Search

This chapter explores anomaly localization in medical images using denoising diffusion models. After providing a brief methodological background of these models, including their application to image reconstruction and their conditioning using guidance mechanisms, we provide an overview of available datasets and evaluation metrics suitable for their application to anomaly localization in medical images. In this context, we discuss supervision schemes ranging from fully supervised segmentation to semi-supervised, weakly supervised, self-supervised, and unsupervised methods, and provide insights into the effectiveness and limitations of these approaches. Furthermore, we highlight open challenges in anomaly localization, including detection bias, domain shift, computational cost, and model interpretability. Our goal is to provide an overview of the current state of the art in the field, outline research gaps, and highlight the potential of diffusion models for robust anomaly localization in medical images.

Electrochemical biosensors ("E-AB" or "E-DNA" type sensors) that utilize square-wave voltammetry originated in academic labs with a few standard experimental configurations for the electrochemical cell and data analysis. We report here on adaptations of these approaches that are friendly to novice scientists such as those in undergraduate laboratories. These approaches utilize commercially available components, low volumes, work over extended periods and enable facile analysis using a custom excel sheet.

Kagome materials are known to be an ideal platform that hosts a plethora of interesting phases such as topological states, electronic correlation, and magnetism, owing to their unique band structure and geometry. We report magnetotransport measurement in Kagome metal Yb$_{0.5}$Co_3Ge$_3$ as a function of pressure. Below $\sim25^\circ$ K the temperature dependence of resistance shows an upturn that is accompanied by a strong negative magnetoresistance, which could be attributed to Kondo effect. Upon pressurization above 1 GPa the resistance shows a reduction as a function of temperature below $4^\circ$ K, along with a further enhanced negative magnetoresistance. This might indicate an onset of a pressure-induced Kondo coherence effect.

Reinforcement learning utilizing kernel ridge regression to predict the expected value function represents a powerful method with great representational capacity. This setting is a highly versatile framework amenable to analytical results. We consider kernel-based function approximation for RL in the infinite horizon average reward setting, also referred to as the undiscounted setting. We propose an optimistic algorithm, similar to acquisition function based algorithms in the special case of bandits. We establish novel no-regret performance guarantees for our algorithm, under kernel-based modelling assumptions. Additionally, we derive a novel confidence interval for the kernel-based prediction of the expected value function, applicable across various RL problems., Comment: 38th Conference on Neural Information Processing Systems (NeurIPS 2024)

We extend the definition of a one-sided identity and a one-sided zero with respect to an entire set with a binary operation to subsets of a set with a binary operation. Given a set $S$ with a binary operation and a subset $A$ of $S$, $lidentity(A)$ is the set of all $b \in S$ such that $ba=a$ for all $a \in A$. There is a similar definition for $ridentity(A)$. Also given a set $S$ with a binary operation and a subset $A$ of $S$, $lzero(A)$ is the set of all $b \in S$ such that $ba=b$ for all $a \in A$. There is a similar definition for $rzero(A)$. In any semigroup or groupoid, every one-sided identity or zero is an idempotent, and every idempotent is a one-sided identity or zero for some subsemigroup or subgroupoid. In particular, every one-sided identity or zero of a maximal subsemigroup of a specific type is an idempotent, and every idempotent is a one-sided identity or zero for some maximal subsemigroup of a specific type. These types include groups, left [right] zero semigroups, and left [right] groups. It has been shown that every idempotent of a semigroup is the identity of a maximal subgroup of that semigroup. We show that every idempotent of a semigroup exists in a maximal left [right] zero subsemigroup. For each idempotent of a semigroup, we describe all rectangular band subsemigroups containing that idempotent. We also show that every idempotent of a semigroup exists in a maximal right [left] subgroup.

Supervised learning has become a cornerstone of modern machine learning, yet a comprehensive theory explaining its effectiveness remains elusive. Empirical phenomena, such as neural analogy-making and the linear representation hypothesis, suggest that supervised models can learn interpretable factors of variation in a linear fashion. Recent advances in self-supervised learning, particularly nonlinear Independent Component Analysis, have shown that these methods can recover latent structures by inverting the data generating process. We extend these identifiability results to parametric instance discrimination, then show how insights transfer to the ubiquitous setting of supervised learning with cross-entropy minimization. We prove that even in standard classification tasks, models learn representations of ground-truth factors of variation up to a linear transformation. We corroborate our theoretical contribution with a series of empirical studies. First, using simulated data matching our theoretical assumptions, we demonstrate successful disentanglement of latent factors. Second, we show that on DisLib, a widely-used disentanglement benchmark, simple classification tasks recover latent structures up to linear transformations. Finally, we reveal that models trained on ImageNet encode representations that permit linear decoding of proxy factors of variation. Together, our theoretical findings and experiments offer a compelling explanation for recent observations of linear representations, such as superposition in neural networks. This work takes a significant step toward a cohesive theory that accounts for the unreasonable effectiveness of supervised deep learning.

We study the implicit bias of the general family of steepest descent algorithms, which includes gradient descent, sign descent and coordinate descent, in deep homogeneous neural networks. We prove that an algorithm-dependent geometric margin starts increasing once the networks reach perfect training accuracy and characterize the late-stage bias of the algorithms. In particular, we define a generalized notion of stationarity for optimization problems and show that the algorithms progressively reduce a (generalized) Bregman divergence, which quantifies proximity to such stationary points of a margin-maximization problem. We then experimentally zoom into the trajectories of neural networks optimized with various steepest descent algorithms, highlighting connections to the implicit bias of Adam.

