Document understanding (VRDU) in regulated domains is particularly challenging, since scanned documents often contain sensitive, evolving, and domain specific knowledge. This leads to two major challenges: the lack of manual annotations for model adaptation and the difficulty for pretrained models to stay up-to-date with domain-specific facts. While Multimodal Large Language Models (MLLMs) show strong zero-shot abilities, they still suffer from hallucination and limited domain grounding. In contrast, discriminative Vision-Language Pre-trained Models (VLPMs) provide reliable grounding but require costly annotations to cover new domains. We introduce Docs2Synth, a synthetic-supervision framework that enables retrieval-guided inference for private and low-resource domains. Docs2Synth automatically processes raw document collections, generates and verifies diverse QA pairs via an agent-based system, and trains a lightweight visual retriever to extract domain-relevant evidence. During inference, the retriever collaborates with an MLLM through an iterative retrieval–generation loop, reducing hallucination and improving response consistency. We further deliver Docs2Synth as an easy-to-use Python package, enabling plug-and-play deployment across diverse real-world scenarios. Experiments on multiple VRDU benchmarks show that Docs2Synth substantially enhances grounding and domain generalization without requiring human annotations.

Multi-page visually-rich document understanding (MP-VRDU) requires managing evidence that is sparse, spread across pages, and often exceeds a model's context window. Prior work has produced competing, largely untested claims about how these systems should be built. We attribute incorrect answers to three failure modes, representation, selection, and reasoning, and isolate each over a multi-page document understanding dataset by intervening on one while holding the others fixed. We find that vision is necessary but does not replace text extraction, that missing pages bound accuracy while distractors cost little, and that reasoners fail to integrate evidence across pages even when it is fully supplied. Prompting can shift reasoning behaviour substantially, improving some outcomes at the expense of others. We translate these findings into guidance for building such systems under a fixed compute budget.

LLM-based text-to-SQL pipelines read the database schema as text, which carries both structural cues (tables, keys, relationships) and semantic cues (table and column names); prior work has studied each axis in isolation, leaving open how they compare in magnitude and whether they substitute for one another. We present a controlled 6 times 3 factorial design crossing structural levels L_1–L_6 (from a denormalised wide table to a 3NF schema with foreign keys and explicit join paths) with semantic levels S_1–S_3 (anonymous, abbreviated, descriptive identifiers), evaluated on 397 corrected BIRD questions with identical gold queries throughout; we materialise 1NF and 2NF variants for nine BIRD databases to support the lowest structural levels. Across nine models from 0.5B to flagship scale we find an asymmetric substitution between the two axes, meaningful names compensate for missing structure but richer structural metadata does not recover performance when names are opaque, which reproduces in 8 of 9 databases and emerges with model scale (negligible below 3B). Within the structural axis the dominant lever is normalisation itself, not metadata layered on top of 3NF, suggesting that for current LLM-based text-to-SQL the practical bottleneck is semantic grounding rather than relational exposure.

3D human reaction generation faces three main challenges: (1) high motion fidelity, (2) real-time inference, and (3) autoregressive adaptability for online scenarios. Existing methods fail to meet all three simultaneously. We propose ARMFlow, a MeanFlow-based autoregressive framework that models temporal dependencies between actor and reactor motions. It consists of a causal context encoder and an MLP-based velocity predictor. We introduce Bootstrap Contextual Encoding (BSCE) in training, encoding generated history instead of the ground-truth ones, to alleviate error accumulation in autoregressive generation. We further introduce the offline variant {ReMFlow}, achieving state-of-the-art performance with the fastest inference among offline methods. Our ARMFlow addresses key limitations of online settings by: (1) enhancing semantic alignment via a global contextual encoder; (2) achieving high accuracy and low latency in a single-step inference; and (3) reducing accumulated errors through BSCE. Our single-step online generation surpasses existing online methods on InterHuman and InterX by over 40% in FID, while matching offline state-of-the-art performance despite using only partial sequence conditions. Code is available in the supplementary.

Answering complex logical queries is a fundamental task in knowledge graph reasoning. Query representation learning models project queries and entities into embedding vectors in low dimensional spaces, commonly referred to as query embeddings (QE). This approach addresses the challenges of complex logical queries on incomplete large knowledge graphs and demands a comprehensive survey. This article presents a comprehensive survey of QE methods according to query syntaxes, representation learning methods, optimization methods, datasets, evaluation metrics and model performance. We propose a taxonomy for existing QE methods and investigate issues in the representation learning of queries within and across methods. Finally, the article concludes with challenges and an outlook of future directions in the field.

