Search

Few-shot learning in medical image classification presents a significant challenge due to the limited availability of annotated data and the complex nature of medical imagery. In this work, we propose Adaptive Vision-Language Fine-tuning with Hierarchical Contrastive Alignment (HiCA), a novel framework that leverages the capabilities of Large Vision-Language Models (LVLMs) for medical image analysis. HiCA introduces a two-stage fine-tuning strategy, combining domain-specific pretraining and hierarchical contrastive learning to align visual and textual representations at multiple levels. We evaluate our approach on two benchmark datasets, Chest X-ray and Breast Ultrasound, achieving state-of-the-art performance in both few-shot and zero-shot settings. Further analyses demonstrate the robustness, generalizability, and interpretability of our method, with substantial improvements in performance compared to existing baselines. Our work highlights the potential of hierarchical contrastive strategies in adapting LVLMs to the unique challenges of medical imaging tasks.

Population annealing is a variant of the simulated annealing algorithm that improves the quality of the thermalization process in systems with rough free-energy landscapes by introducing a resampling process. We consider the diluted Sherrington-Kirkpatrick Ising model using population annealing to study its efficiency in finding solutions to combinatorial optimization problems. From this study, we find an easy-hard-easy transition in the model hardness as the problem instances become more diluted, and associate this behaviour to the clusterization and connectivity of the underlying Erd\H{o}s-R\'enyi graphs. We calculate the efficiency of obtaining minimum energy configurations and find that population annealing outperforms simulated annealing for the cases close to this hardness peak while reaching similar efficiencies in the easy limits. Finally, it is known that population annealing can be used to define an adaptive inverse temperature annealing schedule. We compare this adaptive method to a linear schedule and find that the adaptive method achieves improved efficiencies while being robust against final temperature miscalibrations., Comment: 10 pages, 12 figures

In this paper, we introduce ChatCite, a novel method leveraging large language models (LLMs) for generating comparative literature summaries. The ability to summarize research papers with a focus on key comparisons between studies is an essential task in academic research. Existing summarization models, while effective at generating concise summaries, fail to provide deep comparative insights. ChatCite addresses this limitation by incorporating a multi-step reasoning mechanism that extracts critical elements from papers, incrementally builds a comparative summary, and refines the output through a reflective memory process. We evaluate ChatCite on a custom dataset, CompLit-LongContext, consisting of 1000 research papers with annotated comparative summaries. Experimental results show that ChatCite outperforms several baseline methods, including GPT-4, BART, T5, and CoT, across various automatic evaluation metrics such as ROUGE and the newly proposed G-Score. Human evaluation further confirms that ChatCite generates more coherent, insightful, and fluent summaries compared to these baseline models. Our method provides a significant advancement in automatic literature review generation, offering researchers a powerful tool for efficiently comparing and synthesizing scientific research., Comment: 8 pages

There is growing interest in applying AI to radiology report generation, particularly for chest X-rays (CXRs). This paper investigates whether incorporating pixel-level information through segmentation masks can improve fine-grained image interpretation of multimodal large language models (MLLMs) for radiology report generation. We introduce MAIRA-Seg, a segmentation-aware MLLM framework designed to utilize semantic segmentation masks alongside CXRs for generating radiology reports. We train expert segmentation models to obtain mask pseudolabels for radiology-specific structures in CXRs. Subsequently, building on the architectures of MAIRA, a CXR-specialised model for report generation, we integrate a trainable segmentation tokens extractor that leverages these mask pseudolabels, and employ mask-aware prompting to generate draft radiology reports. Our experiments on the publicly available MIMIC-CXR dataset show that MAIRA-Seg outperforms non-segmentation baselines. We also investigate set-of-marks prompting with MAIRA and find that MAIRA-Seg consistently demonstrates comparable or superior performance. The results confirm that using segmentation masks enhances the nuanced reasoning of MLLMs, potentially contributing to better clinical outcomes., Comment: Accepted as Proceedings Paper at ML4H 2024

In this paper we study the isometric rigidity of certain classes of metric spaces with respect to the $p$-Wasserstein space. We prove that spaces that split a separable Hilbert space are not isometrically rigid with respect to $\mathbb{P}_2$. We then prove that infinite rays are isometrically rigid with respect to $\mathbb{P}_p$ for any $p\geq 1$, whereas taking infinite half-cylinders (i.e.\ product spaces of the form $X\times [0,\infty)$) over compact non-branching geodesic spaces preserves isometric rigidity with respect to $\mathbb{P}_p$, for $p>1$. Finally, we prove that spherical suspensions over compact spaces with diameter less than $\pi/2$ are isometrically rigid with respect to $\mathbb{P}_p$, for $p>1$.