This paper studies the properties of solutions to multi-task shallow ReLU neural network learning problems, wherein the network is trained to fit a dataset with minimal sum of squared weights. Remarkably, the solutions learned for each individual task resemble those obtained by solving a kernel method, revealing a novel connection between neural networks and kernel methods. It is known that single-task neural network training problems are equivalent to minimum norm interpolation problem in a non-Hilbertian Banach space, and that the solutions of such problems are generally non-unique. In contrast, we prove that the solutions to univariate-input, multi-task neural network interpolation problems are almost always unique, and coincide with the solution to a minimum-norm interpolation problem in a Sobolev (Reproducing Kernel) Hilbert Space. We also demonstrate a similar phenomenon in the multivariate-input case; specifically, we show that neural network learning problems with large numbers of diverse tasks are approximately equivalent to an $\ell^2$ (Hilbert space) minimization problem over a fixed kernel determined by the optimal neurons.

Text-to-image (T2I) diffusion models are popular for introducing image manipulation methods, such as editing, image fusion, inpainting, etc. At the same time, image-to-video (I2V) and text-to-video (T2V) models are also built on top of T2I models. We present Kandinsky 3, a novel T2I model based on latent diffusion, achieving a high level of quality and photorealism. The key feature of the new architecture is the simplicity and efficiency of its adaptation for many types of generation tasks. We extend the base T2I model for various applications and create a multifunctional generation system that includes text-guided inpainting/outpainting, image fusion, text-image fusion, image variations generation, I2V and T2V generation. We also present a distilled version of the T2I model, evaluating inference in 4 steps of the reverse process without reducing image quality and 3 times faster than the base model. We deployed a user-friendly demo system in which all the features can be tested in the public domain. Additionally, we released the source code and checkpoints for the Kandinsky 3 and extended models. Human evaluations show that Kandinsky 3 demonstrates one of the highest quality scores among open source generation systems., Comment: Accepted for EMNLP 2024 (Demo track)

Efficiently processing structured point cloud data while preserving multiscale information is a key challenge across domains, from graphics to atomistic modeling. Using a curated dataset of simulated galaxy positions and properties, represented as point clouds, we benchmark the ability of graph neural networks to simultaneously capture local clustering environments and long-range correlations. Given the homogeneous and isotropic nature of the Universe, the data exhibits a high degree of symmetry. We therefore focus on evaluating the performance of Euclidean symmetry-preserving ($E(3)$-equivariant) graph neural networks, showing that they can outperform non-equivariant counterparts and domain-specific information extraction techniques in downstream performance as well as simulation-efficiency. However, we find that current architectures fail to capture information from long-range correlations as effectively as domain-specific baselines, motivating future work on architectures better suited for extracting long-range information., Comment: 19 pages, 3 figures; To appear at the NeurReps Workshop @ NeurIPS 2024

By analyzing the spectral energy distributions (SEDs) of resolved stars in nearby galaxies, we can constrain their stellar properties and line-of-sight dust extinction. From the Scylla survey, we obtain ultraviolet to near-infrared photometry from Wide Field Camera 3 onboard the {\it Hubble Space Telescope} for more than 1.5 million stars in the SMC and LMC. We use the Bayesian Extinction and Stellar Tool (BEAST) to analyze the multi-band SEDs of these sources and characterize their initial masses, ages, metallicities, distances, and line-of-sight extinction properties (e.g.~$A_V$, $R_V$). We apply quality cuts and perform validation simulations to construct a catalog of over 550,000 stars with high-reliability SED fits, which we use to analyze the stellar content and extinction properties of the SMC and LMC. We detect stars with masses as low as 0.6 $M_{\odot}$. Observed stellar age distributions show a jump in stars around 6 Gyrs ago, which is in agreement with other star-formation histories. Extinctions ($A_V$) in both galaxies follow a log-normal distribution. We compare $A_V$ with ancillary gas and dust tracers like $HI$, $H_\alpha$, and far infrared (FIR) dust emission and find positive correlations on a field-by-field basis. We convert observed $A_V$ to predicted dust surface densities using the Draine et. al. (2014) model and find $A_V$-based dust surface densities are a factor of $\sim$2.5 lower than observed FIR-based dust surface densities, a correction factor similar to other studies., Comment: Submitted to ApJ, 31 pages

Concept-based machine learning methods have increasingly gained importance due to the growing interest in making neural networks interpretable. However, concept annotations are generally challenging to obtain, making it crucial to leverage all their prior knowledge. By creating concept-enriched models that incorporate concept information into existing architectures, we exploit their interpretable capabilities to the fullest extent. In particular, we propose Concept-Guided Conditional Diffusion, which can generate visual representations of concepts, and Concept-Guided Prototype Networks, which can create a concept prototype dataset and leverage it to perform interpretable concept prediction. These results open up new lines of research by exploiting pre-existing information in the quest for rendering machine learning more human-understandable., Comment: Interpretable AI: Past, Present and Future Workshop at NeurIPS 2024

We develop a theoretical model to investigate wave propagation in media with random time-varying properties, where temporal fluctuations lead to complex scattering dynamics. Focusing on the ensemble-averaged field, we derive an exact expression for the average Green's function in the presence of finite temporal disorder, and extend our analysis to the thermodynamic limit. In contrast to spatial disorder, causality prevents recurrent scattering, allowing us to achieve a non-perturbative solution. We identify two distinct propagation regimes: one where the field oscillates with constant amplitude, and another where it grows exponentially at long times. Our findings offer new insights into wave dynamics in temporally disordered media, with potential applications in time-varying metamaterials, dynamic sensing, and imaging in turbulent or chaotic environments.

