Grounding large language models (LLMs) in external knowledge sources is a promising method for faithful prediction. While existing grounding approaches work well for simple queries, many real-world information needs require synthesizing multiple pieces of evidence. We introduce "integrative grounding" -- the challenge of retrieving and verifying multiple inter-dependent pieces of evidence to support a hypothesis query. To systematically study this problem, we repurpose data from four domains for evaluating integrative grounding capabilities. Our investigation reveals two critical findings: First, in groundedness verification, while LLMs are robust to redundant evidence, they tend to rationalize using internal knowledge when information is incomplete. Second, in examining retrieval planning strategies, we find that undirected planning can degrade performance through noise introduction, while premise abduction emerges as a promising approach due to its logical constraints. Additionally, LLMs' zero-shot self-reflection capabilities consistently improve grounding quality. These insights provide valuable direction for developing more effective integrative grounding systems.

The development of Natural Language Processing (NLP) tools for low-resource languages is critically hindered by the scarcity of annotated datasets. This paper addresses this fundamental challenge by introducing HausaMovieReview, a novel benchmark dataset comprising 5,000 YouTube comments in Hausa and code-switched English. The dataset was meticulously annotated by three independent annotators, demonstrating a robust agreement with a Fleiss' Kappa score of 0.85 between annotators. We used this dataset to conduct a comparative analysis of classical models (Logistic Regression, Decision Tree, K-Nearest Neighbors) and fine-tuned transformer models (BERT and RoBERTa). Our results reveal a key finding: the Decision Tree classifier, with an accuracy and F1-score 89.72% and 89.60% respectively, significantly outperformed the deep learning models. Our findings also provide a robust baseline, demonstrating that effective feature engineering can enable classical models to achieve state-of-the-art performance in low-resource contexts, thereby laying a solid foundation for future research. Keywords: Hausa, Kannywood, Low-Resource Languages, NLP, Sentiment Analysis

Reward models are essential for aligning language model outputs with human preferences, yet existing approaches often lack both controllability and interpretability. These models are typically optimized for narrow objectives, limiting their generalizability to broader downstream tasks. Moreover, their scalar outputs are difficult to interpret without contextual reasoning. To address these limitations, we introduce $\shortmethodname$, a novel reward modeling framework that is rubric-agnostic, generalizable across evaluation dimensions, and provides interpretable, reasoned score assignments. $\shortmethodname$ enables more transparent and flexible evaluation of language models, supporting robust alignment with diverse human values and use cases. Our models, data, and code are available as open source at https://github.com/rubricreward/r3.

Fairness in ranking models is crucial, as disparities in exposure can disproportionately affect protected groups. Most fairness-aware ranking systems focus on ensuring comparable average exposure for groups across the entire ranked list, which may not fully address real-world concerns. For example, when a ranking model is used for allocating resources among candidates or disaster hotspots, decision-makers often prioritize only the top-$K$ ranked items, while the ranking beyond top-$K$ becomes less relevant. In this paper, we propose a list-wise learning-to-rank framework that addresses the issues of inequalities in top-$K$ rankings at training time. Specifically, we propose a top-$K$ exposure disparity measure that extends the classic exposure disparity metric in a ranked list. We then learn a ranker to balance relevance and fairness in top-$K$ rankings. Since direct top-$K$ selection is computationally expensive for a large number of items, we transform the non-differentiable selection process into a differentiable objective function and develop efficient stochastic optimization algorithms to achieve both high accuracy and sufficient fairness. Extensive experiments demonstrate that our method outperforms existing methods.

Two central capabilities of language models (LMs) are: (i) drawing on prior knowledge about entities, which allows them to answer queries such as "What's the official language of Austria?", and (ii) adapting to new information provided in context, e.g., "Pretend the official language of Austria is Tagalog.", that is pre-pended to the question. In this article, we introduce targeted persuasion score (TPS), designed to quantify how persuasive a given context is to an LM where persuasion is operationalized as the ability of the context to alter the LM's answer to the question. In contrast to evaluating persuasiveness only by inspecting the greedily decoded answer under the model, TPS provides a more fine-grained view of model behavior. Based on the Wasserstein distance, TPS measures how much a context shifts a model's original answer distribution toward a target distribution. Empirically, through a series of experiments, we show that TPS captures a more nuanced notion of persuasiveness than previously proposed metrics.