We introduce a variational algorithm based on Matrix Product States that is trained by minimizing a generalized free energy defined using Tsallis entropy instead of the standard Gibbs entropy. As a result, our model can generate the probability distributions associated with generalized statistical mechanics. The resulting model can be efficiently trained, since the resulting free energy and its gradient can be calculated exactly through tensor network contractions, as opposed to standard methods which require estimating the Gibbs entropy by sampling. We devise a variational annealing scheme by ramping up the inverse temperature, which allows us to train the model while avoiding getting trapped in local minima. We show the validity of our approach in Ising spin-glass problems by comparing it to exact numerical results and quasi-exact analytical approximations. Our work opens up new possibilities for studying generalized statistical physics and solving combinatorial optimization problems with tensor networks., Comment: 6 pages, 6 figures + Supplementary material (5 pages, 5 figures)

We prove Sobolev-Poincar\'e and Poincar\'e inequalities in variable Lebesgue spaces $L^{p(\cdot)}(\Omega)$, with $\Omega\subset{\mathbb R}^n$ a bounded John domain, with weaker regularity assumptions on the exponent $p(\cdot)$ that have been used previously. In particular, we require $p(\cdot)$ to satisfy a new \emph{boundary $\log$-H\"older condition} that imposes some logarithmic decay on the oscillation of $p(\cdot)$ towards the boundary of the domain. Some control over the interior oscillation of $p(\cdot)$ is also needed, but it is given by a very general condition that allows $p(\cdot)$ to be discontinuous at every point of $\Omega$. Our results follows from a local-to-global argument based on the continuity of certain Hardy type operators. We provide examples that show that our boundary $\log$-H\"older condition is essentially necessary for our main results. The same examples are adapted to show that this condition is not sufficient for other related inequalities. Finally, we give an application to a Neumann problem for a degenerate $p(\cdot)$-Laplacian.

We provide a symbolic classification of generalized Seifert fiber spaces, which were introduced by Mitsuishi and Yamaguchi in the classification of collapsing Alexandrov $3$-spaces. Additionally, we show that the canonical double branched cover of a non-manifold generalized Seifert fiber space is a Seifert manifold and compute its symbolic invariants in terms of those of the original space., Comment: 9 pages

The discovery of novel quantum materials within ternary phase spaces containing antagonistic pair such as Fe with Bi, Pb, In, and Ag, presents significant challenges yet holds great potential. In this work, we investigate the stabilization of these immiscible pairs through the integration of Cerium (Ce), an abundant rare-earth and cost-effective element. By employing a machine learning (ML)-guided framework, particularly crystal graph convolutional neural networks (CGCNN), combined with first-principles calculations, we efficiently explore the composition/structure space and predict 9 stable and 37 metastable Ce-Fe-X (X=Bi, Pb, In and Ag) ternary compounds. Our findings include the identification of multiple new stable and metastable phases, which are evaluated for their structural and energetic properties. These discoveries not only contribute to the advancement of quantum materials but also offer viable alternatives to critical rare earth elements, underscoring the importance of Ce-based intermetallic compounds in technological applications.

We prove that the group of isometries preserving a metric foliation on a closed Alexandrov space $X$, or a singular Riemannian foliation on a manifold $M$ is a closed subgroup of the isometry group of $X$ in the case of a metric foliation, or of the isometry group of $M$ for the case of a singular Riemannian foliation. We obtain a sharp upper bound for the dimension of these subgroups and show that, when equality holds, the foliations that realize this upper bound are induced by fiber bundles whose fibers are round spheres or projective spaces. Moreover, singular Riemannian foliations that realize the upper bound are induced by smooth fiber bundles whose fibers are round spheres or projective spaces., Comment: 20 pages, we thank gold essen.trinken in Karlsruhe for its excellent working conditions

