Search

Personalized content recommendations have been pivotal to the content experience in digital media from video streaming to social networks. However, several domain specific challenges have held back adoption of recommender systems in news publishing. To address these challenges, we introduce the Ekstra Bladet News Recommendation Dataset (EB-NeRD). The dataset encompasses data from over a million unique users and more than 37 million impression logs from Ekstra Bladet. It also includes a collection of over 125,000 Danish news articles, complete with titles, abstracts, bodies, and metadata, such as categories. EB-NeRD served as the benchmark dataset for the RecSys '24 Challenge, where it was demonstrated how the dataset can be used to address both technical and normative challenges in designing effective and responsible recommender systems for news publishing. The dataset is available at: https://recsys.eb.dk., Comment: 11 pages, 8 tables, 2 figures, RecSys '24

The RecSys Challenge 2024 aims to advance news recommendation by addressing both the technical and normative challenges inherent in designing effective and responsible recommender systems for news publishing. This paper describes the challenge, including its objectives, problem setting, and the dataset provided by the Danish news publishers Ekstra Bladet and JP/Politikens Media Group ("Ekstra Bladet"). The challenge explores the unique aspects of news recommendation, such as modeling user preferences based on behavior, accounting for the influence of the news agenda on user interests, and managing the rapid decay of news items. Additionally, the challenge embraces normative complexities, investigating the effects of recommender systems on news flow and their alignment with editorial values. We summarize the challenge setup, dataset characteristics, and evaluation metrics. Finally, we announce the winners and highlight their contributions. The dataset is available at: https://recsys.eb.dk., Comment: 5 pages, 3 tables, RecSys' 24

Zero-field optically pumped magnetometers (OPMs) have emerged as an important technology for biomagnetism due to their ulta-sensitive performance, contained within a non-cryogenic small-scale sensor-head. The compactness of such OPMs is often achieved through simplified detection schemes, which typically provide only single-axis magnetic field information. However, multi-axis static magnetic fields on non-measurement axes cause a systematic error that manifests as amplitude and phase errors across the measurement axis. Here we present a triaxial operational technique for a compact zero-field OPM which suppresses multi-axis systematic errors through simultaneous measurement and closed-loop active control of the static magnetic fields across all axes. The demonstrated technique requires magnetic modulation across two axes while providing static field information for all three axes. We demonstrate this technique on a rubidium laboratory-based zero-field magnetometer, achieving a bandwidth of 380 Hz with sensitivities of $<25$ fT/$\sqrt{\rm{Hz}}$ across both transverse axes and $65$ fT/$\sqrt{\rm{Hz}}$ along the beam axis. Using the proposed triaxial technique, we demonstrate precise tracking of a 2 Hz triaxial vector test signal and suppression of systematic cross-axis projection errors over an extended period, $\simeq20$~min.

Score-based models, trained with denoising score matching, are remarkably effective in generating high dimensional data. However, the high variance of their training objective hinders optimisation. We attempt to reduce it with a control variate, derived via a $k$-th order Taylor expansion on the training objective and its gradient. We prove an equivalence between the two and demonstrate empirically the effectiveness of our approach on a low dimensional problem setting; and study its effect on larger problems., Comment: 14 pages, ICML Structured Probabilistic Inference & Generative Modeling 2024

Spatial variation in the intensity of magnetospheric and ionospheric fluctuation during solar storms creates ground-induced currents, of importance in both infrastructure engineering and geophysical science. This activity is currently measured using a network of ground-based magnetometers, typically consisting of extensive installations at established observatory sites. We show that this network can be enhanced by the addition of remote quantum magnetometers which combine high sensitivity with intrinsic calibration. These nodes utilize scalable hardware and run independently of wired communication and power networks. We demonstrate that optically pumped magnetometers, utilizing mass-produced and miniaturized components, offer a single scalable sensor with the sensitivity and stability required for space weather observation. We describe the development and deployment of an off-grid magnetic sensing node, powered by a solar panel, present observed data from periods of low and high geomagnetic activity, and compare it to existing geomagnetic observatories., Comment: 9 pages, 7 figures