Concept Activation Vectors (CAVs) offer insights into neural network decision-making by linking human friendly concepts to the model's internal feature extraction process. However, when a new set of CAVs is discovered, they must still be translated into a human understandable description. For image-based neural networks, this is typically done by visualizing the most relevant images of a CAV, while the determination of the concept is left to humans. In this work, we introduce an approach to aid the interpretation of newly discovered concept sets by suggesting textual descriptions for each CAV. This is done by mapping the most relevant images representing a CAV into a text-image embedding where a joint description of these relevant images can be computed. We propose utilizing the most relevant receptive fields instead of full images encoded. We demonstrate the capabilities of this approach in multiple experiments with and without given CAV labels, showing that the proposed approach provides accurate descriptions for the CAVs and reduces the challenge of concept interpretation., Comment: 19 pages, 7 figures, to be published in ICPR

Users can divulge sensitive information to proprietary LLM providers, raising significant privacy concerns. While open-source models, hosted locally on the user's machine, alleviate some concerns, models that users can host locally are often less capable than proprietary frontier models. Toward preserving user privacy while retaining the best quality, we propose Privacy-Conscious Delegation, a novel task for chaining API-based and local models. We utilize recent public collections of user-LLM interactions to construct a natural benchmark called PUPA, which contains personally identifiable information (PII). To study potential approaches, we devise PAPILLON, a multi-stage LLM pipeline that uses prompt optimization to address a simpler version of our task. Our best pipeline maintains high response quality for 85.5% of user queries while restricting privacy leakage to only 7.5%. We still leave a large margin to the generation quality of proprietary LLMs for future work. Our data and code will be available at https://github.com/siyan-sylvia-li/PAPILLON.

Finding clusters of data points with similar characteristics and generating new cluster-specific samples can significantly enhance our understanding of complex data distributions. While clustering has been widely explored using Variational Autoencoders, these models often lack generation quality in real-world datasets. This paper addresses this gap by introducing TreeDiffusion, a deep generative model that conditions Diffusion Models on hierarchical clusters to obtain high-quality, cluster-specific generations. The proposed pipeline consists of two steps: a VAE-based clustering model that learns the hierarchical structure of the data, and a conditional diffusion model that generates realistic images for each cluster. We propose this two-stage process to ensure that the generated samples remain representative of their respective clusters and enhance image fidelity to the level of diffusion models. A key strength of our method is its ability to create images for each cluster, providing better visualization of the learned representations by the clustering model, as demonstrated through qualitative results. This method effectively addresses the generative limitations of VAE-based approaches while preserving their clustering performance. Empirically, we demonstrate that conditioning diffusion models on hierarchical clusters significantly enhances generative performance, thereby advancing the state of generative clustering models., Comment: 25 pages, submitted to ICLR 2025

Traditional panel data causal inference frameworks, such as difference-in-differences and synthetic control methods, rely on pre-intervention data to estimate counterfactuals, which may not be available in real-world settings when interventions are implemented in response to sudden events. In this paper, we introduce two data fusion methods for causal inference from panel data in scenarios where pre-intervention data is unavailable. These methods leverage auxiliary reference domains with related panel data to estimate causal effects in the target domain, overcoming the limitations imposed by the absence of pre-intervention data. We show the efficacy of these methods by obtaining converging bounds on the absolute bias as well as through simulations, showing their robustness in a variety of panel data settings. Our findings provide a framework for applying causal inference in urgent and data-constrained environments, such as public health crises or epidemiological shocks.

Regularization, whether explicit in terms of a penalty in the loss or implicit in the choice of algorithm, is a cornerstone of modern machine learning. Indeed, controlling the complexity of the model class is particularly important when data is scarce, noisy or contaminated, as it translates a statistical belief on the underlying structure of the data. This work investigates the question of how to choose the regularization norm $\lVert \cdot \rVert$ in the context of high-dimensional adversarial training for binary classification. To this end, we first derive an exact asymptotic description of the robust, regularized empirical risk minimizer for various types of adversarial attacks and regularization norms (including non-$\ell_p$ norms). We complement this analysis with a uniform convergence analysis, deriving bounds on the Rademacher Complexity for this class of problems. Leveraging our theoretical results, we quantitatively characterize the relationship between perturbation size and the optimal choice of $\lVert \cdot \rVert$, confirming the intuition that, in the data scarce regime, the type of regularization becomes increasingly important for adversarial training as perturbations grow in size.

Mobile manipulators are finding use in numerous practical applications. The current issues with mobile manipulation are the large state space owing to the mobile base and the challenge of modeling high degree of freedom systems. It is critical to devise fast and accurate algorithms that generate smooth motion plans for such mobile manipulators. Existing techniques attempt to solve this problem but focus on separating the motion of the base and manipulator. We propose an approach using Lie theory to find the inverse kinematic constraints by converting the kinematic model, created using screw coordinates, between its Lie group and vector representation. An optimization function is devised to solve for the desired joint states of the entire mobile manipulator. This allows the motion of the mobile base and manipulator to be planned and applied in unison resulting in a smooth and accurate motion plan. The performance of the proposed state planner is validated on simulated mobile manipulators in an analytical experiment. Our solver is available with further derivations and results at https://github.com/peleito/slithers., Comment: 8 pages, 9 figures, conference submission

Generative audio models are rapidly advancing in both capabilities and public utilization -- several powerful generative audio models have readily available open weights, and some tech companies have released high quality generative audio products. Yet, while prior work has enumerated many ethical issues stemming from the data on which generative visual and textual models have been trained, we have little understanding of similar issues with generative audio datasets, including those related to bias, toxicity, and intellectual property. To bridge this gap, we conducted a literature review of hundreds of audio datasets and selected seven of the most prominent to audit in more detail. We found that these datasets are biased against women, contain toxic stereotypes about marginalized communities, and contain significant amounts of copyrighted work. To enable artists to see if they are in popular audio datasets and facilitate exploration of the contents of these datasets, we developed a web tool audio datasets exploration tool at https://audio-audit.vercel.app.

