The use of language models for automatically evaluating long-form text (LLM-as-a-judge) is becoming increasingly common, yet most LLM judges are optimized exclusively for English, with strategies for enhancing their multilingual evaluation capabilities remaining largely unexplored in the current literature. This has created a disparity in the quality of automatic evaluation methods for non-English languages, ultimately hindering the development of models with better multilingual capabilities. To bridge this gap, we introduce M-Prometheus, a suite of open-weight LLM judges ranging from 3B to 14B parameters that can provide both direct assessment and pairwise comparison feedback on multilingual outputs. M-Prometheus models outperform state-of-the-art open LLM judges on multilingual reward benchmarks spanning more than 20 languages, as well as on literary machine translation (MT) evaluation covering 4 language pairs. Furthermore, M-Prometheus models can be leveraged at decoding time to significantly improve generated outputs across all 3 tested languages, showcasing their utility for the development of better multilingual models. Lastly, through extensive ablations, we identify the key factors for obtaining an effective multilingual judge, including backbone model selection and training on synthetic multilingual feedback data instead of translated data. We release our models, training dataset, and code.

Equivocation and ambiguity in public speech are well-studied discourse phenomena, especially in political science and analysis of political interviews. Inspired by the well-grounded theory on equivocation, we aim to resolve the closely related problem of response clarity in questions extracted from political interviews, leveraging the capabilities of Large Language Models (LLMs) and human expertise. To this end, we introduce a novel taxonomy that frames the task of detecting and classifying response clarity and a corresponding clarity classification dataset which consists of question-answer (QA) pairs drawn from political interviews and annotated accordingly. Our proposed two-level taxonomy addresses the clarity of a response in terms of the information provided for a given question (high-level) and also provides a fine-grained taxonomy of evasion techniques that relate to unclear, ambiguous responses (lower-level). We combine ChatGPT and human annotators to collect, validate and annotate discrete QA pairs from political interviews, to be used for our newly introduced response clarity task. We provide a detailed analysis and conduct several experiments with different model architectures, sizes and adaptation methods to gain insights and establish new baselines over the proposed dataset and task.

An important challenge in machine translation (MT) is to generate high-quality and diverse translations. Prior work has shown that the estimated likelihood from the MT model correlates poorly with translation quality. In contrast, quality evaluation metrics (such as COMET or BLEURT) exhibit high correlations with human judgments, which has motivated their use as rerankers (such as quality-aware and minimum Bayes risk decoding). However, relying on a single translation with high estimated quality increases the chances of "gaming the metric''. In this paper, we address the problem of sampling a set of high-quality and diverse translations. We provide a simple and effective way to avoid over-reliance on noisy quality estimates by using them as the energy function of a Gibbs distribution. Instead of looking for a mode in the distribution, we generate multiple samples from high-density areas through the Metropolis-Hastings algorithm, a simple Markov chain Monte Carlo approach. The results show that our proposed method leads to high-quality and diverse outputs across multiple language pairs (English$\leftrightarrow${German, Russian}) with two strong decoder-only LLMs (Alma-7b, Tower-7b).

Reinforcement learning (RL) has been proven to be an effective and robust method for training neural machine translation systems, especially when paired with powerful reward models that accurately assess translation quality. However, most research has focused on RL methods that use sentence-level feedback, leading to inefficient learning signals due to the reward sparsity problem -- the model receives a single score for the entire sentence. To address this, we propose a novel approach that leverages fine-grained, token-level quality assessments along with error severity levels using RL methods. Specifically, we use xCOMET, a state-of-the-art quality estimation system, as our token-level reward model. We conduct experiments on small and large translation datasets with standard encoder-decoder and large language models-based machine translation systems, comparing the impact of sentence-level versus fine-grained reward signals on translation quality. Our results show that training with token-level rewards improves translation quality across language pairs over baselines according to both automatic and human evaluation. Furthermore, token-level reward optimization improves training stability, evidenced by a steady increase in mean rewards over training epochs.

We introduce Spire, a speech-augmented language model (LM) capable of both translating and transcribing speech input from English into 10 other languages as well as translating text input in both language directions. Spire integrates the speech modality into an existing multilingual LM via speech discretization and continued pre-training using only 42.5K hours of speech. In particular, we adopt the pretraining framework of multilingual LMs and treat discretized speech input as an additional translation language. This approach not only equips the model with speech capabilities, but also preserves its strong text-based performance. We achieve this using significantly less data than existing speech LMs, demonstrating that discretized speech input integration as an additional language is feasible during LM adaptation. We make our code and models available to the community.

