Longitudinal surveys are follow-up studies, in which participants' information is recorded at different time steps, say every two years. The difference between time series analysis and longitudinal studies is that in the former we have records of, say economic variables over a long period, whereas in the latter, we have a few records of participants, and the number of records depends on the number of participants. One problem with such studies is that we may have a large number of drop-outs, known as right-censoring in survival analysis, due to death, illness, immigration, etc. Another difference is that in time-series analysis, we aim to predict, say future stock prices based on past data, whereas in longitudinal study our goal is to predict a target based on some features. Longitudinal studies are used to study life events such as clinical psychology. Generalised linear models (GLMs) (McCullagh and Nelder, 1989), and generalised linear mixed models (GLMMs) (Frees, 2004) are traditional methods used in longitudinal studies. Qazvini (2023) employs GLMM to analyse the survival rate among the English population using the ELSA dataset.

Large Language Models (LLMs) are distinguished by their architecture, which dictates their parameter size and performance capabilities. Social scientists have increasingly adopted LLMs for text classification tasks, which are difficult to scale with human coders. While very large, closed-source models often deliver superior performance, their use presents significant risks. These include lack of transparency, potential exposure of sensitive data, challenges to replicability, and dependence on proprietary systems. Additionally, their high costs make them impractical for large-scale research projects. In contrast, open-source models, although available in various sizes, may underperform compared to commercial alternatives if used without further fine-tuning. However, open-source models offer distinct advantages: they can be run locally (ensuring data privacy), fine-tuned for specific tasks, shared within the research community, and integrated into reproducible workflows. This study demonstrates that small, fine-tuned open-source LLMs can achieve equal or superior performance to models such as ChatGPT-4. We further explore the relationship between training set size and fine-tuning efficacy in open-source models. Finally, we propose a hybrid workflow that leverages the strengths of both open and closed models, offering a balanced approach to performance, transparency, and reproducibility.

Spoken language models (SLMs) have gained increasing attention with advancements in text-based, decoder-only language models. This paper presents Double-Codebook Speaker-invariant Clustering (DC-Spin), which aims to improve speech tokenization by bridging audio signals and SLM tokens. We propose a chunk-wise approach to enable streamable DC-Spin without retraining and degradation. Comparisons of tokenization methods (self-supervised and neural audio codecs), model scalability, and downstream task proxies show that tokens easily modeled by an n-gram LM or aligned with phonemes offer strong performance, providing insights for designing speech tokenizers for SLMs. Spoken language models (SLMs) and related applications have gained more interest with the advancements of large language models (LLM) and audio tokenization techniques (Wu et al., 2024). These speech LMs resemble causal LMs in natural language processing, but SLMs take speech and, optionally, text as input and generate speech ...

Scaling has been critical in improving model performance and generalization in machine learning. It involves how a model's performance changes with increases in model size or input data, as well as how efficiently computational resources are utilized to support this growth. Despite successes in other areas, the study of scaling in Neural Network Interatomic Potentials (NNIPs) remains limited. NNIPs act as surrogate models for ab initio quantum mechanical calculations. The dominant paradigm here is to incorporate many physical domain constraints into the model, such as rotational equivariance. We contend that these complex constraints inhibit the scaling ability of NNIPs, and are likely to lead to performance plateaus in the long run. In this work, we take an alternative approach and start by systematically studying NNIP scaling strategies. Our findings indicate that scaling the model through attention mechanisms is efficient and improves model expressivity. These insights motivate us to develop an NNIP architecture designed for scalability: the Efficiently Scaled Attention Interatomic Potential (EScAIP). EScAIP leverages a multi-head self-attention formulation within graph neural networks, applying attention at the neighbor-level representations. Implemented with highly-optimized attention GPU kernels, EScAIP achieves substantial gains in efficiency--at least 10x faster inference, 5x less memory usage--compared to existing NNIPs. EScAIP also achieves state-of-the-art performance on a wide range of datasets including catalysts (OC20 and OC22), molecules (SPICE), and materials (MPTrj). We emphasize that our approach should be thought of as a philosophy rather than a specific model, representing a proof-of-concept for developing general-purpose NNIPs that achieve better expressivity through scaling, and continue to scale efficiently with increased computational resources and training data.

This paper proposes a new data-driven methodology for predicting intervals of post-fault voltage trajectories in power systems. We begin by introducing the Quantile Attention-Fourier Deep Operator Network (QAF-DeepONet), designed to capture the complex dynamics of voltage trajectories and reliably estimate quantiles of the target trajectory without any distributional assumptions. The proposed operator regression model maps the observed portion of the voltage trajectory to its unobserved post-fault trajectory. Our methodology employs a pre-training and fine-tuning process to address the challenge of limited data availability. To ensure data privacy in learning the pre-trained model, we use merging via federated learning with data from neighboring buses, enabling the model to learn the underlying voltage dynamics from such buses without directly sharing their data. After pre-training, we fine-tune the model with data from the target bus, allowing it to adapt to unique dynamics and operating conditions. Finally, we integrate conformal prediction into the fine-tuned model to ensure coverage guarantees for the predicted intervals. We evaluated the performance of the proposed methodology using the New England 39-bus test system considering detailed models of voltage and frequency controllers. Two metrics, Prediction Interval Coverage Probability (PICP) and Prediction Interval Normalized Average Width (PINAW), are used to numerically assess the model's performance in predicting intervals. The results show that the proposed approach offers practical and reliable uncertainty quantification in predicting the interval of post-fault voltage trajectories.