Geochemical anomaly detection plays a critical role in mineral exploration as deviations from regional geochemical baselines may indicate mineralization. Existing studies suffer from two key limitations: (1) single region scenarios which limit model generalizability; (2) proprietary datasets, which makes result reproduction unattainable. In this work, we introduce GeoChemAD, an open-source benchmark dataset compiled from government-led geological surveys, covering multiple regions, sampling sources, and target elements. The dataset comprises eight subsets representing diverse spatial scales and sampling conditions. To establish strong baselines, we reproduce and benchmark a range of unsupervised anomaly detection methods, including statistical models, generative and transformer-based approaches. Furthermore, we propose GeoChemFormer, a transformer-based framework that leverages self-supervised pretraining to learn target-element-aware geochemical representations for spatial samples. Extensive experiments demonstrate that GeoChemFormer consistently achieves superior and robust performance across all eight subsets, outperforming existing unsupervised methods in both anomaly detection accuracy and generalization capability. The proposed dataset and framework provide a foundation for reproducible research and future development in this direction.

Generating realistic 3D Human-Human Interaction (HHI) requires coherent modeling of the physical plausibility of the agents and their interaction semantics. Existing methods compress all motion information into a single latent representation, limiting their ability to capture fine-grained actions and inter-agent interactions. This often leads to semantic misalignment and physically implausible artifacts, such as penetration or missed contact. We propose Disentangled Hierarchical Variational Autoencoder (DHVAE) based latent diffusion for structured and controllable HHI generation. DHVAE explicitly disentangles the global interaction context and individual motion patterns into a decoupled latent structure by employing a CoTransformer module. To mitigate implausible and physically inconsistent contacts in HHI, we incorporate contrastive learning constraints with our DHVAE to promote a more discriminative and physically plausible latent interaction space. For high-fidelity interaction synthesis, DHVAE employs a DDIM-based diffusion denoising process in the hierarchical latent space, enhanced by a skip-connected AdaLN-Transformer denoiser. Extensive evaluations show that DHVAE achieves superior motion fidelity, text alignment, and physical plausibility with greater computational efficiency.

Mineral exploration requires integrating heterogeneous geochemical, geophysical, and geological evidence, yet existing prospectivity systems often provide only opaque scores or heatmaps. We present MineTRACE, a web-based system for evidence-grounded exploration of eight commodities: Cu, Au, Ni, W, Sn, Co, Ta, and Mn. Users can explore prospectivity maps, query locations or regions, inspect supporting evidence, and interact through natural language. A transparent expert tree, informed by geological knowledge and known deposits, combines multi-source evidence into interpretable prospectivity scores. For a new location, the conversational assistant retrieves the score and supporting evidence from the analysis pipeline and presents them in natural language. The scorer achieves spatial AUC values of up to 0.917 across different test scenarios, while end-to-end evaluation assesses query accuracy and response grounding. MineTRACE makes public geoscience data easier to access, interpret, and verify, supporting more efficient and transparent mineral exploration.

Visually Rich Document Understanding (VRDU) has become a pivotal area of research, driven by the need to automatically interpret documents that contain intricate visual, textual, and structural elements. Recently, Multimodal Large Language Models (MLLMs) have demonstrated significant promise in this domain, including both OCR-based and OCR-free approaches for information extraction from document images. This survey reviews recent advances in MLLM-based VRDU, highlighting emerging trends and promising research directions with a focus on two key aspects: (1) techniques for representing and integrating textual, visual, and layout features; (2) training paradigms, including pretraining, instruction tuning, and training strategies. Moreover, we address challenges such as data scarcity, handling multi-page and multilingual documents, and integrating emerging trends such as Retrieval-Augmented Generation and agentic frameworks. Our analysis offers a roadmap for advancing MLLM-based VRDU toward more scalable, reliable, and adaptable systems.

Video generation models have been used as a robot policy to predict the future states of executing a task conditioned on task description and observation. Previous works ignore their high computational cost and long inference time. To address this challenge, we propose Draft-and-Target Sampling, a novel diffusion inference paradigm for video generation policy that is training-free and can improve inference efficiency. We introduce a self-play denoising approach by utilizing two complementary denoising trajectories in a single model, draft sampling takes large steps to generate a global trajectory in a fast manner and target sampling takes small steps to verify it. To further speedup generation, we introduce token chunking and progressive acceptance strategy to reduce redundant computation. Experiments on three benchmarks show that our method can achieve up to 2.1x speedup and improve the efficiency of current state-of-the-art methods with minimal compromise to the success rate. Our code is available.

