End-to-end OCR increasingly relies on autoregressive sequence models, where the quadratic cost of Transformer attention limits efficient transcription of long, paragraph-level text. State-Space Models (SSMs) such as Mamba offer linear-time decoding and have recently been shown to match Transformer accuracy on printed historical lines, but their behavior as sequences grow from short lines to full paragraphs, and their generalization to handwriting, remain poorly understood. We study how a Mamba-based OCR recognizer scales from lines to paragraphs. We first conduct a systematic exploration of its four core hyperparameters (decoder depth, state dimension, expansion factor, and connector depth) on synthetic paragraphs from 100 to 1,000 characters, identifying the recurrent state dimension and the expansion factor as the dominant levers for long-sequence accuracy. We then compare the recognizer against a Transformer baseline trained under an identical protocol. On clean synthetic paragraphs, both models stay below 1

While state-of-the-art Handwritten Text Recognition (HTR) models perform well on standard benchmarks, they frequently struggle with writers exhibiting highly specific styles that are underrepresented in the training data. To handle unseen and atypical writers, writer adaptation techniques personalize HTR models to individual handwriting styles. Leading writer adaptation methods require either offline fine-tuning or parameter updates at inference time, both involving gradient computation and backpropagation, which increase computational costs and demand careful hyperparameter tuning. In this work, we propose a novel context-driven HTR framework (Code available at: github.com/TSResearch-hub/FewShotWriterAdaptation) inspired by multimodal in-context learning, enabling inference-time writer adaptation using only a few annotated lines from the target writer without any parameter updates. We further demonstrate the impact of context length, design a compact 8M-parameter CNN-Transformer that enables few-shot in-context adaptation, and show that combining context-driven and standard OCR training strategies leads to complementary improvements. Experiments on IAM and RIMES validate our approach with Character Error Rates of 3.92

In data-scarce scenarios, deep learning models often overfit to noise and irrelevant patterns, which limits their ability to generalize to unseen samples. To address these challenges in medical image segmentation, we introduce Diff-UMamba, a novel architecture that combines the UNet framework with the mamba mechanism to model long-range dependencies. At the heart of Diff-UMamba is a noise reduction module, which employs a signal differencing strategy to suppress noisy or irrelevant activations within the encoder. This encourages the model to filter out spurious features and enhance task-relevant representations, thereby improving its focus on clinically significant regions. As a result, the architecture achieves improved segmentation accuracy and robustness, particularly in low-data settings. Diff-UMamba is evaluated on multiple public datasets, including medical segmentation decathalon dataset (lung and pancreas) and AIIB23, demonstrating consistent performance gains of 1%-3% over baseline methods in various segmentation tasks. To further assess performance under limited data conditions, additional experiments are conducted on the BraTS-21 dataset by varying the proportion of available training samples. The approach is also validated on a small internal non-small cell lung cancer dataset for the segmentation of gross tumor volume in cone beam CT, where it achieves a 4%-5% improvement over baseline. The code is publicly available here.

End-to-end OCR for historical newspapers remains challenging, as models must handle long text sequences, degraded print quality, and complex layouts. While Transformer-based recognizers dominate current research, their quadratic complexity limits efficient paragraph-level transcription and large-scale deployment. We investigate linear-time State-Space Models (SSMs), specifically Mamba, as a scalable alternative to Transformer-based sequence modeling for OCR. We present to our knowledge, the first OCR architecture based on SSMs, combining a CNN visual encoder with bi-directional and autoregressive Mamba sequence modeling, and conduct a large-scale benchmark comparing SSMs with Transformer- and BiLSTM-based recognizers. Multiple decoding strategies (CTC, autoregressive, and non-autoregressive) are evaluated under identical training conditions alongside strong neural baselines (VAN, DAN, DANIEL) and widely used off-the-shelf OCR engines (PERO-OCR, Tesseract OCR, TrOCR, Gemini). Experiments on historical newspapers from the Bibliothèque nationale du Luxembourg, with newly released >99% verified gold-standard annotations, and cross-dataset tests on Fraktur and Antiqua lines, show that all neural models achieve low error rates ( ∼ 2 × vs 2.30 × growth at 1000 chars), reaching 6.07 × faster. We release code, trained models, and standardized evaluation protocols to enable reproducible research and guide practitioners in large-scale cultural heritage OCR available at https://github.com/MarcoPerson/ssm-ocr-benchmark .