We propose a novel benchmark for camera identification via Photo Response Non-Uniformity (PRNU) estimation. The benchmark comprises 13K photos taken with 120+ cameras, where the training and test photos are taken in different scenarios, enabling ``in-the-wild'' evaluation. In addition, we propose a novel PRNU-based camera identification model that employs a hybrid architecture, comprising a denoising autoencoder to estimate the PRNU signal and a convolutional network that can perform 1:N verification of camera devices. Instead of using a conventional approach based on contrastive learning, our method takes the Hadamard product between reference and query PRNU signals as input. This novel design leads to significantly better results compared with state-of-the-art models based on denoising autoencoders and contrastive learning. We release our dataset and code at: https://github.com/CroitoruAlin/PRNU-Bench.

ABSTRACT Free-form, text-based audio editing remains a persistent challenge, despite progress in inversion-based neural methods. Current approaches rely on slow inversion procedures, limiting their practicality. We present a virtual-consistency based audio editing system that bypasses inversion by adapting the sampling process of diffusion models. Our pipeline is model-agnostic, requiring no fine-tuning or architectural changes, and achieves substantial speed-ups over recent neural editing baselines. Crucially, it achieves this efficiency without compromising quality, as demonstrated by quantitative benchmarks and a user study involving 16 participants.

This paper enhances the study of sentiment analysis for the Central Kurdish language by integrating the Bidirectional Encoder Representations from Transformers (BERT) into Natural Language Processing techniques. Kurdish is a low - resourced language, having a high level of linguistic diversity with minimal computational resources, making sentiment analysis somewhat challenging. Earlier, this was done using a traditional w ord embedding model, such as Word2Vec, but with the emergence of new language models, specifically BERT, there is hope for improvements. The better word embedding capabilities of BERT lend to this study, aiding in the capturing of the nuanced semantic pool and the contextual intricacies of the language under study, the Kurdish language, thus setting a new benchmark for sentiment analysis in low - resource languages. The steps include collecting and normalizing a large corpus of Kurdish texts, pretraining BERT with a special tokenizer for Kurdish, and developing different models for sentiment analysis including Bidirectional Long Short - Term Memory ( BiLSTM), Multi - L ayer Perceptron ( MLP), and finetuning the BERT classifier . The proposed approach consists of 3 cla sses: positive, negative, and neutral sentiment analysis using a sentiment embedding of BERT in four different configurations. The accuracy of the best - performing classifier, BiLSTM, is 74.09%. For the BERT with an MLP classifier model, the maximum accuracy achieved is 73.96%, while the fine - tuned BERT model tops the others with 75.37% accuracy. Additionally, the fine - tuned BERT model demonstrates a vast improvement when focused on t wo 2 - class sentiment analyses positive and negative with an accuracy of 86.

A large-scale dataset is essential for training a well-generalized deep-learning model. Most such datasets are collected via scraping from various internet sources, inevitably introducing duplicated data. In the symbolic music domain, these duplicates often come from multiple user arrangements and metadata changes after simple editing. However, despite critical issues such as unreliable training evaluation from data leakage during random splitting, dataset duplication has not been extensively addressed in the MIR community. This study investigates the dataset duplication issues regarding Lakh MIDI Dataset (LMD), one of the largest publicly available sources in the symbolic music domain. To find and evaluate the best retrieval method for duplicated data, we employed the Clean MIDI subset of the LMD as a benchmark test set, in which different versions of the same songs are grouped together. We first evaluated rule-based approaches and previous symbolic music retrieval models for de-duplication and also investigated with a contrastive learning-based BERT model with various augmentations to find duplicate files. As a result, we propose three different versions of the filtered list of LMD, which filters out at least 38,134 samples in the most conservative settings among 178,561 files.

The rapid advancement of large language models (LLMs) has heightened concerns about benchmark data contamination (BDC), where models inadvertently memorize evaluation data during the training process, inflating performance metrics, and undermining genuine generalization assessment. This paper introduces the Data Contamination Risk (DCR) framework, a lightweight, interpretable pipeline designed to detect and quantify BDC risk across four granular levels: semantic, informational, data, and label. By synthesizing contamination scores via a fuzzy inference system, DCR produces a unified DCR Factor that adjusts raw accuracy to reflect contamination-aware performance. Validated on 9 LLMs (0.5B-72B) across sentiment analysis, fake news detection, and arithmetic reasoning tasks, the DCR framework reliably diagnoses contamination severity and with accuracy adjusted using the DCR Factor to within 4% average error across the three benchmarks compared to the uncontaminated baseline. Emphasizing computational efficiency and transparency, DCR provides a practical tool for integrating contamination assessment into routine evaluations, fostering fairer comparisons and enhancing the credibility of LLM benchmarking practices.