Although generative models have made remarkable progress in recent years, their use in critical applications has been hindered by an inability to reliably evaluate the quality of their generated samples. Quality refers to at least two complementary concepts: fidelity and coverage. Current quality metrics often lack reliable, interpretable values due to an absence of calibration or insufficient robustness to outliers. To address these shortcomings, we introduce two novel metrics: Clipped Density and Clipped Coverage. By clipping individual sample contributions, as well as the radii of nearest neighbor balls for fidelity, our metrics prevent out-of-distribution samples from biasing the aggregated values. Through analytical and empirical calibration, these metrics demonstrate linear score degradation as the proportion of bad samples increases. Thus, they can be straightforwardly interpreted as equivalent proportions of good samples. Extensive experiments on synthetic and real-world datasets demonstrate that Clipped Density and Clipped Coverage outperform existing methods in terms of robustness, sensitivity, and interpretability when evaluating generative models.

Synthetic data generation using large language models (LLMs) demonstrates substantial promise in addressing biomedical data challenges and shows increasing adoption in biomedical research. This study systematically reviews recent advances in synthetic data generation for biomedical applications and clinical research, focusing on how LLMs address data scarcity, utility, and quality issues with different modalities. We conducted a scoping review following PRISMA-ScR guidelines and searched literature published between 2020 and 2025 through PubMed, ACM, Web of Science, and Google Scholar. A total of 59 studies were included based on relevance to synthetic data generation in biomedical contexts. Among the reviewed studies, the predominant data modalities were unstructured texts (78.0\%), tabular data (13.6\%), and multimodal sources (8.4\%). Common generation methods included LLM prompting (74.6\%), fine-tuning (20.3\%), and specialized models (5.1\%). Evaluations were heterogeneous: intrinsic metrics (27.1\%), human-in-the-loop assessments (44.1\%), and LLM-based evaluations (13.6\%). However, limitations and key barriers persist in data modalities, domain utility, resource and model accessibility, and standardized evaluation protocols. Future efforts may focus on developing standardized, transparent evaluation frameworks and expanding accessibility to support effective applications in biomedical research.

A public safety Uncrewed Aerial Vehicle (UAV) enhances situational awareness during emergency response. Its agility, mobility optimization, and ability to establish Line-of-Sight (LoS) communication make it increasingly important for managing emergencies such as disaster response, search and rescue, and wildfire monitoring. Although Deep Reinforcement Learning (DRL) has been used to optimize UAV navigation and control, its high training complexity, low sample efficiency, and the simulation-to-reality gap limit its practicality in public safety applications. Recent advances in Large Language Models (LLMs) present a promising alternative. With strong reasoning and generalization abilities, LLMs can adapt to new tasks through In-Context Learning (ICL), enabling task adaptation via natural language prompts and example-based guidance without retraining. Deploying LLMs at the network edge, rather than in the cloud, further reduces latency and preserves data privacy, making them suitable for real-time, mission-critical public safety UAVs. This paper proposes integrating LLM-assisted ICL with public safety UAVs to address key functions such as path planning and velocity control in emergency response. We present a case study on data collection scheduling, demonstrating that the LLM-assisted ICL framework can significantly reduce packet loss compared to conventional approaches while also mitigating potential jailbreaking vulnerabilities. Finally, we discuss LLM optimizers and outline future research directions. The ICL framework enables adaptive, context-aware decision-making for public safety UAVs, offering a lightweight and efficient solution to enhance UAV autonomy and responsiveness in emergencies.

Computer-Aided Design (CAD) plays a central role in engineering and manufacturing, making it possible to create precise and editable 3D models. Using a variety of sensor or user-provided data as inputs for CAD reconstruction can democratize access to design applications. However, existing methods typically focus on a single input modality, such as point clouds, images, or text, which limits their generalizability and robustness. Leveraging recent advances in vision-language models (VLM), we propose a multi-modal CAD reconstruction model that simultaneously processes all three input modalities. Inspired by large language model (LLM) training paradigms, we adopt a two-stage pipeline: supervised fine-tuning (SFT) on large-scale procedurally generated data, followed by reinforcement learning (RL) fine-tuning using online feedback, obtained programatically. Furthermore, we are the first to explore RL fine-tuning of LLMs for CAD tasks demonstrating that online RL algorithms such as Group Relative Preference Optimization (GRPO) outperform offline alternatives. In the DeepCAD benchmark, our SFT model outperforms existing single-modal approaches in all three input modalities simultaneously. More importantly, after RL fine-tuning, cadrille sets new state-of-the-art on three challenging datasets, including a real-world one. Code is avaliable at https://github.com/col14m/cadrille .