Dynamic link prediction is a critical task in the analysis of evolving networks, with applications ranging from recommender systems to economic exchanges. However, the concept of the temporal receptive field, which refers to the temporal context that models use for making predictions, has been largely overlooked and insufficiently analyzed in existing research. In this study, we present a comprehensive analysis of the temporal receptive field in dynamic graph learning. By examining multiple datasets and models, we formalize the role of temporal receptive field and highlight their crucial influence on predictive accuracy. Our results demonstrate that appropriately chosen temporal receptive field can significantly enhance model performance, while for some models, overly large windows may introduce noise and reduce accuracy. We conduct extensive benchmarking to validate our findings, ensuring that all experiments are fully reproducible. Code is available at https://github.com/ykrmm/BenchmarkTW .

We propose a lightweight compressed-domain tracking model that operates directly on video streams, without requiring full RGB video decoding. Using motion vectors and transform coefficients from compressed data, our deep model propagates object bounding boxes across frames, achieving a computational speed-up of order up to 3.7 with only a slight 4

Understanding newspaper images remains a challenging task due to their complex, nested hierarchical structures and dense, heterogeneous layouts. In this paper, we explore two complementary approaches for newspaper structure understanding. First, we present a modular bottom-up pipeline that combines state-of-the-art open-source models: YOLO for layout detection, LayoutReader for reading order prediction, and a custom algorithm for article segmentation. This approach leverages existing robust components while maintaining flexibility and interpretability. Second, we introduce Tiramisu (Tiered Transformers for Hierarchical Structure Understanding), a novel end-to-end transformer-based architecture that explicitly models document hierarchy through an iterative tiered process. Tiramisu performs section and article separation, block localization, semantic categorization, and reading order prediction using highly parallelized attention mechanisms. Finally, we release Finlam La Liberté, a new dataset designed specifically for evaluating hierarchical information retrieval in historical newspapers. Experimental results demonstrate the effectiveness of both approaches in reconstructing complex newspaper hierarchies, with comparative analysis highlighting their respective strengths for scalable document digitization. The Tiramisu training code, including the synthetic newspaper generator, is available at https://git.litislab.fr/tiramisu/tiramisu-newspaper-articles-extractor.

Psoriasis is a common inflammatory skin disease with heterogeneous presentation. Up to 30

The primary objective of the AIIB 2023 competition is to evaluate the predictive significance of airway-related imaging biomarkers in determining the survival outcomes of patients with lung fibrosis.This study introduces a comprehensive three-stage approach. Initially, a segmentation network, namely nn-Unet, is employed to delineate the airway's structural boundaries. Subsequently, key features are extracted from the radiomic images centered around the trachea and an enclosing bounding box around the airway. This step is motivated by the potential presence of critical survival-related insights within the tracheal region as well as pertinent information encoded in the structure and dimensions of the airway. Lastly, radiomic features obtained from the segmented areas are integrated into an SVM classifier. We could obtain an overall-score of 0.8601 for the segmentation in Task 1 while 0.7346 for the classification in Task 2.

Diffusion models have emerged as highly effective techniques for inpainting, however, they remain constrained by slow sampling rates. While recent advances have enhanced generation quality, they have also increased sampling time, thereby limiting scalability in real-world applications. We investigate the generative sampling process of diffusion-based inpainting models and observe that these models make minimal use of the input condition during the initial sampling steps. As a result, the sampling trajectory deviates from the data manifold, requiring complex synchronization mechanisms to realign the generation process. To address this, we propose Time-aware Diffusion Paint (TD-Paint), a novel approach that adapts the diffusion process by modeling variable noise levels at the pixel level. This technique allows the model to efficiently use known pixel values from the start, guiding the generation process toward the target manifold. By embedding this information early in the diffusion process, TD-Paint significantly accelerates sampling without compromising image quality. Unlike conventional diffusion-based inpainting models, which require a dedicated architecture or an expensive generation loop, TD-Paint achieves faster sampling times without architectural modifications. Experimental results across three datasets show that TD-Paint outperforms state-of-the-art diffusion models while maintaining lower complexity.