End-Stage Liver Disease (ESLD), a complex condition, has high rates of co-occurring comorbidities affecting multiple organ systems. There is no clear evidence-based practice (EBP) guidelines addressing the progression of comorbidities in ESLD patients awaiting liver transplantation (LT) and their impact on survival, both pre- and post-transplant. This study aimed to identify and quantify the trajectory of the most common and deteriorating comorbidities in ESLD patients awaiting LT and to analyze their effect on patient outcomes. An initial exploratory phase to identify frequent comorbidities in ESLD patients. Relevant EBP-driven data for diagnosing and measuring the progression of these conditions were collected and organized into five research matrices. In the quantitative phase, a retrospective analysis was conducted using longitudinal de-identified data from electronic health records (EHR) for patients who underwent LT between 2011-2021. Data included demographics, labs, procedures, and medications. Descriptive statistics and survival analysis assessed the association of comorbidities with post-transplant survival. The five most frequent comorbidities identified were Diabetes Mellitus (DM), Chronic Kidney Disease (CKD), Malnutrition, Portal Hypertension (PH), and Ascites. Of the 722 patients analyzed, 68.2% were male, mean age of 54.81 (SD = 11.24), and 19.8% died post-LT. Survival analysis showed a gradual decline over time, with the most significant drop at five years post-LT. Significant predictors of post-LT survival were age at transplant (p=0.01), waitlist time (p=0.004), DM at listing (p=0.02), low albumin (p=0.03), and CKD stage 5 development after listing (p=0.04). This study highlights the variability in diagnosing and measuring comorbidities in ESLD patients and provides insights into their progression and impact on post-LT survival., Comment: 46 pages

Scylla is a deep Hubble Space Telescope survey of the stellar populations, interstellar medium and star formation in the LMC and SMC. As a pure-parallel complement to the Ultraviolet Legacy Library of Young Stars as Essential Standards (ULLYSES) survey, Scylla obtained 342 orbits of ultraviolet (UV) through near-infrared (IR) imaging of the LMC and SMC with Wide Field Camera 3. In this paper, we describe the science objectives, observing strategy, data reduction procedure, and initial results from our photometric analysis of 96 observed fields. Although our observations were constrained by ULYSSES primary exposures, we imaged all fields in at least two filters (F475W and F814W), and 64% of fields in at least three and as many as seven WFC3 filters spanning the UV to IR. Overall, we reach average 50% completeness of $m_{\rm F225W}=26.0$, $m_{\rm F275W}=26.2$, $m_{\rm F336W}=26.9$, $m_{\rm F475W}=27.8$, $m_{\rm F814W}=25.5$, $m_{\rm F110W}=24.7$, and $m_{\rm F160W}=24.0$ Vega magnitudes in our photometric catalogs, which is faintward of the ancient main sequence turnoff in all filters. The primary science goals of Scylla include characterizing the structure and properties of dust in the MCs, as well as their spatially-resolved star formation and chemical enrichment histories. Our images and photometric catalogs, which represent the widest-area coverage of MCs with HST photometry to date, are available as a high-level science product at the Barbara A. Mikulski Archive for Space Telescopes., Comment: 25 pages, 16 figures, 8 tables. Accepted for publication in ApJS

This paper investigates how the topical flow of dyadic conversations emerges over time and how differences in interlocutors' personality traits contribute to this topical flow. Leveraging text embeddings, we map the trajectories of $N = 1655$ conversations between strangers into a high-dimensional space. Using nonlinear projections and clustering, we then identify when each interlocutor enters and exits various topics. Differences in conversational flow are quantified via $\textit{topic entropy}$, a summary measure of the "spread" of topics covered during a conversation, and $\textit{linguistic alignment}$, a time-varying measure of the cosine similarity between interlocutors' embeddings. Our findings suggest that interlocutors with a larger difference in the personality dimension of openness influence each other to spend more time discussing a wider range of topics and that interlocutors with a larger difference in extraversion experience a larger decrease in linguistic alignment throughout their conversation. We also examine how participants' affect (emotion) changes from before to after a conversation, finding that a larger difference in extraversion predicts a larger difference in affect change and that a greater topic entropy predicts a larger affect increase. This work demonstrates how communication research can be advanced through the use of high-dimensional NLP methods and identifies personality difference as an important driver of social influence., Comment: To be published in the Proceedings of the Second Workshop on Social Influence in Conversations (SICon 2024), co-located with EMNLP 2024

Massive stars mainly form in close binaries, where their mutual interactions can profoundly alter their evolutionary paths. Evolved binaries consisting of a massive OB-type main-sequence star with a stripped helium star or a compact companion represent a crucial stage in the evolution towards double compact objects, whose mergers are (potentially) detectable via gravitational waves. The recent detection of X-ray quiet OB+black hole binaries and OB+stripped helium star binaries has set the stage for discovering more of these systems in the near future. In this work, based on 3670 detailed binary-evolution models and using empirical distributions of initial binary parameters, we compute the expected population of such evolved massive binaries in coeval stellar populations, including stars in star clusters and in galaxies with starburst activities, for ages up to 100 Myr. Our results are vividly illustrated in an animation that shows the evolution of these binaries in the color-magnitude diagram over time. We find that the number of OB+black hole binaries peaks around 10 Myr, and OB+neutron star binaries are most abundant at approximately 20 Myr. Both black holes and neutron stars can potentially be found in populations with ages up to 90 Myr. Additionally, we analyze the properties of such binaries at specific ages. We find that OB+helium stars and OB+black hole binaries are likely to be identifiable as single-lined spectroscopic binaries. Our research serves as a guide for future observational efforts to discover such binaries in young star clusters and starburst environments., Comment: 32 pages, 15 figures. Accepted for Publication in Astrophysical Journal Letters (ApJL)

In this article, we introduce and study capacities related to nonlocal Sobolev spaces, with focus on spaces corresponding to zero-order nonlocal operators. In particular, we prove Hardy-type inequalities to obtain Sobolev embeddings and use them to estimate the nonlocal capacities of a ball., Comment: 15 pages

We introduce the condensed fiber product of two $G$-crossed braided fusion categories, generalizing existing constructions in the literature. We show that this product is closely related to the cohomological construction known as zesting. Furthermore, the condensed fiber product defines a monoidal structure on the 2-category of $G$-crossed and braided extensions of braided categories with a fixed transparent symmetric pointed subcategory.

