democratizes quantization research by enabling resource-efficient integration of novel VQ techniques while matching industry-level reconstruction performance.

cs.CV↗
arXiv2026

Algorithms for Adaptive Experiments that Trade-off Statistical Analysis with Reward: Combining Uniform Random Assignment and Reward Maximization

Traditional randomized A/B experiments assign arms with uniform random (UR) probability, such as 50/50 assignment to two versions of a website to discover whether one version engages users more. To more quickly and automatically use data to benefit users, multi-armed bandit algorithms such as Thompson Sampling (TS) have been advocated. While TS is interpretable and incorporates the randomization key to statistical inference, it can cause biased estimates and increase false positives and false negatives in detecting differences in arm means. We introduce a more Statistically Sensitive algorithm, TS-PostDiff (Posterior Probability of Small Difference), that mixes TS with traditional UR by using an additional adaptive step, where the probability of using UR (vs TS) is proportional to the posterior probability that the difference in arms is small. This allows an experimenter to define what counts as a small difference, below which a traditional UR experiment can obtain informative data for statistical inference at low cost, and above which using more TS to maximize user benefits is key. We evaluate TS-PostDiff against UR, TS, and two other TS variants designed to improve statistical inference. We consider results for the common two-armed experiment across a range of settings inspired by real-world applications. Our results provide insight into when and why TS-PostDiff or alternative approaches provide better tradeoffs between benefiting users (reward) and statistical inference (false positive rate and power). TS-PostDiff's adaptivity helps efficiently reduce false positives and increase statistical power when differences are small, while increasing reward more when differences are large. The work highlights important considerations for future Statistically Sensitive algorithm development that balances reward and statistical analysis in adaptive experimentation.

cs.LG↗
arXiv2026

RIE-Greedy: Regularization-Induced Exploration for Contextual Bandits

Real-world contextual bandit problems with complex reward models are often tackled with iteratively trained models, such as boosting trees. However, it is difficult to directly apply simple and effective exploration strategies--such as Thompson Sampling or UCB--on top of those black-box estimators. Existing approaches rely on sophisticated assumptions or intractable procedures that are hard to verify and implement in practice. In this work, we explore the use of an exploration-free (pure-greedy) action selection strategy, that exploits the randomness inherent in model fitting process as an intrinsic source of exploration. More specifically, we note that the stochasticity in cross-validation based regularization process can naturally induce Thompson Sampling-like exploration. We show that this regularization-induced exploration is theoretically equivalent to Thompson Sampling in the two-armed bandit case and empirically leads to reliable exploration in large-scale business environments compared to benchmark methods such as epsilon-greedy and other state-of-the-art approaches. Overall, our work reveals how regularized estimator training itself can induce effective exploration, offering both theoretical insight and practical guidance for contextual bandit design.

stat.ML↗
arXiv2026

WebFactory: Automated Compression of Foundational Language Intelligence into Grounded Web Agents

Current paradigms for training GUI agents are fundamentally limited by a reliance on either unsafe, non-reproducible live web interactions or costly, scarce human-crafted data and environments. We argue this focus on data volume overlooks a more critical factor: the efficiency of compressing a large language model's (LLM) latent knowledge into actionable agent behavior. We introduce WebFactory, a novel, fully automated closed-loop reinforcement learning pipeline for GUI agents, systematically compressing LLM-encoded internet intelligence into efficient, grounded actions. Our pipeline features a process of scalable environment synthesis, knowledge-aware task generation, LLM-powered trajectory collection, decomposed reward RL training, and systematic agent evaluation. Remarkably, our agent demonstrates exceptional data efficiency and generalization. Trained on synthetic data from only 10 websites within WebFactory, it achieves performance comparable to GUI agents trained on the same amount of human-annotated data from a much larger set of environments. This superior performance is consistent across our internal offline and online transfer benchmarks, where our agent also significantly outperforms the base foundation model. We further provide critical insights into the "embodiment potential" of different LLM foundations, offering a new axis for model evaluation. This work presents a scalable and cost-effective paradigm for transforming passive internet knowledge into active, grounded intelligence, marking a critical step towards general-purpose interactive agents.

cs.AI↗
arXiv2026

WebChain: A Large-Scale Human-Annotated Dataset of Real-World Web Interaction Traces

We introduce WebChain, the largest open-source dataset of human-annotated trajectories on real-world websites, designed to accelerate reproducible research in web agents. It contains 31,725 trajectories and 318k steps, featuring a core Triple Alignment of visual, structural, and action data to provide rich, multi-modal supervision. The data is collected via a scalable pipeline that ensures coverage of complex, high-value tasks often missed by synthetic methods. Leveraging this dataset, we propose a Dual Mid-Training recipe that decouples spatial grounding from planning, achieving state-of-the-art performance on our proposed WebChainBench and other public GUI benchmarks. Our work provides the data and insights necessary to build and rigorously evaluate the next generation of scalable web agents.