PLATO (PLAnetary Transits and Oscillations of stars) is ESA's M3 mission designed to detect and characterise extrasolar planets and perform asteroseismic monitoring of a large number of stars. PLATO will detect small planets (down to <2 R_(Earth)) around bright stars (<11 mag), including terrestrial planets in the habitable zone of solar-like stars. With the complement of radial velocity observations from the ground, planets will be characterised for their radius, mass, and age with high accuracy (5 %, 10 %, 10 % for an Earth-Sun combination respectively). PLATO will provide us with a large-scale catalogue of well-characterised small planets up to intermediate orbital periods, relevant for a meaningful comparison to planet formation theories and to better understand planet evolution. It will make possible comparative exoplanetology to place our Solar System planets in a broader context. In parallel, PLATO will study (host) stars using asteroseismology, allowing us to determine the stellar properties with high accuracy, substantially enhancing our knowledge of stellar structure and evolution. The payload instrument consists of 26 cameras with 12cm aperture each. For at least four years, the mission will perform high-precision photometric measurements. Here we review the science objectives, present PLATO's target samples and fields, provide an overview of expected core science performance as well as a description of the instrument and the mission profile at the beginning of the serial production of the flight cameras. PLATO is scheduled for a launch date end 2026. This overview therefore provides a summary of the mission to the community in preparation of the upcoming operational phases.

Radiology reporting is a complex task requiring detailed medical image understanding and precise language generation, for which generative multimodal models offer a promising solution. However, to impact clinical practice, models must achieve a high level of both verifiable performance and utility. We augment the utility of automated report generation by incorporating localisation of individual findings on the image - a task we call grounded report generation - and enhance performance by incorporating realistic reporting context as inputs. We design a novel evaluation framework (RadFact) leveraging the logical inference capabilities of large language models (LLMs) to quantify report correctness and completeness at the level of individual sentences, while supporting the new task of grounded reporting. We develop MAIRA-2, a large radiology-specific multimodal model designed to generate chest X-ray reports with and without grounding. MAIRA-2 achieves state of the art on existing report generation benchmarks and establishes the novel task of grounded report generation., Comment: 72 pages, 21 figures. v2 updates the model and adds results on the PadChest-GR dataset

Nasogastric tubes (NGTs) are feeding tubes that are inserted through the nose into the stomach to deliver nutrition or medication. If not placed correctly, they can cause serious harm, even death to patients. Recent AI developments demonstrate the feasibility of robustly detecting NGT placement from Chest X-ray images to reduce risks of sub-optimally or critically placed NGTs being missed or delayed in their detection, but gaps remain in clinical practice integration. In this study, we present a human-centered approach to the problem and describe insights derived following contextual inquiry and in-depth interviews with 15 clinical stakeholders. The interviews helped understand challenges in existing workflows, and how best to align technical capabilities with user needs and expectations. We discovered the trade-offs and complexities that need consideration when choosing suitable workflow stages, target users, and design configurations for different AI proposals. We explored how to balance AI benefits and risks for healthcare staff and patients within broader organizational and medical-legal constraints. We also identified data issues related to edge cases and data biases that affect model training and evaluation; how data documentation practices influence data preparation and labelling; and how to measure relevant AI outcomes reliably in future evaluations. We discuss how our work informs design and development of AI applications that are clinically useful, ethical, and acceptable in real-world healthcare services.

The development and use of large-scale quantum computers relies on integrating quantum error-correcting (QEC) schemes into the quantum computing pipeline. A fundamental part of the QEC protocol is the decoding of the syndrome to identify a recovery operation with a high success rate. In this work, we implement a decoder that finds the recovery operation with the highest success probability by mapping the decoding problem to a spin system and using Population Annealing to estimate the free energy of the different error classes. We study the decoder performance on a 4.8.8 color code lattice under different noise models, including code capacity with bit-flip and depolarizing noise, and phenomenological noise, which considers noisy measurements, with performance reaching near-optimal thresholds. This decoding algorithm can be applied to a wide variety of stabilizer codes, including surface codes and quantum low-density parity-check (qLDPC) codes., Comment: 11 pages, 9 figures