The growing use of Machine Learning and Artificial Intelligence (AI), particularly Large Language Models (LLMs) like OpenAI's GPT series, leads to disruptive changes across organizations. At the same time, there is a growing concern about how organizations handle personal data. Thus, privacy policies are essential for transparency in data processing practices, enabling users to assess privacy risks. However, these policies are often long and complex. This might lead to user confusion and consent fatigue, where users accept data practices against their interests, and abusive or unfair practices might go unnoticed. LLMss can be used to assess privacy policies for users automatically. In this interdisciplinary work, we explore the challenges of this approach in three pillars, namely technical feasibility, ethical implications, and legal compatibility of using LLMs to assess privacy policies. Our findings aim to identify potential for future research, and to foster a discussion on the use of LLM technologies for enabling users to fulfil their important role as decision-makers in a constantly developing AI-driven digital economy., Comment: Published at AISyS 2024

The Calar Alto Void Integral-field Treasury surveY (CAVITY) is a legacy project aimed at characterising the population of galaxies inhabiting voids, which are the most under-dense regions of the cosmic web, located in the Local Universe. This paper describes the first public data release (DR1) of CAVITY, comprising science-grade optical data cubes for the initial 100 out of a total of $\sim$300 galaxies in the Local Universe ($0.005 < z < 0.050$). These data were acquired using the integral-field spectrograph PMAS/PPak mounted on the 3.5m telescope at the Calar Alto observatory. The DR1 galaxy sample encompasses diverse characteristics in the color-magnitude space, morphological type, stellar mass, and gas ionisation conditions, providing a rich resource for addressing key questions in galaxy evolution through spatially resolved spectroscopy. The galaxies in this study were observed with the low-resolution V500 set-up, spanning the wavelength range 3745-7500 \AA, with a spectral resolution of 6.0 \AA\ (FWHM). Here, we describe the data reduction and characteristics and data structure of the CAVITY datasets essential for their scientific utilisation, highlighting such concerns as vignetting effects, as well as the identification of bad pixels and management of spatially correlated noise. We also provide instructions for accessing the CAVITY datasets and associated ancillary data through the project's dedicated database., Comment: 17 pages, 9 figures, 5 tables. Accepted for publication in A&A (replacement after minor language edits)

The structure of many real-world datasets is intrinsically hierarchical, making the modeling of such hierarchies a critical objective in both unsupervised and supervised machine learning. Recently, novel approaches for hierarchical clustering with deep architectures have been proposed. In this work, we take a critical perspective on this line of research and demonstrate that many approaches exhibit major limitations when applied to realistic datasets, partly due to their high computational complexity. In particular, we show that a lightweight procedure implemented on top of pre-trained non-hierarchical clustering models outperforms models designed specifically for hierarchical clustering. Our proposed approach is computationally efficient and applicable to any pre-trained clustering model that outputs logits, without requiring any fine-tuning. To highlight the generality of our findings, we illustrate how our method can also be applied in a supervised setup, recovering meaningful hierarchies from a pre-trained ImageNet classifier.

One of the most challenging problems in the computational study of localization in quantum manybody systems is to capture the effects of rare events, which requires sampling over exponentially many disorder realizations. We implement an efficient procedure on a quantum processor, leveraging quantum parallelism, to efficiently sample over all disorder realizations. We observe localization without disorder in quantum many-body dynamics in one and two dimensions: perturbations do not diffuse even though both the generator of evolution and the initial states are fully translationally invariant. The disorder strength as well as its density can be readily tuned using the initial state. Furthermore, we demonstrate the versatility of our platform by measuring Renyi entropies. Our method could also be extended to higher moments of the physical observables and disorder learning.

We study the performance of transformers as a function of the number of repetitions of training examples with algorithmically generated datasets. On three problems of mathematics: the greatest common divisor, modular multiplication, and matrix eigenvalues, we show that for a fixed number of training steps, models trained on smaller sets of repeated examples outperform models trained on larger sets of single-use examples. We also demonstrate that two-set training - repeated use of a small random subset of examples, along normal sampling on the rest of the training set - provides for faster learning and better performance. This highlights that the benefits of repetition can outweigh those of data diversity. These datasets and problems provide a controlled setting to shed light on the still poorly understood interplay between generalization and memorization in deep learning.

Scientific research often seeks to understand the causal structure underlying high-level variables in a system. For example, climate scientists study how phenomena, such as El Ni\~no, affect other climate processes at remote locations across the globe. However, scientists typically collect low-level measurements, such as geographically distributed temperature readings. From these, one needs to learn both a mapping to causally-relevant latent variables, such as a high-level representation of the El Ni\~no phenomenon and other processes, as well as the causal model over them. The challenge is that this task, called causal representation learning, is highly underdetermined from observational data alone, requiring other constraints during learning to resolve the indeterminacies. In this work, we consider a temporal model with a sparsity assumption, namely single-parent decoding: each observed low-level variable is only affected by a single latent variable. Such an assumption is reasonable in many scientific applications that require finding groups of low-level variables, such as extracting regions from geographically gridded measurement data in climate research or capturing brain regions from neural activity data. We demonstrate the identifiability of the resulting model and propose a differentiable method, Causal Discovery with Single-parent Decoding (CDSD), that simultaneously learns the underlying latents and a causal graph over them. We assess the validity of our theoretical results using simulated data and showcase the practical validity of our method in an application to real-world data from the climate science field., Comment: 33 pages, 17 figures