Masked generative models (MGMs) can generate tokens in parallel and in any order, unlike autoregressive models (ARMs), which decode one token at a time, left-to-right. However, MGMs process the full-length sequence at every sampling step, including mask tokens that carry no information. In contrast, ARMs process only the previously generated tokens. We introduce ``Partition Generative Models'' (PGMs), which replace masking with partitioning. Tokens are split into two groups that cannot attend to each other, and the model learns to predict each group conditioned on the other, eliminating mask tokens entirely. Because the groups do not interact, PGMs can process only the clean tokens during sampling, like ARMs, while retaining parallel, any-order generation, like MGMs. On OpenWebText, PGMs achieve $5-5.5\times$ higher throughput than MDLM while producing samples with lower Generative Perplexity. On ImageNet, PGMs reach comparable FID to MaskGIT with a $7.5\times$ throughput improvement. With twice as many steps, the FID improves to 4.56 while remaining $3.9\times$ faster than MGMs. Finally, PGMs remain compatible with existing MGM samplers and distillation methods.

Chain-of-Thought (CoT) reasoning has advanced the capabilities and transparency of language models (LMs); however, reasoning chains can contain inaccurate statements that reduce performance and trustworthiness. To address this, we propose to augment each reasoning step in a CoT with a latent veracity (or correctness) variable. To efficiently explore this expanded space, we introduce Veracity Search (VS), a discrete search algorithm over veracity assignments. It performs otherwise intractable inference in the posterior distribution over latent veracity values by leveraging the LM's joint likelihood over veracity and the final answer as a proxy reward. This efficient inference-time verification method facilitates supervised fine-tuning of an Amortized Veracity Inference (AVI) machine by providing pseudo-labels for veracity. AVI generalizes VS, enabling accurate zero-shot veracity inference in novel contexts. Empirical results demonstrate that VS reliably identifies errors in logical (ProntoQA), mathematical (GSM8K), and commonsense (CommonsenseQA) reasoning benchmarks, with AVI achieving comparable zero-shot accuracy. Finally, we demonstrate the utility of latent veracity inference for providing feedback during self-correction and self-improvement.

Climate-smart and biodiversity-preserving forestry demands precise information on forest resources, extending to the individual tree level. Multispectral airborne laser scanning (ALS) has shown promise in automated point cloud processing, but challenges remain in leveraging deep learning techniques and identifying rare tree species in class-imbalanced datasets. This study addresses these gaps by conducting a comprehensive benchmark of deep learning and traditional shallow machine learning methods for tree species classification. For the study, we collected high-density multispectral ALS data ($>1000$ $\mathrm{pts}/\mathrm{m}^2$) at three wavelengths using the FGI-developed HeliALS system, complemented by existing Optech Titan data (35 $\mathrm{pts}/\mathrm{m}^2$), to evaluate the species classification accuracy of various algorithms in a peri-urban study area located in southern Finland. We established a field reference dataset of 6326 segments across nine species using a newly developed browser-based crowdsourcing tool, which facilitated efficient data annotation. The ALS data, including a training dataset of 1065 segments, was shared with the scientific community to foster collaborative research and diverse algorithmic contributions. Based on 5261 test segments, our findings demonstrate that point-based deep learning methods, particularly a point transformer model, outperformed traditional machine learning and image-based deep learning approaches on high-density multispectral point clouds. For the high-density ALS dataset, a point transformer model provided the best performance reaching an overall (macro-average) accuracy of 87.9% (74.5%) with a training set of 1065 segments and 92.0% (85.1%) with a larger training set of 5000 segments.

