Large language models (LLMs) and their multimodal variants can now process visual inputs, including images of text. This raises an intriguing question: can we compress textual inputs by feeding them as images to reduce token usage while preserving performance? In this paper, we show that visual text representations are a practical and surprisingly effective form of input compression for decoder LLMs. We exploit the idea of rendering long text inputs as a single image and provide it directly to the model. This leads to dramatically reduced number of decoder tokens required, offering a new form of input compression. Through experiments on two distinct benchmarks RULER (long-context retrieval) and CNN/DailyMail (document summarization) we demonstrate that this text-as-image method yields substantial token savings (often nearly half) without degrading task performance.

This paper addresses the challenges in developing language models for less-represented languages, with a focus on Luxembourgish. Despite its active development, Luxembourgish faces a digital data scarcity, exacerbated by Luxembourg's multilingual context. We propose a novel text generation model based on the T5 architecture, combining limited Luxembourgish data with equal amounts, in terms of size and type, of German and French data. We hypothesise that a model trained on Luxembourgish, German, and French will improve the model's cross-lingual transfer learning capabilities and outperform monolingual and large multilingual models. To verify this, the study at hand explores whether multilingual or monolingual training is more beneficial for Luxembourgish language generation. For the evaluation, we introduce LuxGen, a text generation benchmark that is the first of its kind for Luxembourgish.

Orthographic variation is very common in Luxembourgish texts due to the absence of a fully-fledged standard variety. Additionally, developing NLP tools for Luxembourgish is a difficult task given the lack of annotated and parallel data, which is exacerbated by ongoing standardization. In this paper, we propose the first sequence-to-sequence normalization models using the ByT5 and mT5 architectures with training data obtained from word-level real-life variation data. We perform a fine-grained, linguistically-motivated evaluation to test byte-based, word-based and pipeline-based models for their strengths and weaknesses in text normalization. We show that our sequence model using real-life variation data is an effective approach for tailor-made normalization in Luxembourgish.

Recent advances have extended the context window of frontier LLMs dramatically, from a few thousand tokens up to millions, enabling entire books and codebases to fit into context. However, the compute costs of inferencing long-context LLMs are massive and often prohibitive in practice. RAG offers an efficient and effective alternative: retrieve and process only the subset of the context most important for the current task. Although promising, recent work applying RAG to long-context tasks has two core limitations: 1) there has been little focus on making the RAG pipeline compute efficient, and 2) such works only test on simple QA tasks, and their performance on more challenging tasks is unclear. To address this, we develop an algorithm based on PageRank, a graph-based retrieval algorithm, which we call mixture-of-PageRanks (MixPR). MixPR uses a mixture of PageRank-based graph-retrieval algorithms implemented using sparse matrices for efficent, cheap retrieval that can deal with a variety of complex tasks. Our MixPR retriever achieves state-of-the-art results across a wide range of long-context benchmark tasks, outperforming both existing RAG methods, specialized retrieval architectures, and long-context LLMs despite being far more compute efficient. Due to using sparse embeddings, our retriever is extremely compute efficient, capable of embedding and retrieving millions of tokens within a few seconds and runs entirely on CPU.

The Universal Dependencies (UD) project has significantly expanded linguistic coverage across 161 languages, yet Luxembourgish, a West Germanic language spoken by approximately 400,000 people, has remained absent until now. In this paper, we introduce LuxBank, the first UD Treebank for Luxembourgish, addressing the gap in syntactic annotation and analysis for this `low-research' language. We establish formal guidelines for Luxembourgish language annotation, providing the foundation for the first large-scale quantitative analysis of its syntax. LuxBank serves not only as a resource for linguists and language learners but also as a tool for developing spell checkers and grammar checkers, organising existing text archives and even training large language models. By incorporating Luxembourgish into the UD framework, we aim to enhance the understanding of syntactic variation within West Germanic languages and offer a model for documenting smaller, semi-standardised languages. This work positions Luxembourgish as a valuable resource in the broader linguistic and NLP communities, contributing to the study of languages with limited research and resources.