The microfabrication of cesium vapor cells for optically pumped magnetometry relies on optimization of buffer gas pressure in order to maximize atomic coherence time and sensitivity to external magnetic signals. We demonstrate post-bond nitrogen buffer gas pressure tuning through localized heating of an integrated micro-pill dispenser. We characterize the variation in the intrinsic longitudinal relaxation rate, $\gamma_{10}$, and magnetic sensitivity, as a function of the resulting nitrogen buffer gas pressure. Measurements are conducted through employing an optically pumped magnetometer operating in a free-induction-decay configuration. $\gamma_{10}$ is extracted across a range of nitrogen pressures between $\sim$~60~-~700~Torr, measuring a minimum of 140~Hz at 115~Torr. Additionally, we achieve sensitivities as low as 130 ~fT/$\sqrt{\text{Hz}}$ at a bias field amplitude of $\sim 50~\mu$T. With the optimal nitrogen buffer gas pressure now quantified and achievable post-fabrication, these mass-producible cells can be tailored to suit a variety of sensing applications, ensuring peak magnetometer performance., Comment: 7 pages (inc references), 3 figures

Bose-Einstein condensates (BECs) have been proposed for many applications in atom interferometry, as their coherence over long evolution times promises unprecedented sensitivity. To date, BECs can be efficiently created in devices using atom chips, but these are still complex and place high demands on size, weight and power. To further simplify these setups, we equipped an atom chip with a nano-structured diffraction-grating to derive all beams for the magneto-optical trap (MOT) from a single laser beam. Moreover, using a 2D$^+$-MOT as an atomic source and a beam with uniform intensity for the grating illumination, we capture $1\times10^9$ atoms in one second, cool them to $14\,\mu$K, and demonstrate magnetic trapping using the atom chip. This is a major step towards the simplification of portable BEC devices for quantum sensing on earth and in space., Comment: 12 pages, 3 figures

Ultracold atoms are crucial for unlocking truly precise and accurate quantum metrology, and provide an essential platform for quantum computing, communication and memories. One of the largest ongoing challenges is the miniaturization of these quantum devices. Here, we show that the typically macroscopic optical lattice architecture at the heart of many ultra-precise quantum technologies can be realized with a single input laser beam on the same diffractive chip already used to create the ultracold atoms. Moreover, this inherently ultra-stable platform enables access to a plethora of new lattice dimensionalities and geometries, ideally suited for the design of high-accuracy, portable quantum devices., Comment: 7 pages, 2 figures

Lasso-type estimators are routinely used to estimate high-dimensional time series models. The theoretical guarantees established for Lasso typically require the penalty level to be chosen in a suitable fashion often depending on unknown population quantities. Furthermore, the resulting estimates and the number of variables retained in the model depend crucially on the chosen penalty level. However, there is currently no theoretically founded guidance for this choice in the context of high-dimensional time series. Instead one resorts to selecting the penalty level in an ad hoc manner using, e.g., information criteria or cross-validation. We resolve this problem by considering estimation of the perhaps most commonly employed multivariate time series model, the linear vector autoregressive (VAR) model, and propose a weighted Lasso estimator with penalization chosen in a fully data-driven way. The theoretical guarantees that we establish for the resulting estimation and prediction error match those currently available for methods based on infeasible choices of penalization. We thus provide a first solution for choosing the penalization in high-dimensional time series models., Comment: 54 pages, 2 figures

Uncertainty Quantification (UQ) is vital to safety-critical model-based analyses, but the widespread adoption of sophisticated UQ methods is limited by technical complexity. In this paper, we introduce UM-Bridge (the UQ and Modeling Bridge), a high-level abstraction and software protocol that facilitates universal interoperability of UQ software with simulation codes. It breaks down the technical complexity of advanced UQ applications and enables separation of concerns between experts. UM-Bridge democratizes UQ by allowing effective interdisciplinary collaboration, accelerating the development of advanced UQ methods, and making it easy to perform UQ analyses from prototype to High Performance Computing (HPC) scale. In addition, we present a library of ready-to-run UQ benchmark problems, all easily accessible through UM-Bridge. These benchmarks support UQ methodology research, enabling reproducible performance comparisons. We demonstrate UM-Bridge with several scientific applications, harnessing HPC resources even using UQ codes not designed with HPC support., Comment: Add Benjamin Kent as co-author in accordance with the paper's published version