Neural networks are widely adopted to solve complex and challenging tasks. Especially in high-stakes decision-making, understanding their reasoning process is crucial, yet proves challenging for modern deep networks. Feature visualization (FV) is a powerful tool to decode what information neurons are responding to and hence to better understand the reasoning behind such networks. In particular, in FV we generate human-understandable images that reflect the information detected by neurons of interest. However, current methods often yield unrecognizable visualizations, exhibiting repetitive patterns and visual artifacts that are hard to understand for a human. To address these problems, we propose to guide FV through statistics of real image features combined with measures of relevant network flow to generate prototypical images. Our approach yields human-understandable visualizations that both qualitatively and quantitatively improve over state-of-the-art FVs across various architectures. As such, it can be used to decode which information the network uses, complementing mechanistic circuits that identify where it is encoded. Code is available at: https://github.com/adagorgun/VITAL

We present an ngram model-based logit scaling technique that effectively transfers extreme subword stylistic variation to large language models at inference time. We demonstrate its efficacy by tracking the perplexity of generated text with respect to the ngram interpolated and original versions of an evaluation model. Minimizing the former measure while the latter approaches the perplexity of a text produced by a target author or character lets us select a sufficient degree of adaptation while retaining fluency.

Bandit optimization usually refers to the class of online optimization problems with limited feedback, namely, a decision maker uses only the objective value at the current point to make a new decision and does not have access to the gradient of the objective function. While this name accurately captures the limitation in feedback, it is somehow misleading since it does not have any connection with the multi-armed bandits (MAB) problem class. We propose two new classes of problems: the functional multi-armed bandit problem (FMAB) and the best function identification problem. They are modifications of a multi-armed bandit problem and the best arm identification problem, respectively, where each arm represents an unknown black-box function. These problem classes are a surprisingly good fit for modeling real-world problems such as competitive LLM training. To solve the problems from these classes, we propose a new reduction scheme to construct UCB-type algorithms, namely, the F-LCB algorithm, based on algorithms for nonlinear optimization with known convergence rates. We provide the regret upper bounds for this reduction scheme based on the base algorithms' convergence rates. We add numerical experiments that demonstrate the performance of the proposed scheme.

This work presents SSL4EO-S12 v1.1, a multimodal, multitemporal Earth Observation dataset designed for pretraining large-scale foundation models. Building on the success of SSL4EO-S12, this extension updates the previous version to fix geospatial alignment inaccuracies and the inefficent data structure. The dataset allows low-barrier, analysis-ready data loading while maintaining the predecessor's spatial coverage of the world's 10,000 largest cities and surrounding geographies, resulting in 246k time series with nearly one million image patches. We package each time series in Zarr file format stored in WebDataset tar shards for efficient data loading and representation of meta-information such as cloud masks. We add new modalities for elevation, land-cover, and vegetation to support multimodal pre-training. Released under the CC-BY-4.0 license, SSL4EO-S12 v1.1 facilitates open research and provides a robust foundation for future advancements in self-supervised learning and geospatial analysis. The dataset is available online through https://huggingface.co/datasets/embed2scale/SSL4EO-S12-v1.1.

Despite its real-world significance, model performance on tabular data remains underexplored, leaving uncertainty about which model to rely on and which prompt configuration to adopt. To address this gap, we create ToRR, a benchmark for Table Reasoning and Robustness, measuring model performance and robustness on table-related tasks. The benchmark includes 10 datasets that cover different types of table reasoning capabilities across varied domains. ToRR goes beyond model performance rankings, and is designed to reflect whether models can handle tabular data consistently and robustly, across a variety of common table representation formats. We present a leaderboard as well as comprehensive analyses of the results of leading models over ToRR. Our results reveal a striking pattern of brittle model behavior, where even strong models are unable to perform robustly on tabular data tasks. Although no specific table format leads to consistently better performance, we show that testing over multiple formats is crucial for reliably estimating model capabilities. Moreover, we show that the reliability boost from testing multiple prompts can be equivalent to adding more test examples. Overall, our findings show that table understanding and reasoning tasks remain a significant challenge.

