Motions submitted to the Swedish Parliament are important data for social science and humanities researchers. We introduce a new research corpus, the Swedish Parliamentary Motions Corpus, which is larger and more developed than previously available research corpora for the Swedish motions. The corpus contains annotated and structured parliamentary motions over more than 150 years, through the bicameral parliament (1867–1970) and Sweden’s current unicameral parliament (1971–). Along with the corpus, we describe procedures to measure and ensure transparency around issues related to the data quality of the corpus. In addition, we link motions’ authors to a rich metadata set, ensuring the corpus’s utility in various research applications.

The general applicability and robustness of posterior inference algorithms is critical to widely used probabilistic programming languages such as Stan, PyMC, Pyro, and Turing.jl. When designing a new inference algorithm, whether it involves Monte Carlo sampling or variational approximation, the fundamental problem is evaluating its accuracy and efficiency across a range of representative target posteriors. To solve this problem, we propose posteriordb, a database of models and data sets defining target densities along with reference Monte Carlo draws. We further provide a guide to the best practices in using posteriordb for algorithm evaluation and comparison. To provide a wide range of realistic posteriors, posteriordb currently comprises 120 representative models with data, and has been instrumental in developing several inference algorithms.

Legal citation detection in court judgments underpins reliable precedent mapping, citation analytics, and document retrieval.Extracting references to legislation and case law in the United Kingdom is especially challenging: citation styles have evolved over centuries, and judgments routinely cite foreign or historical authorities.We conduct the first systematic comparison of three modelling paradigms on this task using the Cambridge Law Corpus: (i) rule-based regular expressions; (ii) transformer-based encoders (BERT, RoBERTa, LEGAL-BERT, ModernBERT); and (iii) large language models (GPT-4.1).We produced a gold-standard high-quality corpus of 190 court judgments containing 45,179 fine-grained annotations for UK and non-UK legislation and case references.ModernBERT achieves a macroaveraged F1 of 93.3%,only marginally ahead of the other encoder-only models, yet significantly outperforming the strongest regularexpression baseline (35.42% F1) and GPT-4.1 (76.57%F1).

The Swedish parliamentary records are an important source material for social science and humanities researchers. We introduce a new research corpus, the Swedish Parliament Corpus, which is larger and more developed than previously available research corpora for the Swedish parliament. The corpus contains annotated and structured parliamentary records over more than 150 years, through the bicameral parliament (1867–1970) and the unicameral parliament (1971–). In addition to the records, which contain all speeches in the parliament, we also provide a database of all members of parliament over the same period. Along with the corpus, we describe procedures to ensure data quality. The corpus facilitates detailed analysis of parliamentary speeches in several research fields.

Recently, new methods for model assessment, based on subsampling and posterior approximations, have been proposed for scaling leave-one-out cross-validation (LOO-CV) to large datasets. Although these methods work well for estimating predictive performance for individual models, they are less powerful in model comparison. We propose an efficient method for estimating differences in predictive performance by combining fast approximate LOO surrogates with exact LOO sub-sampling using the difference estimator and supply proofs with regards to scaling characteristics. The resulting approach can be orders of magnitude more efficient than previous approaches, as well as being better suited to model comparison.

To scale non-parametric extensions of probabilistic topic models such as Latent Dirichlet allocation to larger data sets, practitioners rely increasingly on parallel and distributed systems.In this work, we study data-parallel training for the hierarchical Dirichlet process (HDP) topic model.Based upon a representation of certain conditional distributions within an HDP, we propose a doubly sparse data-parallel sampler for the HDP topic model.This sampler utilizes all available sources of sparsity found in natural language-an important way to make computation efficient.We benchmark our method on a well-known corpus (PubMed) with 8m documents and 768m tokens, using a single multi-core machine in under four days.

We examine the accuracy of black box variational posterior approximations for parametric models in a probabilistic programming context. The performance of these approximations depends on (1) how well the variational family approximates the true posterior distribution, (2) the choice of divergence, and (3) the optimization of the variational objective. We show that even when the true variational family is used, high-dimensional posteriors can be very poorly approximated using common stochastic gradient descent (SGD) optimizers. Motivated by recent theory, we propose a simple and parallel way to improve SGD estimates for variational inference. The approach is theoretically motivated and comes with a diagnostic for convergence and a novel stopping rule, which is robust to noisy objective functions evaluations. We show empirically, the new workflow works well on a diverse set of models and datasets, or warns if the stochastic optimization fails or if the used variational distribution is not good.

Abstract We show that the eccentricities, diameter, radius, and Wiener index of an undirected n-vertex graph with nonnegative edge lengths can be computed in time $$O(n\cdot \left( {\begin{array}{c}k+\lceil \log n\rceil \\ k\end{array}}\right) \cdot 2^k \log n)$$ O ( n · k + ⌈ log n ⌉ k · 2 k log n ) , where k is linear in the treewidth of the graph. For every $$\epsilon >0$$ ϵ > 0 , this bound is $$n^{1+\epsilon }\exp O(k)$$ n 1 + ϵ exp O ( k ) , which matches a hardness result of Abboud et al. (in: Proceedings of the Twenty-Seventh Annual ACM-SIAM Symposium on Discrete Algorithms, 2016. 10.1137/1.9781611974331.ch28 ) and closes an open problem in the multivariate analysis of polynomial-time computation. To this end, we show that the analysis of an algorithm of Cabello and Knauer (Comput Geom 42:815–824, 2009. 10.1016/j.comgeo.2009.02.001 ) in the regime of non-constant treewidth can be improved by revisiting the analysis of orthogonal range searching, improving bounds of the form $$\log ^d n$$ log d n to $$\left( {\begin{array}{c}d+\lceil \log n\rceil \\ d\end{array}}\right)$$ d + ⌈ log n ⌉ d , as originally observed by Monier (J Algorithms 1:60–74, 1980. 10.1016/0196-6774(80)90005-X ). We also investigate the parameterization by vertex cover number.