We present a novel method to characterize dust depletion, namely, the depletion of metals into dust grains. We used observed correlations among relative abundances combining a total of 17 metals in diverse galactic environments, including the Milky Way (MW), Large Magellanic Cloud (LMC), Small Magellanic Cloud (SMC), and damped Lyman-$\alpha$ absorbers (DLAs) towards quasars and gamma-ray bursts (GRBs). We only considered the relative abundances of metals that qualify as tracers of dust and we used all available dust tracers. We find linear correlations among all studied dust tracers in a multidimensional space, where each dimension corresponds to an individual dust tracer. The fit to the linear correlations among the dust tracers describes the tendencies of different elements when depleting into dust grains. We determined the overall strength of dust depletion, $\Delta$, along individual lines of sight, based on the correlations among different dust tracers. We avoided any preference for specific dust tracers or any other assumptions by including all available dust tracers in this multidimensional space. We also determined the dust depletion of Kr, C, O, Cl, P, Zn, Ge, Mg, Cu, Si, Fe, Ni, and Ti. Finally, we offer simple guidelines for the application of the method to the study of the observed patterns of abundances and relative abundances. This has allowed for a straightforward determination of the overall strength of depletion and the dust depletion of individual elements. We also obtained an estimate for the gas-phase metallicity and identified any additional deviations due to the nucleosynthesis of specific stellar populations. Thus, we have established a unified methodology for characterizing dust depletion across cosmic time and diverse galactic environments, offering a valuable new approach to the study of dust depletion in studies of the chemical evolution of galaxies., Comment: Accepted for publication in A&A

As artificial intelligence (AI) continues to evolve, the current paradigm of treating AI as a passive tool no longer suffices. As a human-AI team, we together advocate for a shift toward viewing AI as a learning partner, akin to a student who learns from interactions with humans. Drawing from interdisciplinary concepts such as ecorithms, order from chaos, and cooperation, we explore how AI can evolve and adapt in unpredictable environments. Arising from these brief explorations, we present two key recommendations: (1) foster ethical, cooperative treatment of AI to benefit both humans and AI, and (2) leverage the inherent heterogeneity between human and AI minds to create a synergistic hybrid intelligence. By reframing AI as a dynamic partner, a model emerges in which AI systems develop alongside humans, learning from human interactions and feedback loops including reflections on team conversations. Drawing from a transpersonal and interdependent approach to consciousness, we suggest that a "third mind" emerges through collaborative human-AI relationships. Through design interventions such as interactive learning and conversational debriefing and foundational interventions allowing AI to model multiple types of minds, we hope to provide a path toward more adaptive, ethical, and emotionally healthy human-AI relationships. We believe this dynamic relational learning-partner (DRLP) model for human-AI teaming, if enacted carefully, will improve our capacity to address powerful solutions to seemingly intractable problems., Comment: White Paper

Within the scaling laws paradigm, which underpins the training of large neural networks like ChatGPT and Llama, we consider a supervised regression setting and establish the existance of a strong form of the model collapse phenomenon, a critical performance degradation due to synthetic data in the training corpus. Our results show that even the smallest fraction of synthetic data (e.g., as little as 1\% of the total training dataset) can still lead to model collapse: larger and larger training sets do not enhance performance. We further investigate whether increasing model size, an approach aligned with current trends in training large language models, exacerbates or mitigates model collapse. In a simplified regime where neural networks are approximated via random projections of tunable size, we both theoretically and empirically show that larger models can amplify model collapse. Interestingly, our theory also indicates that, beyond the interpolation threshold (which can be extremely high for very large datasets), larger models may mitigate the collapse, although they do not entirely prevent it. Our theoretical findings are empirically verified through experiments on language models and feed-forward neural networks for images.

Environmentally-benign, non-toxic electrolytes with combinatorial design spaces are excellent candidates for green solvents, green leaching agents, and carbon capture sources. Here, we examine one particular green solvent, ethaline, a 2:1 molar ratio of ethylene glycol and choline chloride. Despite its touted green credentials, we find partial decomposition of ethaline into toxic chloromethane and dimethylaminoethanol at room temperature, limiting its sustainable advantage. We experimentally characterize these decomposition products and computationally develop a general, quantum chemically-accurate workflow to understand decomposition. We find that fluctuations of the hydrogen bonds bind chloride near reaction sites, initiating the reaction between choline cations and chloride anions. In summary, in the design of green solvents, we do not recommend the use of choline chloride due to its susceptibility to undergo decomposition in strongly hydrogen-bound mixtures.

Dual active galactic nuclei (AGN) offer a unique opportunity to probe the relationship between super massive black holes (SMBH) and their host galaxies as well as the role of major mergers in triggering AGN activity. The confirmed dual AGN Mrk 266 has been studied extensively with multi-wavelength imaging. Now, high spatial resolution IFU spectroscopy of Mrk 266 provides an opportunity to probe the kinematics of both the merger event and AGN feedback. We present for the first time high spatial resolution kinematic maps for both nuclei of Mrk 266 obtained with the Keck OSIRIS IFU spectrograph, utilizing adaptive optics to achieve a resolution of 0.31" and 0.20" for the NE and SW nuclei, respectively. Using the M-sigma relation for mergers, we infer a SMBH mass of approximately 7e7 solar masses for the southwestern nucleus. Additionally, we report that the molecular gas kinematics of the southwestern nucleus are dominated by rotation rather than large-scale chaotic motions. The southwest nucleus also contains both a circumnuclear ring of star formation from which an inflow of molecular gas is likely fueling the AGN and a compact, AGN-dominated outflow of highly ionized gas with a timescale of approximately 2 Myr, significantly shorter than the timescale of the merger. The northeastern nucleus, on the other hand, exhibits complex kinematics related to the merger, including molecular gas that appears to have decoupled from the rotation of the stars. Our results suggest that while the AGN activity in Mrk 266 was likely triggered during the merger, AGN feeding is currently the result of processes internal to each host galaxy, thus resulting in a strong asymmetry between the two nuclei., Comment: 18 pages, 18 figures, 3 tables. Accepted for publication in MNRAS