Time series anomaly detection (TSAD) has been a long-standing pillar problem in Web-scale systems and online infrastructures, such as service reliability monitoring, system fault diagnosis, and performance optimization. Large language models (LLMs) have demonstrated unprecedented capabilities in time series analysis, the potential of multimodal LLMs (MLLMs), particularly vision-language models, in TSAD remains largely under-explored. One natural way for humans to detect time series anomalies is through visualization and textual description. It motivates our research question: Can multimodal LLMs perform time series anomaly detection? Existing studies often oversimplify the problem by treating point-wise anomalies as special cases of range-wise ones or by aggregating point anomalies to approximate range-wise scenarios. They limit our understanding for realistic scenarios such as multi-granular anomalies and irregular time series. To address the gap, we build a VisualTimeAnomaly benchmark to comprehensively investigate zero-shot capabilities of MLLMs for TSAD, progressively from point-, range-, to variate-wise anomalies, and extends to irregular sampling conditions. Our study reveals several key insights in multimodal MLLMs for TSAD. Built on these findings, we propose a MLLMs-based multi-agent framework TSAD-Agents to achieve automatic TSAD. Our framework comprises scanning, planning, detection, and checking agents that synergistically collaborate to reason, plan, and self-reflect to enable automatic TSAD. These agents adaptively invoke tools such as traditional methods and MLLMs and dynamically switch between text and image modalities to optimize detection performance.

We develop a novel method for personalized off-policy learning in scenarios with unobserved confounding. Thereby, we address a key limitation of standard policy learning: standard policy learning assumes unconfoundedness, meaning that no unobserved factors influence both treatment assignment and outcomes. However, this assumption is often violated, because of which standard policy learning produces biased estimates and thus leads to policies that can be harmful. To address this limitation, we employ causal sensitivity analysis and derive a semi-parametrically efficient estimator for a sharp bound on the value function under unobserved confounding. Our estimator has three advantages: (1) Unlike existing works, our estimator avoids unstable minimax optimization based on inverse propensity weighted outcomes. (2) Our estimator is semi-parametrically efficient. (3) We prove that our estimator leads to the optimal confounding-robust policy. Finally, we extend our theory to the related task of policy improvement under unobserved confounding, i.e., when a baseline policy such as the standard of care is available. We show in experiments with synthetic and real-world data that our method outperforms simple plug-in approaches and existing baselines. Our method is highly relevant for decision-making where unobserved confounding can be problematic, such as in healthcare and public policy.

Although adversarial robustness has been extensively studied in white-box settings, recent advances in black-box attacks (including transfer- and query-based approaches) are primarily benchmarked against weak defenses, leaving a significant gap in the evaluation of their effectiveness against more recent and moderate robust models (e.g., those featured in the Robustbench leaderboard). In this paper, we question this lack of attention from black-box attacks to robust models. We establish a framework to evaluate the effectiveness of recent black-box attacks against both top-performing and standard defense mechanisms, on the ImageNet dataset. Our empirical evaluation reveals the following key findings: (1) the most advanced black-box attacks struggle to succeed even against simple adversarially trained models; (2) robust models that are optimized to withstand strong white-box attacks, such as AutoAttack, also exhibits enhanced resilience against black-box attacks; and (3) robustness alignment between the surrogate models and the target model plays a key factor in the success rate of transfer-based attacks

We present a dataset of 19th century American literary orthovariant tokens with a novel layer of human-annotated dialect group tags designed to serve as the basis for computational experiments exploring literarily meaningful orthographic variation. We perform an initial broad set of experiments over this dataset using both token (BERT) and character (CANINE)-level contextual language models. We find indications that the "dialect effect" produced by intentional orthographic variation employs multiple linguistic channels, and that these channels are able to be surfaced to varied degrees given particular language modelling assumptions. Specifically, we find evidence showing that choice of tokenization scheme meaningfully impact the type of orthographic information a model is able to surface.