Bayesian active learning is based on information theoretical approaches that focus on maximising the information that new observations provide to the model parameters. This is commonly done by maximising the Bayesian Active Learning by Disagreement (BALD) acquisitions function. However, we highlight that it is challenging to estimate BALD when the new data points are subject to censorship, where only clipped values of the targets are observed. To address this, we derive the entropy and the mutual information for censored distributions and derive the BALD objective for active learning in censored regression ($\mathcal{C}$-BALD). We propose a novel modelling approach to estimate the $\mathcal{C}$-BALD objective and use it for active learning in the censored setting. Across a wide range of datasets and models, we demonstrate that $\mathcal{C}$-BALD outperforms other Bayesian active learning methods in censored regression.

In this paper we extend the study of Arrow's generalisation of Black's single-peaked domain and connect this to domains where voting rules satisfy different versions of independence of irrelevant alternatives. First we report on a computational generation of all non-isomorphic Arrow's single-peaked domains on $n\leq 9$ alternatives. Next, we introduce a quantitative measure of richness for domains, as the largest number $r$ such that every alternative is given every rank between 1 and $r$ by the orders in the domain. We investigate the richness of Arrow's single-peaked domains and prove that Black's single-peaked domain has the highest possible richness, but it is not the only domain which attains the maximum. After this we connect Arrow's single-peaked domains to the discussion by Dasgupta, Maskin and others of domains on which plurality and the Borda count satisfy different versions of Independence of Irrelevant alternatives (IIA). For Nash's version of IIA and plurality, it turns out the domains are exactly the duals of Arrow's single-peaked domains. As a consequence there can be at most two alternatives which are ranked first in any such domain. For the Borda count both Arrow's and Nash's versions of IIA lead to a maximum domain size which is exponentially smaller than $2^{n-1}$, the size of Black's single-peaked domain.

Several of the classical results in social choice theory demonstrate that in order for many voting systems to be well-behaved the set domain of individual preferences must satisfy some kind of restriction, such as being single-peaked on a political axis. As a consequence it becomes interesting to measure how diverse the preferences in a well-behaved domain can be. In this paper we introduce an egalitarian approach to measuring preference diversity, focusing on the abundance of distinct suborders one subsets of the alternative. We provide a common generalisation of the frequently used concepts of ampleness and copiousness. We give a detailed investigation of the abundance for Condorcet domains. Our theorems imply a ceiling for the local diversity in domains on large sets of alternatives, which show that in this measure Black's single-peaked domain is in fact optimal. We also demonstrate that for some numbers of alternatives, there are Condorcet domains which have largest local diversity without having maximum order.

This paper delves into applying reinforcement learning (RL) in strategy games, particularly those characterized by parity challenges, as seen in specific positions of Go and Chess and a broader range of impartial games. We propose a simulated learning process, structured within a curriculum learning framework and augmented with noisy labels, to mirror the intricacies of self-play learning scenarios. This approach thoroughly analyses how neural networks (NNs) adapt and evolve from elementary to increasingly complex game positions. Our empirical research indicates that even minimal label noise can significantly impede NNs' ability to discern effective strategies, a difficulty that intensifies with the growing complexity of the game positions. These findings underscore the urgent need for advanced methodologies in RL training, specifically tailored to counter the obstacles imposed by noisy evaluations. The development of such methodologies is crucial not only for enhancing NN proficiency in strategy games with significant parity elements but also for broadening the resilience and efficiency of RL systems across diverse and complex environments.

In this work, we address the problem of assessing and constructing feedback for early-stage writing automatically using machine learning. Early-stage writing is typically vastly different from conventional writing due to phonetic spelling and lack of proper grammar, punctuation, spacing etc. Consequently, early-stage writing is highly non-trivial to analyze using common linguistic metrics. We propose to use sequence-to-sequence models for "translating" early-stage writing by students into "conventional" writing, which allows the translated text to be analyzed using linguistic metrics. Furthermore, we propose a novel robust likelihood to mitigate the effect of noise in the dataset. We investigate the proposed methods using a set of numerical experiments and demonstrate that the conventional text can be predicted with high accuracy., Comment: 9 pages, 1 figure, 1 table, to be published in the proceedings of the Northern Lights Deep Learning Conference 2024