Multi-modal document retrieval is a critical bottleneck for Retrieval-Augmented Generation (RAG), particularly for visually rich documents like financial reports and academic papers. Existing text-only retrievers fail to capture spatial and structural semantics, while vision-based approaches necessitate computationally expensive vision-language generation, shifting the bottleneck from offline indexing to online RAG inference. To address this, we propose LMS-Retrieval, a layout-aware, modality-aware, and structure-aware framework that augments a ColBERT backbone with fine-grained spatial feature fusion and optional structural path encodings. This approach enables effective multi-modal modeling while maintaining the efficiency of text-based inference. To stabilize training, we implement Conditional Ring Negative Sampling, which dynamically selects negatives within an optimal difficulty range to mitigate false negatives and prevent training instability. Experiments on the MMDocIR benchmark reveal a clear duality: the core Layout-Modality-aware variant achieves the highest Macro and Micro Average Recall across all baselines, demonstrating superior robustness and generalization. Meanwhile, the structural component acts as a specialized enhancement for complex, hierarchical documents, enabling the full LMS-Retrieval model to excel in rigid domains by achieving a Recall@1 of 56.8 on Laws and 48.4 on News, significantly outperforming both ColBERT and ColPali. These results establish LMS-Retrieval as a scalable solution that reconciles the precision of text with the structural awareness of vision. Our code and pre-trained models are publicly available on github( https://github.com/Lumanman9/LMS-Retrieval ).

Document parsing is a foundational step for document understanding tasks such as visual question answering and key information extraction, as it transforms unstructured scanned images into structured representations by extracting textual, visual, and layout information. While numerous Optical Character Recognition (OCR) engines and multimodal large language models (MLLMs) have been developed for this purpose, selecting an appropriate document parsing solution for a given document collection remains challenging, particularly in label-scarce settings. In this work, we conduct a systematic evaluation of text recognition performance across a diverse set of OCR engines and state-of-the-art MLLMs on multiple scanned document benchmarks spanning different domains and languages. Motivated by the limited contextual reasoning capabilities of many OCR engines and the high cost of manual annotations, we propose DocOCR-Eval, an annotation-free evaluation framework for automatic OCR assessment and selection. DocOCR-Eval employs a three-staged correction and ranking strategy to approximate annotation-based tool ordering without ground-truth labels. We show that aggregating across multiple MLLMs progressively improves alignment with annotation-based rankings. Extensive experiments further demonstrate that reliable OCR tool selection can be achieved in realistic, label-limited settings, providing practical guidance for deploying document parsing systems across diverse real-world document collections.

Extracting structured knowledge from unstructured data still faces practical limitations: entity and event extraction pipelines remain brittle, knowledge graph construction requires costly ontology engineering, and cross-domain generalization is rarely production-ready. In contrast, space and time provide universal contextual anchors that naturally align heterogeneous information and benefit downstream tasks such as retrieval and reasoning. We introduce STIndex, an end-to-end system that structures unstructured content into a multidimensional spatiotemporal data warehouse. Users define domain-specific analysis dimensions with configurable hierarchies, while large language models perform context-aware extraction and grounding. STIndex integrates document-level memory, geocoding correction, and quality validation, and offers an interactive analytics dashboard for visualization, clustering, burst detection, and entity network analysis. In evaluation on a public health benchmark, STIndex improves spatiotemporal entity extraction F1 by 4.37% (GPT-4o-mini) and 3.60% (Qwen3-8B). A live demonstration and open-source code are available at https://stindex.ai4wa.com/dashboard.

Causal reinforcement learning (RL) lacks benchmarks for complex systems that combine sequential decision making, hidden information, large masked action spaces, and explicit causal structure. We introduce MTG-Causal-RL, a Gymnasium benchmark built on Magic: The Gathering with a 3,077-dimensional partial observation, a 478-action masked discrete action space, five competitive Standard archetypes, three reward schemes, and a hand-specified Structural Causal Model (SCM) over strategic variables. Every episode exposes causal variables, SCM-predicted intervention effects, and per-factor credit traces, making causal credit assignment, leave-one-out cross-archetype transfer, and policy auditability first-class metrics. We adapt a panel of reference baselines: random, heuristic, masked PPO, a causal-world-model PPO variant, and an architecture-matched scalar control. We propose Causal Graph-Factored Advantage PPO (CGFA-PPO) as a reference causal agent that uses SCM parents of win probability as factor-aligned critic targets with an intervention-calibration loss. All comparisons use paired seeds, paired-bootstrap confidence intervals, and Holm-Bonferroni correction within pre-registered families. Masked PPO and CGFA-PPO reach competitive in-distribution win rates and exceed the random baseline; per-factor calibration trajectories and leave-one-out transfer gaps expose diagnostic structure that scalar win rate alone cannot. We release the benchmark, reference-baseline results, and full evaluation protocol openly. By coupling a strategically rich, partially observed domain with an explicit causal interface and statistical protocol, MTG-Causal-RL gives causal-RL, world-model, and LLM-agent research a shared testbed for questions current benchmarks cannot pose together: causal credit assignment under masked action spaces, structural transfer across archetypes, and SCM-grounded policy auditability.