The prosody of a spoken utterance, including features like stress, intonation and rhythm, can significantly affect the underlying semantics, and as a consequence can also affect its textual translation. Nevertheless, prosody is rarely studied within the context of speech-to-text translation (S2TT) systems. In particular, end-to-end (E2E) systems have been proposed as well-suited for prosody-aware translation because they have direct access to the speech signal when making translation decisions, but the understanding of whether this is successful in practice is still limited. A main challenge is the difficulty of evaluating prosody awareness in translation. To address this challenge, we introduce an evaluation methodology and a focused benchmark (named ContraProST) aimed at capturing a wide range of prosodic phenomena. Our methodology uses large language models and controllable text-to-speech (TTS) to generate contrastive examples. Through experiments in translating English speech into German, Spanish, and Japanese, we find that (a) S2TT models possess some internal representation of prosody, but the prosody signal is often not strong enough to affect the translations, (b) E2E systems outperform cascades of speech recognition and text translation systems, confirming their theoretical advantage in this regard, and (c) certain cascaded systems also capture prosodic information in the translation, but only to a lesser extent that depends on the particulars of the transcript's surface form.

Surgical scene segmentation is essential for enhancing surgical precision, yet it is frequently compromised by the scarcity and imbalance of available data. To address these challenges, semantic image synthesis methods based on generative adversarial networks and diffusion models have been developed. However, these models often yield non-diverse images and fail to capture small, critical tissue classes, limiting their effectiveness. In response, we propose the Class-Aware Semantic Diffusion Model (CASDM), a novel approach which utilizes segmentation maps as conditions for image synthesis to tackle data scarcity and imbalance. Novel class-aware mean squared error and class-aware self-perceptual loss functions have been defined to prioritize critical, less visible classes, thereby enhancing image quality and relevance. Furthermore, to our knowledge, we are the first to generate multi-class segmentation maps using text prompts in a novel fashion to specify their contents. These maps are then used by CASDM to generate surgical scene images, enhancing datasets for training and validating segmentation models. Our evaluation, which assesses both image quality and downstream segmentation performance, demonstrates the strong effectiveness and generalisability of CASDM in producing realistic image-map pairs, significantly advancing surgical scene segmentation across diverse and challenging datasets.

Self-supervised learning (SSL) models have become crucial in speech processing, with recent advancements concentrating on developing architectures that capture representations across multiple timescales. The primary goal of these multi-scale architectures is to exploit the hierarchical nature of speech, where lower-resolution components aim to capture representations that align with increasingly abstract concepts (e.g., from phones to words to sentences). Although multi-scale approaches have demonstrated some improvements over single-scale models, the precise reasons for these enhancements have poor empirical support. In this study, we present an initial analysis of layer-wise representations in multi-scale architectures, with a focus on Canonical Correlation Analysis (CCA) and Mutual Information (MI). We apply this analysis to Multi-Resolution HuBERT (MR-HuBERT) and find that (1) the improved performance on SUPERB tasks is primarily due to the auxiliary low-resolution loss rather than the downsampling itself, and (2) downsampling to lower resolutions neither improves downstream performance nor correlates with higher-level information (e.g., words), though it does improve computational efficiency. These findings challenge assumptions about the multi-scale nature of MR-HuBERT and motivate the importance of disentangling computational efficiency from learning better representations.

Climate decision making is constrained by the complexity and inaccessibility of key information within lengthy, technical, and multi-lingual documents. Generative AI technologies offer a promising route for improving the accessibility of information contained within these documents, but suffer from limitations. These include (1) a tendency to hallucinate or mis-represent information, (2) difficulty in steering or guaranteeing properties of generated output, and (3) reduced performance in specific technical domains. To address these challenges, we introduce a novel evaluation framework with domain-specific dimensions tailored for climate-related documents. We then apply this framework to evaluate Retrieval-Augmented Generation (RAG) approaches and assess retrieval- and generation-quality within a prototype tool that answers questions about individual climate law and policy documents. In addition, we publish a human-annotated dataset and scalable automated evaluation tools, with the aim of facilitating broader adoption and robust assessment of these systems in the climate domain. Our findings highlight the key components of responsible deployment of RAG to enhance decision-making, while also providing insights into user experience (UX) considerations for safely deploying such systems to build trust with users in high-risk domains.

This study examines the influence of Google's search algorithm on news diversity by analyzing search results in Brazil, the UK, and the US. It explores how Google's system preferentially favors a limited number of news outlets. Utilizing algorithm auditing techniques, the research measures source concentration with the Herfindahl-Hirschman Index (HHI) and Gini coefficient, revealing significant concentration trends. The study underscores the importance of conducting horizontal analyses across multiple search queries, as focusing solely on individual results pages may obscure these patterns. Factors such as popularity, political bias, and recency were evaluated for their impact on news rankings. Findings indicate a slight leftward bias in search outcomes and a preference for popular, often national outlets. This bias, combined with a tendency to prioritize recent content, suggests that Google's algorithm may reinforce existing media inequalities. By analyzing the largest dataset to date -- 221,863 search results -- this research provides comprehensive, longitudinal insights into how algorithms shape public access to diverse news sources.