We introduce Rosetta, a multimodal model that leverages Multimodal In-Context Learning (MICL) to classify sequences of novel script patterns in documents by leveraging minimal examples, thus eliminating the need for explicit retraining. To enhance contextual learning, we designed a dataset generation process that ensures varying degrees of contextual informativeness, improving the model’s adaptability in leveraging context across different scenarios. A key strength of our method is the use of a Context-Aware Tokenizer (CAT), which enables open-vocabulary classification. This allows the model to classify text and symbol patterns across an unlimited range of classes, extending its classification capabilities beyond the scope of its training alphabet of patterns. As a result, it unlocks applications such as the recognition of new alphabets and languages. Experiments on synthetic datasets demonstrate the potential of Rosetta to successfully classify Out-Of-Distribution visual patterns and diverse sets of alphabets and scripts, including but not limited to Chinese, Greek, Russian, French, Spanish, and Japanese. (The code from this study will soon be accessible as open-source on the GitHub repository: https://github.com/TSResearch-hub/Rosetta-ICL-classif

While significant advances have been made in understanding renal pathophysiology, less is known about the role of glycosphingolipid (GSL) metabolism in driving organ dysfunction. Here, we used a small molecule inhibitor of glucosylceramide synthase to modulate GSL levels in three mouse models of distinct renal pathologies: Alport syndrome (Col4a3 KO), polycystic kidney disease (Nek8jck), and steroid-resistant nephrotic syndrome (Nphs2 cKO). At the tissue level, we identified a core immune-enriched transcriptional signature that was shared across models and enriched in human polycystic kidney disease. Single nuclei analysis identified robust transcriptional changes across multiple kidney cell types, including epithelial and immune lineages. To further explore the role of GSL modulation in macrophage biology, we performed in vitro studies with homeostatic and inflammatory bone marrow-derived macrophages. Cumulatively, this study provides a comprehensive overview of renal dysfunction and the effect of GSL modulation on kidney-derived cells in the setting of renal dysfunction. This study identifies shared immune-related gene signatures across kidney cell types in three distinct mouse models of renal disease and examines the role of glycosphingolipid metabolism in renal dysfunction.

Automatic traffic surveillance usually relies on the estimation of traffic flow parameters through either dedicated sensors or the processing of road surveillance cameras. However, dedicated sensors are expensive to deploy and maintain. Moreover, available video processing algorithms usually require a complex multi-step pipeline, unsuited for large scale deployment. Herein, we address the problem of automatically estimating the flow rate (number of vehicles/unit of time) from surveillance cameras at low computation cost. To do so, we rely on end-to-end deep architectures applied to compressed MPEG4 part-2 video streams issued from road surveillance cameras. By leveraging the approximate flow representation induced by the compression, we heavily reduce the computation and memory requirements. We propose three end-to-end deep architectures using this coarse pixel flow representation as input. We also release two datasets, one based on synthetic videos and one collected on industrial tunnel cameras. By training the deep models on the newly introduced datasets, we evidence the effectiveness of predicting the flow rate directly from MPEG4 part-2 compressed video streams. We demonstrate an improved accuracy in comparison with a more classical RGB-based architecture and show an impressive speed up of x 2065 at prediction time.

In recent years, Dynamic Graph (DG) representations have been increasingly used for modeling dynamic systems due to their ability to integrate both topological and temporal information in a compact representation. Dynamic graphs efficiently handle applications such as social network prediction, recommender systems, traffic forecasting, or electroencephalography analysis, which cannot be addressed using standard numerical representations. As a direct consequence, dynamic graph learning has emerged as a new machine learning problem, combining challenges from both sequential/temporal data processing and static graph learning. In this research area, the Dynamic Graph Neural Network (DGNN) has become the state-of-the-art approach and a plethora of models have been proposed in the very recent years. This paper aims to provide a review of the problems and models related to dynamic graph learning. The various dynamic graph supervised learning settings are analyzed and discussed. We identify the similarities and differences between existing models concerning the way time information is modeled. Finally, we provide guidelines for DGNN design and optimization, and review public datasets for evaluating model performance on various tasks, along with the corresponding publications.