The rapid increase in textual information means we need more efficient methods to sift through, organize, and understand it all. While retrieval-augmented generation (RAG) models excel in accessing information from large document collections, they struggle with complex tasks that require aggregation and reasoning over information spanning across multiple documents--what we call holistic reasoning. Long-context language models (LCLMs) have great potential for managing large-scale documents, but their holistic reasoning capabilities remain unclear. In this work, we introduce HoloBench, a novel framework that brings database reasoning operations into text-based contexts, making it easier to systematically evaluate how LCLMs handle holistic reasoning across large documents. Our approach adjusts key factors such as context length, information density, distribution of information, and query complexity to evaluate LCLMs comprehensively. Our experiments show that the amount of information in the context has a bigger influence on LCLM performance than the actual context length. Furthermore, the complexity of queries affects performance more than the amount of information, particularly for different types of queries. Interestingly, queries that involve finding maximum or minimum values are easier for LCLMs and are less affected by context length, even though they pose challenges for RAG systems. However, tasks requiring the aggregation of multiple pieces of information show a noticeable drop in accuracy as context length increases. Additionally, we find that while grouping relevant information generally improves performance, the optimal positioning varies across models. Our findings surface both the advancements and the ongoing challenges in achieving a holistic understanding of long contexts.

The study aims to investigate the similarities and differences in the brain damage caused by Hypoxia-Ischemia (HI), Hypoglycemia, and Epilepsy. Hypoglycemia poses a significant challenge in improving glycemic regulation for insulin-treated patients, while HI brain disease in neonates is associated with low oxygen levels. The study examines the possibility of using a combination of medical data and Electroencephalography (EEG) measurements to predict outcomes over a two-year period. The study employs a multilevel fusion of data features to enhance the accuracy of the predictions. Therefore this paper suggests a hybridized classification model for Hypoxia-Ischemia and Hypoglycemia, Epilepsy brain injury (HCM-BI). A Support Vector Machine is applied with clinical details to define the Hypoxia-Ischemia outcomes of each infant. The newborn babies are assessed every two years again to know the neural development results. A selection of four attributes is derived from the Electroencephalography records, and SVM does not get conclusions regarding the classification of diseases. The final feature extraction of the EEG signal is optimized by the Bayesian Neural Network (BNN) to get the clear health condition of Hypoglycemia and Epilepsy patients. Through monitoring and assessing physical effects resulting from Electroencephalography, The Bayesian Neural Network (BNN) is used to extract the test samples with the most log data and to report hypoglycemia and epilepsy Keywords- Hypoxia-Ischemia , Hypoglycemia , Epilepsy , Multilevel Fusion of Data Features , Bayesian Neural Network (BNN) , Support Vector Machine (SVM)

Graph neural networks form a class of deep learning architectures specifically designed to work with graph-structured data. As such, they share the inherent limitations and problems of deep learning, especially regarding the issues of explainability and trustworthiness. We propose $\mu\mathcal{G}$, an original domain-specific language for the specification of graph neural networks that aims to overcome these issues. The language's syntax is introduced, and its meaning is rigorously defined by a denotational semantics. An equivalent characterization in the form of an operational semantics is also provided and, together with a type system, is used to prove the type soundness of $\mu\mathcal{G}$. We show how $\mu\mathcal{G}$ programs can be represented in a more user-friendly graphical visualization, and provide examples of its generality by showing how it can be used to define some of the most popular graph neural network models, or to develop any custom graph processing application.

ProMoAI is a novel tool that leverages Large Language Models (LLMs) to automatically generate process models from textual descriptions, incorporating advanced prompt engineering, error handling, and code generation techniques. Beyond automating the generation of complex process models, ProMoAI also supports process model optimization. Users can interact with the tool by providing feedback on the generated model, which is then used for refining the process model. ProMoAI utilizes the capabilities LLMs to offer a novel, AI-driven approach to process modeling, significantly reducing the barrier to entry for users without deep technical knowledge in process modeling.

Verification of properties expressed as-regular languages such as LTL can benefit hugely from stutter-insensitivity, using a diverse set of reduction strategies. However properties that are not stutter-insensitive, for instance due to the use of the neXt operator of LTL or to some form of counting in the logic, are not covered by these techniques in general. We propose in this paper to study a weaker property than stutter-insensitivity. In a stutter insensitive language both adding and removing stutter to a word does not change its acceptance, any stuttering can be abstracted away; by decomposing this equivalence relation into two implications we obtain weaker conditions. We define a shortening insensitive language where any word that stutters less than a word in the language must also belong to the language. A lengthening insensitive language has the dual property. A semi-decision procedure is then introduced to reliably prove shortening insensitive properties or deny lengthening insensitive properties while working with a reduction of a system. A reduction has the property that it can only shorten runs. Lipton's transaction reductions or Petri net agglomerations are examples of eligible structural reduction strategies. An implementation and experimental evidence is provided showing most nonrandom properties sensitive to stutter are actually shortening or lengthening insensitive. Performance of experiments on a large (random) benchmark from the model-checking competition indicate that despite being a semi-decision procedure, the approach can still improve state of the art verification tools.