While machine translation (MT) systems are achieving increasingly strong performance on benchmarks, they often produce translations with errors and anomalies. Understanding these errors can potentially help improve the translation quality and user experience. This paper introduces xTower, an open large language model (LLM) built on top of TowerBase designed to provide free-text explanations for translation errors in order to guide the generation of a corrected translation. The quality of the generated explanations by xTower are assessed via both intrinsic and extrinsic evaluation. We ask expert translators to evaluate the quality of the explanations across two dimensions: relatedness towards the error span being explained and helpfulness in error understanding and improving translation quality. Extrinsically, we test xTower across various experimental setups in generating translation corrections, demonstrating significant improvements in translation quality. Our findings highlight xTower's potential towards not only producing plausible and helpful explanations of automatic translations, but also leveraging them to suggest corrected translations.

The target space geometry of abelian vector multiplets in ${\cal N}= 2$ theories in four and five space-time dimensions is called special geometry. It can be elegantly formulated in terms of Hessian geometry. In this review, we introduce Hessian geometry, focussing on aspects that are relevant for the special geometries of four- and five-dimensional vector multiplets. We formulate ${\cal N}= 2$ theories in terms of Hessian structures and give various concrete applications of Hessian geometry, ranging from static BPS black holes in four and five space-time dimensions to topological string theory, emphasizing the role of the Hesse potential. We also discuss the r-map and c-map which relate the special geometries of vector multiplets to each other and to hypermultiplet geometries. By including time-like dimensional reductions, we obtain theories in Euclidean signature, where the scalar target spaces carry para-complex versions of special geometry.

The spectrum of coherent gluon radiation from a quark-anti-quark pair experiencing multiple scatterings within a coloured medium is central for understanding in-medium parton cascades. Despite its foundational importance, current results are limited by reliance on simplified scattering rates, such as the harmonic oscillator approximation, valid only in restricted phase-space regions. Using the formalism introduced in a previous article, we express the gluon emission spectrum as a set of differential equations that can be solved numerically, circumventing conventional approximations. We present the transverse momentum and energy distributions of emitted gluons for realistic interaction models, illustrating the breakdown of colour coherence across the entire accessible phase-space, and consequently enabling a higher-precision description of jet observables.

Large language models (LLMs) have demonstrated strong performance in sentence-level machine translation, but scaling to document-level translation remains challenging, particularly in modeling long-range dependencies and discourse phenomena across sentences and paragraphs. In this work, we propose a method to improve LLM-based long-document translation through targeted fine-tuning on high-quality document-level data, which we curate and introduce as DocBlocks. Our approach supports multiple translation paradigms, including direct document-to-document and chunk-level translation, by integrating instructions both with and without surrounding context. This enables models to better capture cross-sentence dependencies while maintaining strong sentence-level translation performance. Experimental results show that incorporating multiple translation paradigms improves document-level translation quality and inference speed compared to prompting and agent-based methods.

The spin point groups are finite groups whose elements act on both real space and spin space. Among these groups are the magnetic point groups in the case where the real and spin space operations are locked to one another. The magnetic point groups are central to magnetic crystallography for strong spin-orbit coupled systems and the spin point groups generalize these to the intermediate and weak spin-orbit coupled cases. The spin point groups were introduced in the 1960's in the context of condensed matter physics and enumerated shortly thereafter. In this paper, we complete the theory ofcrystallographic spin point groups by presenting an account of these groups and their representation theory. Our main findings are that the so-called nontrivial spin point groups (numbering $598$ groups) have co-irreps corresponding exactly to the (co-)-irreps of regular or black and white groups and we tabulate this correspondence for each nontrivial group. However a total spin group, comprising the product of a nontrivial group and a spin-only group, has new co-irreps in cases where there is continuous rotational freedom. We provide explicit co-irrep tables for all these instances. We also discuss new forms of spin-only group extending the Litvin-Opechowski classes. To exhibit the usefulness of these groups to physically relevant problems we discuss a number of examples from electronic band structures of altermagnets to magnons.