Empirical research on \textit{construals}--social affinity groups that share similar patterns of meaning--has advanced significantly in recent years. This progress is largely driven by the development of \textit{Construal Clustering Methods} (CCMs), which group survey respondents into construal clusters based on similarities in their response patterns. We identify key limitations of existing CCMs, which affect their accuracy when applied to the typical structures of available data, and introduce Bipolar Class Analysis (BCA), a CCM designed to address these shortcomings. BCA measures similarity in response shifts between expressions of support and rejection across survey respondents, addressing conceptual and measurement challenges in existing methods. We formally define BCA and demonstrate its advantages through extensive simulation analyses, where it consistently outperforms existing CCMs in accurately identifying construals. Along the way, we develop a novel data-generation process that approximates more closely how individuals map latent opinions onto observable survey responses, as well as a new metric to evaluate the performance of CCMs. Additionally, we find that applying BCA to previously studied real-world datasets reveals substantively different construal patterns compared to those generated by existing CCMs in prior empirical analyses. Finally, we discuss limitations of BCA and outline directions for future research.

Recent advances in AI combine large language models (LLMs) with vision encoders that bring forward unprecedented technical capabilities to leverage for a wide range of healthcare applications. Focusing on the domain of radiology, vision-language models (VLMs) achieve good performance results for tasks such as generating radiology findings based on a patient's medical image, or answering visual questions (e.g., 'Where are the nodules in this chest X-ray?'). However, the clinical utility of potential applications of these capabilities is currently underexplored. We engaged in an iterative, multidisciplinary design process to envision clinically relevant VLM interactions, and co-designed four VLM use concepts: Draft Report Generation, Augmented Report Review, Visual Search and Querying, and Patient Imaging History Highlights. We studied these concepts with 13 radiologists and clinicians who assessed the VLM concepts as valuable, yet articulated many design considerations. Reflecting on our findings, we discuss implications for integrating VLM capabilities in radiology, and for healthcare AI more generally., Comment: to appear at CHI 2024

Language-supervised pre-training has proven to be a valuable method for extracting semantically meaningful features from images, serving as a foundational element in multimodal systems within the computer vision and medical imaging domains. However, the computed features are limited by the information contained in the text, which is particularly problematic in medical imaging, where the findings described by radiologists focus on specific observations. This challenge is compounded by the scarcity of paired imaging-text data due to concerns over leakage of personal health information. In this work, we fundamentally challenge the prevailing reliance on language supervision for learning general-purpose biomedical imaging encoders. We introduce RAD-DINO, a biomedical image encoder pre-trained solely on unimodal biomedical imaging data that obtains similar or greater performance than state-of-the-art biomedical language-supervised models on a diverse range of benchmarks. Specifically, the quality of learned representations is evaluated on standard imaging tasks (classification and semantic segmentation), and a vision-language alignment task (text report generation from images). To further demonstrate the drawback of language supervision, we show that features from RAD-DINO correlate with other medical records (e.g., sex or age) better than language-supervised models, which are generally not mentioned in radiology reports. Finally, we conduct a series of ablations determining the factors in RAD-DINO's performance; notably, we observe that RAD-DINO's downstream performance scales well with the quantity and diversity of training data, demonstrating that image-only supervision is a scalable approach for training a foundational biomedical image encoder. Model weights of RAD-DINO trained on publicly available datasets are available at https://huggingface.co/microsoft/rad-dino.

Superconductivity (SC) is absent in Cr-substituted BaFe$_{2}$As$_{2}$ (CrBFA), a well-established but poorly understood topic. Additionally, the suppression of the spin density wave transition temperature ($T_{\text{SDW}}$) in CrBFA and Mn-substituted BaFe$_{2}$As$_{2}$ (MnBFA) coincides as a function of Cr/Mn content, despite the distinct electronic effects of these substitutions. In this work, we employ angle-resolved photoemission spectroscopy (ARPES) and combined density functional theory plus dynamical mean field theory calculations (DFT+DMFT) to address the evolution of the Fermi surface (FS) and electronic correlations in CrBFA. Our findings reveal that incorporating Cr leads to an effective hole doping of the states near the FS, which is well described within the virtual crystal approximation (VCA). We analyzed the electronic band spectra with main $d_{yz}$-orbital character and found a fractional scaling of the imaginary part of self-energy as a function of the binding energy, a signature property of Hund's correlations. We conclude that CrBFA is a correlated electron system and the changes in the FS as a function of Cr are unrelated to the suppression of $T_{\text{SDW}}$. We suggest that the absence of SC is primarily due to the competition between Cr local moments and the Fe-derived itinerant spin fluctuations., Comment: 18 pages, 5 figures Submission to SciPost Physics