We derive a policy gradient theorem for Cumulative Prospect Theory (CPT) objectives in finite-horizon Reinforcement Learning (RL), generalizing the standard policy gradient theorem and encompassing distortion-based risk objectives as special cases. Motivated by behavioral economics, CPT combines an asymmetric utility transformation around a reference point with probability distortion. Building on our theorem, we design a first-order policy gradient algorithm for CPT-RL using a Monte Carlo gradient estimator based on order statistics. We establish statistical guarantees for the estimator and prove asymptotic convergence of the resulting algorithm to first-order stationary points of the (generally non-convex) CPT objective. Simulations illustrate qualitative behaviors induced by CPT and compare our first-order approach to existing zeroth-order methods.

We propose an algorithm for nonparametric online change point detection based on sequential score function estimation and the tracking the best expert approach. The core of the procedure is a version of the fixed share forecaster tailored to the case of infinite number of experts and quadratic loss functions. The algorithm shows promising results in numerical experiments on artificial and real-world data sets. Its performance is supported by rigorous high-probability bounds describing behaviour of the test statistic in the pre-change and post-change regimes.

Multi-turn dialogues and context-intensive tasks challenge Large Language Models (LLMs) to integrate long histories without sacrificing generation quality. Although prefix LLMs can better exploit historical context via bidirectional attention on prefix tokens, they are rarely used in practice because multi-turn training requires many duplicated triplets, and its bidirectional prefix prevents KV-cache reuse at inference time, driving up high cost and latency. To retain the contextual understanding of prefix mask while preserving the inference-time efficiency of causal mask, we introduce Intermittent Semi-working Mask (ISM), a masking scheme that injects sparse bidirectional attention into the causal backbone. ISM alternates bidirectional attention over query segments with unidirectional attention over answer segments, enabling the synthesis of in-context while preserving global causality. This design eliminates triplet expansion during training and maintains KV-cache reuse during inference, yielding latency comparable to standard causal LLMs. ISM is architecture-agnostic and parameter-free, adding only minimal latency. Across extensive evaluations, ISM outperforms causal baselines not only on multi-turn dialogue, but also on context-intensive tasks like mathematical reasoning.

In surgical skill assessment, the Objective Structured Assessments of Technical Skills (OSATS) and Global Rating Scale (GRS) are well-established tools for evaluating surgeons during training. These metrics, along with performance feedback, help surgeons improve and reach practice standards. Recent research on the open-source JIGSAWS dataset, which includes both GRS and OSATS labels, has focused on regressing GRS scores from kinematic data, video, or their combination. However, we argue that regressing GRS alone is limiting, as it aggregates OSATS scores and overlooks clinically meaningful variations during a surgical trial. To address this, we developed a weakly-supervised recurrent transformer model that tracks a surgeon's performance throughout a session by mapping hidden states to six OSATS, derived from kinematic data. These OSATS scores are averaged to predict GRS, allowing us to compare our model's performance against state-of-the-art (SOTA) methods. We report Spearman's Correlation Coefficients (SCC) demonstrating that our model outperforms SOTA using kinematic data (SCC 0.83-0.88), and matches performance with video-based models. Our model also surpasses SOTA in most tasks for average OSATS predictions (SCC 0.46-0.70) and specific OSATS (SCC 0.56-0.95). The generation of pseudo-labels at the segment level translates quantitative predictions into qualitative feedback, vital for automated surgical skill assessment pipelines. A senior surgeon validated our model's outputs, agreeing with 77\% of the weakly-supervised predictions \(p=0.006\).

Estimating potential outcomes for treatments over time based on observational data is important for personalized decision-making in medicine. However, many existing methods for this task fail to properly adjust for time-varying confounding and thus yield biased estimates. There are only a few neural methods with proper adjustments, but these have inherent limitations (e.g., division by propensity scores that are often close to zero), which result in poor performance. As a remedy, we introduce the iterative G-computation network (IGC-Net). Our IGC-Net is a novel, neural end-to-end model which adjusts for time-varying confounding in order to estimate conditional average potential outcomes (CAPOs) over time. Specifically, our IGC-Net is the first neural model to perform fully regression-based iterative G-computation for CAPOs in the time-varying setting. We evaluate the effectiveness of our IGC-Net across various experiments. In sum, this work represents a significant step towards personalized decision-making from electronic health records.