Data clustering, the task of grouping observations according to their similarity, is a key component of unsupervised learning -- with real world applications in diverse fields such as biology, medicine, and social science. Often in these fields the data comes with complex interdependencies between the dimensions of analysis, for instance the various characteristics and opinions people can have live on a complex social network. Current clustering methods are ill-suited to tackle this complexity: deep learning can approximate these dependencies, but not take their explicit map as the input of the analysis. In this paper, we aim at fixing this blind spot in the unsupervised learning literature. We can create network-aware embeddings by estimating the network distance between numeric node attributes via the generalized Euclidean distance. Differently from all methods in the literature that we know of, we do not cluster the nodes of the network, but rather its node attributes. In our experiments we show that having these network embeddings is always beneficial for the learning task; that our method scales to large networks; and that we can actually provide actionable insights in applications in a variety of fields such as marketing, economics, and political science. Our method is fully open source and data and code are available to reproduce all results in the paper.

In this paper, we introduce CDL, a software library designed for the analysis of permutations and linear orders subject to various structural restrictions. Prominent examples of these restrictions include pattern avoidance, a topic of interest in both computer science and combinatorics, and "never conditions" utilized in social choice and voting theory. CDL offers a range of fundamental functionalities, including identifying the permutations that meet specific restrictions and determining the isomorphism of such sets. To facilitate exploration of large permutation sets or domains, CDL incorporates multiple search strategies and heuristics., Comment: 7 pages

The use of Air traffic management (ATM) simulators for planing and operations can be challenging due to their modelling complexity. This paper presents XALM (eXplainable Active Learning Metamodel), a three-step framework integrating active learning and SHAP (SHapley Additive exPlanations) values into simulation metamodels for supporting ATM decision-making. XALM efficiently uncovers hidden relationships among input and output variables in ATM simulators, those usually of interest in policy analysis. Our experiments show XALM's predictive performance comparable to the XGBoost metamodel with fewer simulations. Additionally, XALM exhibits superior explanatory capabilities compared to non-active learning metamodels. Using the `Mercury' (flight and passenger) ATM simulator, XALM is applied to a real-world scenario in Paris Charles de Gaulle airport, extending an arrival manager's range and scope by analysing six variables. This case study illustrates XALM's effectiveness in enhancing simulation interpretability and understanding variable interactions. By addressing computational challenges and improving explainability, XALM complements traditional simulation-based analyses. Lastly, we discuss two practical approaches for reducing the computational burden of the metamodelling further: we introduce a stopping criterion for active learning based on the inherent uncertainty of the metamodel, and we show how the simulations used for the metamodel can be reused across key performance indicators, thus decreasing the overall number of simulations needed.

In this paper, we introduce the class of bipartite peak-pit domains. This is a class of Condorcet domains which include both the classical single-peaked and single-dipped domains. Our class of domains can be used to model situations where some alternatives are ranked based on a most preferred location on a societal axis, and some are ranked based on a least preferred location. This makes it possible to model situations where agents have different rationales for their ranking depending on which of two subclasses of the alternatives one is considering belong to. The class of bipartite peak-pit domains includes most peak-pit domains for $n\leq 7$ alternatives, and the largest Condorcet domains for each $n\leq 8$. In order to study the maximum possible size of a bipartite peak-pit domain we introduce set-alternating schemes. This is a method for constructing well-structured peak-pit domains which are copious and connected. We show that domains based on these schemes always have size at least $2^{n-1}$ and some of them have sizes larger than the domains of Fishburn's alternating scheme. We show that the maximum domain size for sufficiently high $n$ exceeds $2.1973^n$. This improves the previous lower bound for peak-pit domains $2.1890^n$ from \cite{karpov2023constructing}, which was also the highest asymptotic lower bound for the size of the largest Condorcet domains.