Solution-based soft matter self-assembly (SA) promises unique materials properties from approaches including additive manufacturing/three-dimensional (3D) printing. We report direct ink writing derived, hierarchically porous transition metal nitride superconductors (SCs) and precursor oxides, structure-directed by Pluronics-family block copolymer (BCP) SA and heat treated in various environments. SCs with periodic lattices on three length scales show record nanoconfinement-induced upper critical field enhancements correlated with BCP molar mass, attaining values of 50 T for NbN and 8.1 T for non-optimized TiN samples, the first mapping of a tailorable SC property onto a macromolecular parameter. They reach surface areas above 120 m$^2$/g, the highest reported for compound SCs to date. Embedded printing enables the first BCP directed mesoporous non-self-supporting helical SCs. Results suggest that additive manufacturing may open pathways to mesoporous SCs with not only a variety of macroscopic form factors but enhanced properties from intrinsic, SA-derived mesostructures with substantial academic and technological promise.

This work introduces TrajDiffuser, a compositional diffusion-based flexible and concurrent trajectory generator for 6 degrees of freedom powered descent guidance. TrajDiffuser is a statistical model that learns the multi-modal distributions of a dataset of simulated optimal trajectories, each subject to only one or few constraints that may vary for different trajectories. During inference, the trajectory is generated simultaneously over time, providing stable long-horizon planning, and constraints can be composed together, increasing the model's generalizability and decreasing the training data required. The generated trajectory is then used to initialize an optimizer, increasing its robustness and speed., Comment: Full manuscript submitted to IEEE Aerospace 2025 on 4-Oct-2024

Self-anthropomorphism in robots manifests itself through their display of human-like characteristics in dialogue, such as expressing preferences and emotions. Our study systematically analyzes self-anthropomorphic expression within various dialogue datasets, outlining the contrasts between self-anthropomorphic and non-self-anthropomorphic responses in dialogue systems. We show significant differences in these two types of responses and propose transitioning from one type to the other. We also introduce Pix2Persona, a novel dataset aimed at developing ethical and engaging AI systems in various embodiments. This dataset preserves the original dialogues from existing corpora and enhances them with paired responses: self-anthropomorphic and non-self-anthropomorphic for each original bot response. Our work not only uncovers a new category of bot responses that were previously under-explored but also lays the groundwork for future studies about dynamically adjusting self-anthropomorphism levels in AI systems to align with ethical standards and user expectations., Comment: Findings of EMNLP 2024, 19 pages

Deep learning models in medical imaging often encounter challenges when adapting to new clinical settings unseen during training. Test-time adaptation offers a promising approach to optimize models for these unseen domains, yet its application in anomaly detection (AD) remains largely unexplored. AD aims to efficiently identify deviations from normative distributions; however, full adaptation, including pathological shifts, may inadvertently learn the anomalies it intends to detect. We introduce a novel concept of selective test-time adaptation that utilizes the inherent characteristics of deep pre-trained features to adapt selectively in a zero-shot manner to any test image from an unseen domain. This approach employs a model-agnostic, lightweight multi-layer perceptron for neural implicit representations, enabling the adaptation of outputs from any reconstruction-based AD method without altering the source-trained model. Rigorous validation in brain AD demonstrated that our strategy substantially enhances detection accuracy for multiple conditions and different target distributions. Specifically, our method improves the detection rates by up to 78% for enlarged ventricles and 24% for edemas., Comment: MICCAIw ADSMI 2024 [Best Paper Award]

One-to-one tutoring is one of the most efficient methods of teaching. Following the rise in popularity of Large Language Models (LLMs), there have been efforts to use them to create conversational tutoring systems, which can make the benefits of one-to-one tutoring accessible to everyone. However, current LLMs are primarily trained to be helpful assistants and thus lack crucial pedagogical skills. For example, they often quickly reveal the solution to the student and fail to plan for a richer multi-turn pedagogical interaction. To use LLMs in pedagogical scenarios, they need to be steered towards using effective teaching strategies: a problem we introduce as Pedagogical Steering and believe to be crucial for the efficient use of LLMs as tutors. We address this problem by formalizing a concept of tutoring strategy, and introducing StratL, an algorithm to model a strategy and use prompting to steer the LLM to follow this strategy. As a case study, we create a prototype tutor for high school math following Productive Failure (PF), an advanced and effective learning design. To validate our approach in a real-world setting, we run a field study with 17 high school students in Singapore. We quantitatively show that StratL succeeds in steering the LLM to follow a Productive Failure tutoring strategy. We also thoroughly investigate the existence of spillover effects on desirable properties of the LLM, like its ability to generate human-like answers. Based on these results, we highlight the challenges in Pedagogical Steering and suggest opportunities for further improvements. We further encourage follow-up research by releasing a dataset of Productive Failure problems and the code of our prototype and algorithm., Comment: 18 pages, 9 figures, 6 tables

The Visual Monitoring Camera (VMC) is a small imaging instrument onboard Mars Express with a field of view of ~40x30 degrees. The camera was initially intended to provide visual confirmation of the separation of the Beagle 2 lander and has similar technical specifications to a typical webcam of the 2000s. In 2007, a few years after the end of its original mission, VMC was turned on again to obtain full-disk images of Mars to be used for outreach purposes. As VMC obtained more images, the scientific potential of the camera became evident, and in 2018 the camera was given an upgraded status of a new scientific instrument, with science goals in the field of Martian atmosphere meteorology. The wide Field of View of the camera combined with the orbit of Mars Express enable the acquisition of full-disk images of the planet showing different local times, which for a long time has been rare among orbital missions around Mars. The small data volume of images also allows videos that show the atmospheric dynamics of dust and cloud systems to be obtained. This paper is intended to be the new reference paper for VMC as a scientific instrument, and thus provides an overview of the updated procedures to plan, command and execute science observations of the Martian atmosphere. These observations produce valuable science data that is calibrated and distributed to the community for scientific use.