We present a novel corpus consisting of orthographically variant words found in works of 19th century U.S. literature annotated with their corresponding "standard" word pair. We train a set of neural edit distance models to pair these variants with their standard forms, and compare the performance of these models to the performance of a set of neural edit distance models trained on a corpus of orthographic errors made by L2 English learners. Finally, we analyze the relative performance of these models in the light of different negative training sample generation strategies, and offer concluding remarks on the unique challenge literary orthographic variation poses to string pairing methodologies.

The generative adversarial network (GAN) is a well-known model for learning high-dimensional distributions, but the mechanism for its generalization ability is not understood. In particular, GAN is vulnerable to the memorization phenomenon, the eventual convergence to the empirical distribution. We consider a simplified GAN model with the generator replaced by a density, and analyze how the discriminator contributes to generalization. We show that with early stopping, the generalization error measured by Wasserstein metric escapes from the curse of dimensionality, despite that in the long term, memorization is inevitable. In addition, we present a hardness of learning result for WGAN.

Efficiently training control policies for robots is a major challenge that can greatly benefit from utilizing knowledge gained from training similar systems through cross-embodiment knowledge transfer. In this work, we focus on accelerating policy training using a library-based initialization scheme that enables effective knowledge transfer across multirotor configurations. By leveraging a physics-aware neural control architecture that combines a reinforcement learning-based controller and a supervised control allocation network, we enable the reuse of previously trained policies. To this end, we utilize a policy evaluation-based similarity measure that identifies suitable policies for initialization from a library. We demonstrate that this measure correlates with the reduction in environment interactions needed to reach target performance and is therefore suited for initialization. Extensive simulation and real-world experiments confirm that our control architecture achieves state-of-the-art control performance, and that our initialization scheme saves on average up to $73.5\%$ of environment interactions (compared to training a policy from scratch) across diverse quadrotor and hexarotor designs, paving the way for efficient cross-embodiment transfer in reinforcement learning.

The LLM community often reports benchmark results as if they are synonymous with general model capabilities. However, benchmarks can have problems that distort performance, like test set contamination and annotator error. How can we know that a benchmark is a reliable indicator of some capability that we want to measure? This question concerns the construct validity of LLM benchmarks, and it requires separating benchmark results from capabilities when we model and predict LLM performance. Both social scientists and computer scientists propose formal models - latent factor models and scaling laws - for identifying the capabilities underlying benchmark scores. However, neither technique is satisfactory for construct validity. Latent factor models ignore scaling laws, and as a result, the capabilities they extract often proxy model size. Scaling laws ignore measurement error, and as a result, the capabilities they extract are both uninterpretable and overfit to the observed benchmarks. This thesis presents the structured capabilities model, the first model to extract interpretable and generalisable capabilities from a large collection of LLM benchmark results. I fit this model and its two alternatives on a large sample of results from the OpenLLM Leaderboard. Structured capabilities outperform latent factor models on parsimonious fit indices, and exhibit better out-of-distribution benchmark prediction than scaling laws. These improvements are possible because neither existing approach separates model scale from capabilities in the appropriate way. Model scale should inform capabilities, as in scaling laws, and these capabilities should inform observed results up to measurement error, as in latent factor models. In combining these two insights, structured capabilities demonstrate better explanatory and predictive power for quantifying construct validity in LLM evaluations.