With the advent of Large Language Models (LLMs), assessing their causal reasoning capabilities has attracted a rapid growth in research efforts. Much of the existing work, however, mirrors the evaluation of commonsense reasoning, which is a task with overly open-ended questions and lacks distinction from information retrieval. Determining whether LLMs inherently possess causal reasoning capabilities or merely reproduce patterns from training data is the key research question this research is focusing on. First, to produce suitable benchmark datasets, we leverage extensive prior work in causal inference through Directed Acyclic Graphs (DAGs) and develop a Causal Structure-Guided Causal Reasoning Dataset Generation Framework. This framework facilitates the generation of causal reasoning datasets from any existing causal graphs represented as DAGs. Second, a DAG-Think-Twice structured prompting technique is proposed, introducing two structured tags ( and ) in the prompts with explicit guidance on the roles of concepts and events played in a causal reasoning task (i.e. confounders, mediators, colliders). This has achieved a remarkable improvement of 40 https://github.com/22856226/Research .

Enterprises generate vast amounts of unstructured documents, posing challenges for knowledge extraction and representation. Large language models (LLMs) offer strong potential for processing such data but struggle with factual accuracy and provenance. Knowledge graphs (KGs) provide a structured framework to address these limitations [6], yet constructing high-quality KGs from heterogeneous data remains a challenge. To address this issue, we present Docs2KG, a modular framework to build high-quality KGs from diverse unstructured documents. We first employs state-of-the-art document processing techniques to extract textual content, tabular data, and figures. The extracted information is then unified into a multifaceted KG with three aspects: (1) a Layout KG capturing document structural hierarchies, (2) a Metadata KG preserving document properties, and (3) a Semantic KG representing domain-specific entities and relationships. Docs2KG supports multiple construction paradigms for Semantic KG: ontology-based approaches, hybrid NLP pipelines with LLM verification, LLM-guided ontology generation, and specialized models for named entity recognition, event extraction, and causal relationship identification to enhance semantic coverage and accuracy. A key feature of Docs2KG is its human-in-the-loop verification interface, enabling iterative quality assessment and refinement of the resulting KGs. Docs2KGis openly available at https://docs2kg.ai4wa.com, with the aim of advancing knowledge graph construction research and accelerating enterprise applications through high-quality knowledge graph construction.

Illegal, unreported, and unregulated (IUU) fishing causes global economic losses of 10-25 billion USD annually and undermines marine sustainability and governance. Synthetic Aperture Radar (SAR) provides reliable maritime surveillance under all weather and lighting conditions, but classifying small maritime objects in SAR imagery remains challenging. We investigate quantum machine learning for this task, focusing on quantum kernel methods (QKMs) applied to real and complex SAR chips extracted from the SARFish dataset. We tackle two binary classification problems, the first for distinguishing vessels from non-vessels, and the second for distinguishing fishing vessels from other types of vessels. We compare QKMs applied to real and complex SAR chips against classical Laplacian, RBF, and linear kernels applied to real SAR chips. We restrict the comparison to be between just kernel based models so that the comparison is as fair and meaningful as possible. Using noiseless numerical simulations of the quantum kernels, we find that with the real SAR chips, QKMs are capable of obtaining equal or better performance than the classical kernels in the best case. However, the specific quantum kernel used to encode the complex SAR data overfits and performs poorly. This work presents the first application of QKMs to maritime classification in SAR imagery and offers insight into the potential and current limitations of quantum-enhanced learning for maritime surveillance.

Quantum machine learning methods often rely on fixed, hand-crafted quantum encodings that may not capture optimal features for downstream tasks. In this work, we study the power of quantum autoencoders in learning data-driven quantum representations. We first theoretically demonstrate that the quantum autoencoder method is efficient in terms of sample complexity throughout the entire training process. Then we numerically train the quantum autoencoder on 3 million peptide sequences, and evaluate their effectiveness across multiple peptide classification problems including antihypertensive peptide prediction, blood-brain barrier-penetration, and cytotoxic activity detection. The learned representations were compared against Hamiltonian-evolved baselines using a quantum kernel with support vector machines. Results show that quantum autoencoder learned representations achieve accuracy improvements ranging from 0.4% to 8.1% over Hamiltonian baselines across seven datasets, demonstrating effective generalization to diverse downstream datasets with pre-training enabling effective transfer learning without task-specific fine-tuning. This work establishes that quantum autoencoder architectures can effectively learn from large-scale datasets (3 million samples) with compact parameterizations (∼900 parameters), demonstrating their viability for practical quantum applications.