We present a radiology-specific multimodal model for the task for generating radiological reports from chest X-rays (CXRs). Our work builds on the idea that large language model(s) can be equipped with multimodal capabilities through alignment with pre-trained vision encoders. On natural images, this has been shown to allow multimodal models to gain image understanding and description capabilities. Our proposed model (MAIRA-1) leverages a CXR-specific image encoder in conjunction with a fine-tuned large language model based on Vicuna-7B, and text-based data augmentation, to produce reports with state-of-the-art quality. In particular, MAIRA-1 significantly improves on the radiologist-aligned RadCliQ metric and across all lexical metrics considered. Manual review of model outputs demonstrates promising fluency and accuracy of generated reports while uncovering failure modes not captured by existing evaluation practices. More information and resources can be found on the project website: https://aka.ms/maira., Comment: 18 pages, 9 tables, 5 figures. v2 adds test IDs and image encoder citation. v3 fixes error in NPV/specificity

The recent success of general-domain large language models (LLMs) has significantly changed the natural language processing paradigm towards a unified foundation model across domains and applications. In this paper, we focus on assessing the performance of GPT-4, the most capable LLM so far, on the text-based applications for radiology reports, comparing against state-of-the-art (SOTA) radiology-specific models. Exploring various prompting strategies, we evaluated GPT-4 on a diverse range of common radiology tasks and we found GPT-4 either outperforms or is on par with current SOTA radiology models. With zero-shot prompting, GPT-4 already obtains substantial gains ($\approx$ 10% absolute improvement) over radiology models in temporal sentence similarity classification (accuracy) and natural language inference ($F_1$). For tasks that require learning dataset-specific style or schema (e.g. findings summarisation), GPT-4 improves with example-based prompting and matches supervised SOTA. Our extensive error analysis with a board-certified radiologist shows GPT-4 has a sufficient level of radiology knowledge with only occasional errors in complex context that require nuanced domain knowledge. For findings summarisation, GPT-4 outputs are found to be overall comparable with existing manually-written impressions., Comment: EMNLP 2023 main

In recent years, the field of autonomous driving has witnessed remarkable advancements, driven by the integration of a multitude of sensors, including cameras and LiDAR systems, in different prototypes. However, with the proliferation of sensor data comes the pressing need for more sophisticated information processing techniques. This research paper introduces a novel modification to an object detection network that uses camera and lidar information, incorporating an additional branch designed for the task of re-identifying objects across adjacent cameras within the same vehicle while elevating the quality of the baseline 3D object detection outcomes. The proposed methodology employs a two-step detection pipeline: initially, an object detection network is employed, followed by a 3D box estimator that operates on the filtered point cloud generated from the network's detections. Extensive experimental evaluations encompassing both 2D and 3D domains validate the effectiveness of the proposed approach and the results underscore the superiority of this method over traditional Non-Maximum Suppression (NMS) techniques, with an improvement of more than 5\% in the car category in the overlapping areas.

We extend basic results in $3$-manifold topology to general three-dimensional Alexandrov spaces (or Alexandrov $3$-spaces for short), providing a unified framework for manifold and non-manifold spaces. We generalize the connected sum to non-manifold $3$-spaces and prove a prime decomposition theorem, exhibit an infinite family of closed, prime non-manifold $3$-spaces which are not irreducible, and establish a conjecture of Mitsuishi and Yamaguchi on the structure of closed, simply-connected Alexandrov $3$-spaces with non-negative curvature. Additionally, we define a notion of generalized Dehn surgery for Alexandrov $3$-spaces and show that any closed Alexandrov $3$-space may be obtained by performing generalized Dehn surgery on a link in $S^3$ or the non-trivial $S^2$-bundle over $S^1$. As an application of this result, we show that every closed Alexandrov $3$-space is homeomorphic to the boundary of a $4$-dimensional Alexandrov space., Comment: 24 pages, 6 figures