BACKGROUND: Exome and genome sequencing is becoming the method of choice for rare disease diagnostics. One of the key challenges remaining is distinguishing the disease causing variants from the benign background variation. After analysis and annotation of the sequencing data there are typically thousands of candidate variants requiring further investigation. One of the most effective and least biased ways to reduce this number is to assess the rarity of a variant in any population. Currently, there are a number of reliable sources of information for major population frequencies when considering single nucleotide variants (SNVs) and small insertion and deletions (INDELs), with gnomAD as the most prominent public resource available. However, local variation or frequencies in sub-populations may be underrepresented in these public resources. In contrast, for structural variation (SV), the background frequency in the general population is more or less unknown mostly due to challenges in calling SVs in a consistent way. Keeping track of local variation is one way to overcome these problems and significantly reduce the number of potential disease causing variants retained for manual inspection, both for SNVs and SVs. RESULTS: Here, we present loqusdb, a tool to solve the challenge of keeping track of any type of variant observations from genome sequencing data. Loqusdb was designed to handle a large flow of samples and unlike other solutions, samples can be added continuously to the database without rebuilding it, facilitating improvements and additions. We assessed the added value of a local observations database using 98 samples annotated with information from a background of 888 unrelated individuals. CONCLUSIONS: We show both how powerful SV analysis can be when filtering for population frequencies and how the number of apparently rare SNVs/INDELs can be reduced by adding local population information even after annotating the data with other large frequency databases, such as gnomAD. In conclusion, we show that a local frequency database is an attractive, and a necessary addition to the publicly available databases that facilitate the analysis of exome and genome data in a clinical setting.

Miriam Hurtado Bodell, Martin Arvidsson, Måns Magnusson. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019.

Model inference, such as model comparison, model checking, and model selection, is an important part of model development. Leave-one-out cross-validation (LOO) is a general approach for assessing the generalizability of a model, but unfortunately, LOO does not scale well to large datasets. We propose a combination of using approximate inference techniques and probability-proportional-to-size-sampling (PPS) for fast LOO model evaluation for large datasets. We provide both theoretical and empirical results showing good properties for large data.

Latent Dirichlet Allocation (LDA) is a topic model widely used in natural language processing and machine learning. Most approaches to training the model rely on iterative algorithms, which makes it difficult to run LDA on big corpora that are best analyzed in parallel and distributed computational environments. Indeed, current approaches to parallel inference either don't converge to the correct posterior or require storage of large dense matrices in memory. We present a novel sampler that overcomes both problems, and we show that this sampler is faster, both empirically and theoretically, than previous Gibbs samplers for LDA. We do so by employing a novel Pólya-urn-based approximation in the sparse partially collapsed sampler for LDA. We prove that the approximation error vanishes with data size, making our algorithm asymptotically exact, a property of importance for large-scale topic models. In addition, we show, via an explicit example, that-contrary to popular belief in the topic modeling literature-partially collapsed samplers can be more efficient than fully collapsed samplers. We conclude by comparing the performance of our algorithm with that of other approaches on well-known corpora.

We show that the eccentricities, diameter, radius, and Wiener index of an undirected n-vertex graph with nonnegative edge lengths can be computed in time O(n * binom{k+ceil[log n]}{k} * 2^k k^2 log n), where k is the treewidth of the graph. For every epsilon>0, this bound is n^{1+epsilon}exp O(k), which matches a hardness result of Abboud, Vassilevska Williams, and Wang (SODA 2015) and closes an open problem in the multivariate analysis of polynomial-time computation. To this end, we show that the analysis of an algorithm of Cabello and Knauer (Comp. Geom., 2009) in the regime of non-constant treewidth can be improved by revisiting the analysis of orthogonal range searching, improving bounds of the form log^d n to binom{d+ceil[log n]}{d}, as originally observed by Monier (J. Alg. 1980). We also investigate the parameterization by vertex cover number.

We suggest a Bayesian approach to the problem of reducing bug turn-around time in large software development organizations. Our approach is to use classification to predict where bugs are located in components. This classification is a form of automatic fault localization (AFL) at the component level. The approach only relies on historical bug reports and does not require detailed analysis of source code or detailed test runs. Our approach addresses two problems identified in user studies of AFL tools. The first problem concerns the trust in which the user can put in the results of the tool. The second problem concerns understanding how the results were computed. The proposed model quantifies the uncertainty in its predictions and all estimated model parameters. Additionally, the output of the model explains why a result was suggested. We evaluate the approach on more than 50000 bugs.

MOTIVATION: Fast and accurate quality control is essential for studies involving next-generation sequencing data. Whilst numerous tools exist to quantify QC metrics, there is no common approach to flexibly integrate these across tools and large sample sets. Assessing analysis results across an entire project can be time consuming and error prone; batch effects and outlier samples can easily be missed in the early stages of analysis. RESULTS: We present MultiQC, a tool to create a single report visualising output from multiple tools across many samples, enabling global trends and biases to be quickly identified. MultiQC can plot data from many common bioinformatics tools and is built to allow easy extension and customization. AVAILABILITY AND IMPLEMENTATION: MultiQC is available with an GNU GPLv3 license on GitHub, the Python Package Index and Bioconda. Documentation and example reports are available at http://multiqc.info CONTACT: [email protected].