Sulfur is depleted with respect to its cosmic standard abundance in dense star-forming regions. It has been suggested that this depletion is caused by the freeze-out of sulfur on interstellar dust grains, but the observed abundances and upper limits of sulfur-bearing ices remain too low to account for all of the missing sulfur. Toward the same environments, a strong absorption feature at 6.85 $\mu$m is observed, but its long-standing assignment to the NH4+ cation remains tentative. We investigate the plausibility of NH4SH salt serving as a sulfur reservoir and a carrier of the 6.85 $\mu$m band in interstellar ices by characterizing its IR signatures and apparent band strengths in water-rich laboratory ice mixtures and using this laboratory data to constrain NH4SH abundances in observations of 4 protostars and 2 cold dense clouds. The observed 6.85 $\mu$m feature is fit well with the laboratory NH4SH:H2O ice spectra. NH4+ column densities obtained from the 6.85 $\mu$m band range from 8-23% with respect to H2O toward the sample of protostars and dense clouds. The redshift of the 6.85 $\mu$m feature correlates with higher abundances of NH4+ with respect to H2O in both the laboratory data presented here and observational data of dense clouds and protostars. The apparent band strength of the SH- feature is likely too low for the feature to be detectable in the spectrally busy 3.9 $\mu$m region, but the 5.3 $\mu$m NH4+ $\nu_{4}$ + SH- R combination mode may be an alternative means of detection. Its tentative assignment adds to mounting evidence supporting the presence of NH4+ salts in ices and is the first tentative observation of the SH- anion toward interstellar ices. If the majority ($\gtrsim$80-85%) of the NH4+ cations quantified toward the investigated sources in this work are bound to SH- anions, then NH4SH salts could account for up to 17-18% of their sulfur budgets., Comment: Accepted for publication in A&A. 20 pages, 14 figures, and 7 tables in the main text; 15 pages, 17 figures, and 10 tables in the appendix

Clinnova, a collaborative initiative involving France, Germany, Switzerland, and Luxembourg, is dedicated to unlocking the power of precision medicine through data federation, standardization, and interoperability. This European Greater Region initiative seeks to create an interoperable European standard using artificial intelligence (AI) and data science to enhance healthcare outcomes and efficiency. Key components include multidisciplinary research centers, a federated biobanking strategy, a digital health innovation platform, and a federated AI strategy. It targets inflammatory bowel disease, rheumatoid diseases, and multiple sclerosis (MS), emphasizing data quality to develop AI algorithms for personalized treatment and translational research. The IHU Strasbourg (Institute of Minimal-invasive Surgery) has the lead in this initiative to develop the federated learning (FL) proof of concept (POC) that will serve as a foundation for advancing AI in healthcare. At its core, Clinnova-MS aims to enhance MS patient care by using FL to develop more accurate models that detect disease progression, guide interventions, and validate digital biomarkers across multiple sites. This technical report presents insights and key takeaways from the first cross-border federated POC on MS segmentation of MRI images within the Clinnova framework. While our work marks a significant milestone in advancing MS segmentation through cross-border collaboration, it also underscores the importance of addressing technical, logistical, and ethical considerations to realize the full potential of FL in healthcare settings.

Medical image analysis tasks often focus on regions or structures located in a particular location within the patient's body. Often large parts of the image may not be of interest for the image analysis task. When using deep-learning based approaches, this causes an unnecessary increases the computational burden during inference and raises the chance of errors. In this paper, we introduce CTARR, a novel generic method for CT Anatomical Region Recognition. The method serves as a pre-processing step for any deep learning-based CT image analysis pipeline by automatically identifying the pre-defined anatomical region that is relevant for the follow-up task and removing the rest. It can be used in (i) image segmentation to prevent false positives in anatomically implausible regions and speeding up the inference, (ii) image classification to produce image crops that are consistent in their anatomical context, and (iii) image registration by serving as a fast pre-registration step. Our proposed method is based on atlas registration and provides a fast and robust way to crop any anatomical region encoded as one or multiple bounding box(es) from any unlabeled CT scan of the brain, chest, abdomen and/or pelvis. We demonstrate the utility and robustness of the proposed method in the context of medical image segmentation by evaluating it on six datasets of public segmentation challenges. The foreground voxels in the regions of interest are preserved in the vast majority of cases and tasks (97.45-100%) while taking only fractions of a seconds to compute (0.1-0.21s) on a deep learning workstation and greatly reducing the segmentation runtime (2.0-12.7x). Our code is available at https://github.com/ThomasBudd/ctarr.

This paper is devoted to analyzing the observer convergence rate for a class of linear control systems in a Hilbert space. To characterize the polynomial stability of the observer error system, we apply the spectral theory of linear operators and explicitly construct the resolvent of the corresponding infinitesimal generator. The asymptotic behavior of the resolvent on the imaginary axis is studied to describe the rate of decay of the observation error. The estimated decay rate is illustrated through an example of an oscillating flexible structure with one-dimensional output.

The African School of Fundamental Physics and Applications, also known as the African School of Physics (ASP), was initiated in 2010, as a three-week biennial event, to offer additional training in fundamental and applied physics to African students with a minimum of three-year university education. Since its inception, ASP has grown to be much more than a school. ASP has become a series of activities and events to support academic development of African students, teachers and faculties. We report on the eighth African School of Physics, ASP2024, organized in Morocco, on April 15--19 and July 7--21, 2024. ASP2024 included programs for university students, high school teachers and high school pupils., Comment: 23 pages, 24 figures. arXiv admin note: substantial text overlap with arXiv:2302.13940