Chemical substitution is commonly used to explore new ground states in materials, yet the role of disorder is often overlooked. In Mn-substituted BaFe$_{2}$As$_{2}$ (MnBFA), superconductivity (SC) is absent, despite being observed for nominal hole-doped phases. Instead, a glassy magnetic phase emerges, associated with the $S=5/2$ Mn local spins. In this work, we present a comprehensive investigation of the electronic structure of MnBFA using angle-resolved photoemission spectroscopy (ARPES). We find that Mn causes a small and orbital-specific reduction of the electron pockets, only partially disrupting nesting conditions. Based upon the analysis of the spectral properties, we observe, for all bands, an increase in the electronic scattering rate as a function of Mn content. This is interpreted as increasing band incoherence, which we propose as the primary contributor to the suppression of the magnetic order in MnBFA. This finding connects the MnBFA electronic band structure properties to the glassy magnetic behavior observed in these materials and suggests that SC is absent because of the collective magnetic impurity behavior that scatters the Fe-derived excitations. Additionally, our analysis shows that the binding energy ($E_{B}$) dependence of the imaginary part of the self-energy [$\text{Im}\Sigma(E_{B})$] is best described by a fractional scaling ($\text{Im}\Sigma(E_{B})\propto\sqrt{-E_{B}}$). These results indicate that Mn tunes MnBFA into an electronic disordered phase between the correlated Hund's metal in BaFe$_{2}$As$_{2}$ and the Hund's insulator in BaMn$_{2}$As$_{2}$., Comment: main 8 pages, 5 figures + supp 5 pages, 4 figures

We express the total space of a principal circle bundle over a connected sum of two manifolds in terms of the total spaces of circle bundles over each summand, provided certain conditions hold. We then apply this result to provide sufficient conditions for the existence of free circle and torus actions on connected sums of products of spheres and obtain a topological classification of closed, simply-connected manifolds with a free cohomogeneity-four torus action. As a corollary, we obtain infinitely-many manifolds with Riemannian metrics of positive Ricci curvature and isometric torus actions., Comment: Minor changes, final version, to appear in Forum of Mathematics, Sigma

We introduce a novel Region-based contrastive pretraining for Medical Image Retrieval (RegionMIR) that demonstrates the feasibility of medical image retrieval with similar anatomical regions. RegionMIR addresses two major challenges for medical image retrieval i) standardization of clinically relevant searching criteria (e.g., anatomical, pathology-based), and ii) localization of anatomical area of interests that are semantically meaningful. In this work, we propose an ROI image retrieval image network that retrieves images with similar anatomy by extracting anatomical features (via bounding boxes) and evaluate similarity between pairwise anatomy-categorized features between the query and the database of images using contrastive learning. ROI queries are encoded using a contrastive-pretrained encoder that was fine-tuned for anatomy classification, which generates an anatomical-specific latent space for region-correlated image retrieval. During retrieval, we compare the anatomically encoded query to find similar features within a feature database generated from training samples, and retrieve images with similar regions from training samples. We evaluate our approach on both anatomy classification and image retrieval tasks using the Chest ImaGenome Dataset. Our proposed strategy yields an improvement over state-of-the-art pretraining and co-training strategies, from 92.24 to 94.12 (2.03%) classification accuracy in anatomies. We qualitatively evaluate the image retrieval performance demonstrating generalizability across multiple anatomies with different morphology.

Label scarcity is a bottleneck for improving task performance in specialised domains. We propose a novel compositional transfer learning framework (DoT5 - domain compositional zero-shot T5) for zero-shot domain transfer. Without access to in-domain labels, DoT5 jointly learns domain knowledge (from MLM of unlabelled in-domain free text) and task knowledge (from task training on more readily available general-domain data) in a multi-task manner. To improve the transferability of task training, we design a strategy named NLGU: we simultaneously train NLG for in-domain label-to-data generation which enables data augmentation for self-finetuning and NLU for label prediction. We evaluate DoT5 on the biomedical domain and the resource-lean subdomain of radiology, focusing on NLI, text summarisation and embedding learning. DoT5 demonstrates the effectiveness of compositional transfer learning through multi-task learning. In particular, DoT5 outperforms the current SOTA in zero-shot transfer by over 7 absolute points in accuracy on RadNLI. We validate DoT5 with ablations and a case study demonstrating its ability to solve challenging NLI examples requiring in-domain expertise., Comment: Accepted at TACL, pre-MIT Press publication version. 16 pages, 4 figures