cs.AI↗
arXiv2026

Modeling Cell Developmental Trajectory using Multinomial Unbalanced Optimal Transport

Single-cell trajectory analysis aims to reconstruct the biological developmental processes of cells as they evolve over time, leveraging temporal correlations in gene expression. During cellular development, gene expression patterns typically change and vary across different cell types. A significant challenge in this analysis is that RNA sequencing destroys the cell, making it impossible to track gene expression across multiple stages for the same cell. Recent advances have introduced the use of optimal transport tools to model the trajectory of individual cells. In this paper, our focus shifts to a question of greater practical importance: we examine the differentiation of cell types over time. Specifically, we propose a novel method based on discrete unbalanced optimal transport to model the developmental trajectory of cell types. Our method detects biological changes in cell types and infers their transitions to different states by analyzing the transport matrix. We validated our method using single-cell RNA sequencing data from mouse embryonic fibroblasts. The results accurately identified major developmental changes in cell types, which were corroborated by experimental evidence. Furthermore, the inferred transition probabilities between cell types are highly congruent to biological ground truth.

stat.AP↗
arXiv2026

A Statistically Reliable Optimization Framework for Bandit Experiments in Scientific Discovery

Scientific experimentation is largely driven by statistical hypothesis testing to determine significant differences in interventions. Traditionally, experimenters allocate samples uniformly between each intervention. However, such an approach may lead to suboptimal outcomes - multi-armed bandits (MABs) addresses this problem by allocating samples adaptively to maximize outcomes. Yet, two challenges have hindered the use of MABs in scientific domains. First, common hypothesis tests (e.g., $t$-tests) become invalid under adaptive sampling without correction, leading to inflated type~I and type~II errors. This is an understudied problem, and prior solutions suffer from issues such as low statistical power which prevent adoption in many practical settings. Second, practitioners must explicitly balance cumulative reward with statistical efficiency, yet no general methodology exists to quantify this trade-off across algorithms. In this paper, we study assumption modification and critical region correction approaches for hypothesis testing that enable common tests to be applied to adaptively collected data. We provide heuristic justification for its power efficiency and show in simulation that it achieves higher power than existing approaches. Further, we derive a theoretically and practically motivated objective function for adaptive experiment evaluation, which we integrate into a unified experimental framework. Our framework asks experimenters to specify an experiment extension cost for their problem, and based on that enables our proposed optimization procedure to select the bandit algorithm that best balances reward and power in their setting. We show that our approach enables practitioners to improve outcomes with only slightly more steps than uniform randomization, while retaining statistical validity.

stat.AP↗
arXiv2026

Successive classification learning for estimating quantile optimal treatment regimes

Quantile optimal treatment regimes (OTRs) aim to assign treatments that maximize a specified quantile of patients' outcomes. Compared to treatment regimes that target the mean outcomes, quantile OTRs offer fairer regimes when a lower quantile is selected, as it improves outcomes for vulnerable patients. In this paper, we propose a novel method for estimating quantile OTRs by reformulating the problem as a successive classification task, solvable via training a sequence of classifiers, each successive classifier built on the output of its predecessors. This reformulation enables us to leverage the powerful machine learning technique to enhance computational efficiency and handle complex decision boundaries. We also investigate the estimation of quantile OTRs when outcomes are discrete, a setting that has received limited attention in the literature. A key challenge is that direct extensions of existing methods to discrete outcomes often lead to inconsistency and ineffectiveness issues. To overcome this, we introduce a smoothing technique that maps discrete outcomes to continuous surrogates, enabling consistent and effective estimation. We provide theoretical guarantees to support our methodology, and demonstrate its superior performance through comprehensive simulation studies and real-data analysis.

stat.ME↗
arXiv2025

VenusBench-GD: A Comprehensive Multi-Platform GUI Benchmark for Diverse Grounding Tasks

GUI grounding is a critical component in building capable GUI agents. However, existing grounding benchmarks suffer from significant limitations: they either provide insufficient data volume and narrow domain coverage, or focus excessively on a single platform and require highly specialized domain knowledge. In this work, we present VenusBench-GD, a comprehensive, bilingual benchmark for GUI grounding that spans multiple platforms, enabling hierarchical evaluation for real-word applications. VenusBench-GD contributes as follows: (i) we introduce a large-scale, cross-platform benchmark with extensive coverage of applications, diverse UI elements, and rich annotated data, (ii) we establish a high-quality data construction pipeline for grounding tasks, achieving higher annotation accuracy than existing benchmarks, and (iii) we extend the scope of element grounding by proposing a hierarchical task taxonomy that divides grounding into basic and advanced categories, encompassing six distinct subtasks designed to evaluate models from complementary perspectives. Our experimental findings reveal critical insights: general-purpose multimodal models now match or even surpass specialized GUI models on basic grounding tasks. In contrast, advanced tasks, still favor GUI-specialized models, though they exhibit significant overfitting and poor robustness. These results underscore the necessity of comprehensive, multi-tiered evaluation frameworks.

cs.CV↗