This study proposes a novel semi-supervised multi-label emotion classification approach for French tweets based on pseudo-labeling. Human subjectivity in emotional expression makes it difficult for a machine to learn. Therefore, it necessitates training supervised learning models on large datasets annotated by multiple annotators. However, creating such datasets can be costly and time-consuming. Moreover, aggregating annotations from multiple annotators to capture their collective emotional state adds complexity to the task. Semi-supervised learning techniques have shown effectiveness with limited datasets. Furthermore, Large language Models (LLMs), particularly Chat-GPT, have demonstrated superior annotation accuracy compared to annotations obtained from crowdsourcing platforms, when both are evaluated against expert-annotated data. This work introduces a novel approach for multi-label emotion classification of French tweets by leveraging pseudo-labels generated through Chat-GPT, a robust large language model. Using zero-shot, one-shot, and few-shot learning techniques, Chat-GPT annotates the unlabelled part of our dataset. These Chat-GPT-annotated pseudo-labels are then merged with manual annotations, facilitating the training of a multi-label emotion classification model via semi-supervised learning. Furthermore, within the context of our research, we present a new French tweet dataset, containing testimonials from people affected by an urban industrial incident. This dataset features 2,350 tweets, each manually annotated by three human annotators based on 8 pre-identified emotions. Benchmark results are presented for multi-label emotion classification models employing both fully supervised and semi-supervised approaches with pseudo-labeling. Our findings demonstrate the superiority of our approach for multi-label emotion classification over standard pseudo-labeling and an established baseline.

Chronic inflammatory demyelinating polyneuropathy (CIDP) is a rare, immune-mediated disorder in which an aberrant immune response causes demyelination and axonal damage of the peripheral nerves. Genetic contribution to CIDP is unclear and no genome-wide association study (GWAS) has been reported so far. In this study, we aimed to identify CIDP-related risk loci, genes, and pathways. We first focused on CIDP, and 516 CIDP cases and 403,545 controls were included in the GWAS analysis. We also investigated genetic risk for inflammatory polyneuropathy (IP), in which we performed a GWAS study using FinnGen data and combined the results with GWAS from the UK Biobank using a fixed-effect meta-analysis. A total of 1,261 IP cases and 823,730 controls were included in the analysis. Stratified analyses by gender were performed. Mendelian randomization (MR), colocalization, and transcriptome-wide association study (TWAS) analyses were performed to identify associated genes. Gene-set analyses were conducted to identify associated pathways. We identified one genome-wide significant locus at 20q13.33 for CIDP risk among women, the top variant located at the intron region of gene CDH4. Sex-combined MR, colocalization, and TWAS analyses identified three candidate pathogenic genes for CIDP and five genes for IP. MAGMA gene-set analyses identified a total of 18 pathways related to IP or CIDP. Sex-stratified analyses identified three genes for IP among males and two genes for IP among females. Our study identified suggestive risk genes and pathways for CIDP and IP. Functional analyses should be conducted to further confirm these associations.

ABSTRACT BACKGROUND : Therapeutic targets supported by genetic evidence from genome-wide association studies (GWAS) show higher probability of success in clinical trials. GWAS is a powerful approach to identify links between genetic variants and phenotypic variation; however, identifying the genes driving associations identified in GWAS remains challenging. Integration of molecular quantitative trait loci (molQTL) such as expression QTL (eQTL) using mendelian randomization (MR) and colocalization analyses can help with the identification of causal genes. Careful interpretation remains warranted because eQTL can affect the expression of multiple genes within the same locus. METHODS : We used a combination of genomic features that include variant annotation, activity-by-contact maps, MR, and colocalization with molQTL to prioritize causal genes across 4,611 disease GWAS and meta-analyses from biobank studies, namely FinnGen, Estonian Biobank and UK Biobank. RESULTS : Genes identified using this approach are enriched for gold standard causal genes and capture known biological links between disease genetics and biology. In addition, we find that eQTLs colocalizing with GWAS are statistically enriched for corresponding disease-relevant tissues. We show that predicted directionality from MR is generally consistent with matched drug mechanism of actions (>78% for approved drugs). Compared to the nearest gene mapping method our approach also shows a higher enrichment in approved therapeutic targets (risk ratio 1.38 vs 2.06). Finally, using this approach, we detected a novel association between the IL6 receptor signal transduction gene IL6ST and polymyalgia rheumatica, an indication for which sarilumab, a monoclonal antibody against IL-6, has been recently approved. CONCLUSIONS : Combining variant annotation and activity-by-contact maps to molQTL increases performance to identify causal genes, while informing on directionality which can be translated to successful target identification and drug development.