Software Defined Networking (SDN) is a widely deployed technology enabling the agile and flexible management of networks and services. This paradigm represents an appropriate candidate to address the dynamic and secure management of large and complex Authentication, Authorization and Accounting (AAA) infrastructures. In those infrastructures, there are several nodes which must exchange information securely to interconnect different realms. This article describes a novel SDN-based framework with a data model-driven approach following the standard YANG, named SDN-AAA, which can be used to dynamically manage routing and security configuration in AAA scenarios., Comment: 7 pages, 5 figures

Combinatorial optimization problems are ubiquitous in industry. In addition to finding a solution with minimum cost, problems of high relevance involve a number of constraints that the solution must satisfy. Variational quantum algorithms have emerged as promising candidates for solving these problems in the noisy intermediate-scale quantum stage. However, the constraints are often complex enough to make their efficient mapping to quantum hardware difficult or even infeasible. An alternative standard approach is to transform the optimization problem to include these constraints as penalty terms, but this method involves additional hyperparameters and does not ensure that the constraints are satisfied due to the existence of local minima. In this paper, we introduce a new method for solving combinatorial optimization problems with challenging constraints using variational quantum algorithms. We propose the Multi-Objective Variational Constrained Optimizer (MOVCO) to classically update the variational parameters by a multiobjective optimization performed by a genetic algorithm. This optimization allows the algorithm to progressively sample only states within the in-constraints space, while optimizing the energy of these states. We test our proposal on a real-world problem with great relevance in finance: the Cash Management problem. We introduce a novel mathematical formulation for this problem, and compare the performance of MOVCO versus a penalty based optimization. Our empirical results show a significant improvement in terms of the cost of the achieved solutions, but especially in the avoidance of local minima that do not satisfy any of the mandatory constraints., Comment: 19 pages, 9 figures + appendix (4 pages, 2 figures)

Self-supervised learning in vision-language processing exploits semantic alignment between imaging and text modalities. Prior work in biomedical VLP has mostly relied on the alignment of single image and report pairs even though clinical notes commonly refer to prior images. This does not only introduce poor alignment between the modalities but also a missed opportunity to exploit rich self-supervision through existing temporal content in the data. In this work, we explicitly account for prior images and reports when available during both training and fine-tuning. Our approach, named BioViL-T, uses a CNN-Transformer hybrid multi-image encoder trained jointly with a text model. It is designed to be versatile to arising challenges such as pose variations and missing input images across time. The resulting model excels on downstream tasks both in single- and multi-image setups, achieving state-of-the-art performance on (I) progression classification, (II) phrase grounding, and (III) report generation, whilst offering consistent improvements on disease classification and sentence-similarity tasks. We release a novel multi-modal temporal benchmark dataset, MS-CXR-T, to quantify the quality of vision-language representations in terms of temporal semantics. Our experimental results show the advantages of incorporating prior images and reports to make most use of the data., Comment: To appear in CVPR 2023

We study certain inequalities and a related result on weighted Sobolev spaces on bounded John domains in $\mathbb{R}^n$. Namely, we prove the existence of a right inverse for the divergence operator, along with the corresponding a priori estimate, the improved and the fractional Poincar\'e inequalities, the Korn inequality and the local Fefferman-Stein inequality. All these results are obtained on weighted Sobolev spaces, where the weight is a power of the distance to the boundary. In all cases the exponent of the weight $d(\cdot,\partial\Omega)^{\beta p}$ is only required to satisfy the restriction: $\beta p>-(n-\dim_A(\partial\Omega))$, where $p$ is the exponent of the Sobolev space and $\dim_A(\partial\Omega)$ is the Assouad dimension of the boundary of the domain. According to our best knowledge, this condition is less restrictive than the ones in the literature., Comment: 21 pages, 2 figures