Mixture-of-Experts (MoE) effectively scales model capacity while preserving computational efficiency through sparse expert activation. However, training high-quality MoEs from scratch is prohibitively expensive. A promising alternative is to convert pretrained dense models into sparse MoEs. Existing dense-to-MoE methods fall into two categories: \textbf{dynamic structural pruning} that converts dense models into MoE architectures with moderate sparsity to balance performance and inference efficiency, and \textbf{downcycling} approaches that use pretrained dense models to initialize highly sparse MoE architectures. However, existing methods break the intrinsic activation patterns within dense models, leading to suboptimal expert construction. In this work, we argue that the Gated Linear Unit (GLU) mechanism provides a natural blueprint for dense-to-MoE conversion. We show that the fine-grained neural-wise activation patterns of GLU reveal a coarse-grained structure, uncovering an inherent MoE architecture composed of consistently activated universal neurons and dynamically activated specialized neurons. Leveraging this discovery, we introduce ExpertWeaver, a training-free framework that partitions neurons according to their activation patterns and constructs shared experts and specialized routed experts with layer-adaptive configurations. Our experiments demonstrate that ExpertWeaver significantly outperforms existing methods, both as a training-free dynamic structural pruning technique and as a downcycling strategy for superior MoE initialization.

Transient objects in casual multi-view captures cause ghosting artifacts in 3D Gaussian Splatting (3DGS) reconstruction. Existing solutions relied on scene decomposition at significant memory cost or on motion-based heuristics that were vulnerable to parallax ambiguity. A semantic filtering framework was proposed for category-aware transient removal using vision-language models. CLIP similarity scores between rendered views and distractor text prompts were accumulated per-Gaussian across training iterations. Gaussians exceeding a calibrated threshold underwent opacity regularization and periodic pruning. Unlike motion-based approaches, semantic classification resolved parallax ambiguity by identifying object categories independently of motion patterns. Experiments on the RobustNeRF benchmark demonstrated consistent improvement in reconstruction quality over vanilla 3DGS across four sequences, while maintaining minimal memory overhead and real-time rendering performance. Threshold calibration and comparisons with baselines validated semantic guidance as a practical strategy for transient removal in scenarios with predictable distractor categories.

As large language models (LLMs) become increasingly prevalent, reliable methods for detecting AI-generated text are critical for mitigating potential risks. We introduce DependencyAI, a simple and interpretable approach for detecting AI-generated text using only the labels of linguistic dependency relations. Our method achieves competitive performance across monolingual, multi-generator, and multilingual settings. To increase interpretability, we analyze feature importance to reveal syntactic structures that distinguish AI-generated from human-written text. We also observe a systematic overprediction of certain models on unseen domains, suggesting that generator-specific writing styles may affect cross-domain generalization. Overall, our results demonstrate that dependency relations alone provide a robust signal for AI-generated text detection, establishing DependencyAI as a strong linguistically grounded, interpretable, and non-neural network baseline.

The tendency for hallucination in current large language models (LLMs) negatively impacts dialogue systems. Such hallucinations produce factually incorrect responses that may mislead users and undermine system trust. Existing refinement methods for dialogue systems typically operate at the response level, overlooking the fact that a single response may contain multiple verifiable or unverifiable facts. To address this gap, we propose Fine-Refine, a fine-grained refinement framework that decomposes responses into atomic units, verifies each unit using external knowledge, assesses fluency via perplexity, and iteratively corrects granular errors. We evaluate factuality across the HybriDialogue and OpendialKG datasets in terms of factual accuracy (fact score) and coverage (Not Enough Information Proportion), and experiments show that Fine-Refine substantially improves factuality, achieving up to a 7.63-point gain in dialogue fact score, with a small trade-off in dialogue quality.

We introduce LuxMT, a machine translation system based on Gemma 3 27B and fine-tuned for translation from Luxembourgish (LB) into French (FR) and English (EN). To assess translation performance, we construct a novel benchmark covering LB-FR, LB-EN, and LB-FR using human-translated data from Luci, a tourist magazine about Luxembourg. Training data stems from LuxAlign, a parallel corpus of multilingual Luxembourgish news articles, and LB parliamentary transcripts augmented with Google Translate. We filter the data using LuxEmbedder, LB sentence embeddings, to remove low-equivalence segment-pairs. Overall, LuxMT's results suggest strong improvements over the Gemma 3 baseline, even for translating LB to German (DE), despite the training data not containing any DE. We also explore LuxEmbedder's potential to be used as a quality estimation metric and find strong correlations with other reference-based metrics. However, we call for further research to fully assess the metric's utility and advise using it with caution.