Spin preparation prior to a free-induction-decay (FID) measurement can be adversely affected by transverse bias fields, particularly in the geophysical field range. A strategy that enhances the spin polarization accumulated before readout is demonstrated, by synchronizing optical pumping with a magnetic field pulse that supersedes any transverse fields by over two order of magnitude. The pulsed magnetic field is generated along the optical pumping axis using a compact electromagnetic coil pair encompassing a micro-electromechanical systems (MEMS) vapor cell. The coils also resistively heat the cesium (Cs) vapor to the optimal atomic density without spurious magnetic field contributions as they are rapidly demagnetized to approximately zero field during spin readout. The demagnetization process is analyzed electronically, and directly with a FID measurement, to confirm that the residual magnetic field is minimal during detection. The sensitivity performance of this technique is compared to existing optical pumping modalities across a wide magnetic field range. A noise floor sensitivity of $238\,\mathrm{fT/\surd{Hz}}$ was achieved in a field of approximately $\mathrm{50\,\mu{T}}$, in close agreement with the Cram\'{e}r-Rao lower bound (CRLB) predicted noise density of $258\,\mathrm{fT/\surd{Hz}}$., Comment: 10 pages, 7 figures

The ability to learn polynomials and generalize out-of-distribution is essential for simulation metamodels in many disciplines of engineering, where the time step updates are described by polynomials. While feed forward neural networks can fit any function, they cannot generalize out-of-distribution for higher-order polynomials. Therefore, this paper collects and proposes multiplicative neural network (MNN) architectures that are used as recursive building blocks for approximating higher-order polynomials. Our experiments show that MNNs are better than baseline models at generalizing, and their performance in validation is true to their performance in out-of-distribution tests. In addition to MNN architectures, a simulation metamodeling approach is proposed for simulations with polynomial time step updates. For these simulations, simulating a time interval can be performed in fewer steps by increasing the step size, which entails approximating higher-order polynomials. While our approach is compatible with any simulation with polynomial time step updates, a demonstration is shown for an epidemiology simulation model, which also shows the inductive bias in MNNs for learning and generalizing higher-order polynomials.

We report initial research to develop a compact and practical primary thermometer based on Doppler broadening thermometry (DBT). The DBT sensor uses an intrinsic property of thermalized atoms, namely, the Doppler width of a spectral line characteristic of the atoms being probed. The DBT sensor, being founded on a primary thermometry approach, requires no calibration or reference, and so in principle could achieve reliable long-term in-situ thermodynamic temperature measurement. Here we describe our approach and report on initial proof-of-concept investigations with alkali metal vapour cells. Our focus is to develop long-term stable thermometers based on DBT that can be used to reliably measure temperatures for long periods and in environments where sensor retrieval for re-calibration is impractical such as in nuclear waste storage facilities., Comment: 7 pages, 4 figures, referees' comments incorporated

Grating magneto-optical traps are an enabling quantum technology for portable metrological devices with ultracold atoms. However, beam diffraction efficiency and angle are affected by wavelength, creating a single-optic design challenge for laser cooling in two stages at two distinct wavelengths - as commonly used for loading e.g. Sr or Yb atoms into optical lattice or tweezer clocks. Here, we optically characterize a wide variety of binary gratings at different wavelengths to find a simple empirical fit to experimental grating diffraction efficiency data in terms of dimensionless etch depth and period for various duty cycles. The model avoids complex 3D light-grating surface calculations, yet still yields results accurate to a few percent across a broad range of parameters. Gratings optimized for two (or more) wavelengths can now be designed in an informed manner suitable for a wide class of atomic species enabling advanced quantum technologies., Comment: 11 pages, 3 figures, 2 tables. See also 4-page supplementary document

In this paper we give the first explicit enumeration of all maximal Condorcet domains on $n\leq 7$ alternatives. This has been accomplished by developing a new algorithm for constructing Condorcet domains, and an implementation of that algorithm which has been run on a supercomputer. We follow this up by the first survey of the properties of all maximal Condorcet domains up to degree 7, with respect to many properties studied in the social sciences and mathematical literature. We resolve several open questions posed by other authors, both by examples from our data and theorems. We give a new set of results on the symmetry properties of Condorcet domains which unify earlier works. Finally we discuss connections to other domain types such as non-dictatorial domains and generalisations of single-peaked domains. All our data is made freely available for other researches via a new website., Comment: Typos corrected, some open problems, a conjecture, and a new figure has been added in the latest version