Conformal prediction is a distribution-free framework for uncertainty quantification that replaces point predictions with sets, offering marginal coverage guarantees (i.e., ensuring that the prediction sets contain the true label with a specified probability, in expectation). In this paper, we uncover a novel connection between conformal prediction and sparse softmax-like transformations, such as sparsemax and $\gamma$-entmax (with $\gamma > 1$), which may assign nonzero probability only to a subset of labels. We introduce new non-conformity scores for classification that make the calibration process correspond to the widely used temperature scaling method. At test time, applying these sparse transformations with the calibrated temperature leads to a support set (i.e., the set of labels with nonzero probability) that automatically inherits the coverage guarantees of conformal prediction. Through experiments on computer vision and text classification benchmarks, we demonstrate that the proposed method achieves competitive results in terms of coverage, efficiency, and adaptiveness compared to standard non-conformity scores based on softmax.

This work addresses the problem of sensing the world: how to learn a multimodal representation of a reinforcement learning agent's environment that allows the execution of tasks under incomplete perceptual conditions. To address such problem, we argue for hierarchy in the design of representation models and contribute with a novel multimodal representation model, MUSE. The proposed model learns hierarchical representations: low-level modality-specific representations, encoded from raw observation data, and a high-level multimodal representation, encoding joint-modality information to allow robust state estimation. We employ MUSE as the sensory representation model of deep reinforcement learning agents provided with multimodal observations in Atari games. We perform a comparative study over different designs of reinforcement learning agents, showing that MUSE allows agents to perform tasks under incomplete perceptual experience with minimal performance loss. Finally, we evaluate the performance of MUSE in literature-standard multimodal scenarios with higher number and more complex modalities, showing that it outperforms state-of-the-art multimodal variational autoencoders in single and cross-modality generation.

The sensitivity of the Deep Underground Neutrino Experiment (DUNE) to neutrino oscillation is determined, based on a full simulation, reconstruction, and event selection of the far detector and a full simulation and parameterized analysis of the near detector. Detailed uncertainties due to the flux prediction, neutrino interaction model, and detector effects are included. DUNE will resolve the neutrino mass ordering to a precision of 5$\sigma$, for all $\delta_{\mathrm{CP}}$ values, after 2 years of running with the nominal detector design and beam configuration. It has the potential to observe charge-parity violation in the neutrino sector to a precision of 3$\sigma$ (5$\sigma$) after an exposure of 5 (10) years, for 50\% of all $\delta_{\mathrm{CP}}$ values. It will also make precise measurements of other parameters governing long-baseline neutrino oscillation, and after an exposure of 15 years will achieve a similar sensitivity to $\sin^{2} 2\theta_{13}$ to current reactor experiments.

Deep Learning for medical imaging faces challenges in adapting and generalizing to new contexts. Additionally, it often lacks sufficient labeled data for specific tasks requiring significant annotation effort. Continual Learning (CL) tackles adaptability and generalizability by enabling lifelong learning from a data stream while mitigating forgetting of previously learned knowledge. Active Learning (AL) reduces the number of required annotations for effective training. This work explores both approaches (CAL) to develop a novel framework for robust medical image analysis. Based on the automatic recognition of shifts in image characteristics, Replay-Base Architecture for Context Adaptation (RBACA) employs a CL rehearsal method to continually learn from diverse contexts, and an AL component to select the most informative instances for annotation. A novel approach to evaluate CAL methods is established using a defined metric denominated IL-Score, which allows for the simultaneous assessment of transfer learning, forgetting, and final model performance. We show that RBACA works in domain and class-incremental learning scenarios, by assessing its IL-Score on the segmentation and diagnosis of cardiac images. The results show that RBACA outperforms a baseline framework without CAL, and a state-of-the-art CAL method across various memory sizes and annotation budgets. Our code is available in this https URL .

Quantum computers, using efficient Hamiltonian evolution routines, have the potential to simulate Green's functions of classically-intractable quantum systems. However, the decoherence errors of near-term quantum processors prohibit large evolution times, posing limits to the spectrum resolution. In this work, we show that Atomic Norm Minimization, a well-known super-resolution technique, can significantly reduce the minimum circuit depth for accurate spectrum reconstruction. We demonstrate this technique by recovering the spectral function of an impurity model from measurements of its Green's function on an IBM quantum computer. The reconstruction error with the Atomic Norm Minimization is one order of magnitude smaller than with more standard signal processing methods. Super-resolution methods can facilitate the simulation of large and previously unexplored quantum systems, and may constitute a useful non-variational tool to establish a quantum advantage in a nearer future.