Search

Many active galaxies harbor powerful relativistic jets, however, the detailed mechanisms of their formation and acceleration remain poorly understood. To investigate the area of jet acceleration and collimation with the highest available angular resolution, we study the innermost region of the bipolar jet in the nearby low-ionization nuclear emission-line region (LINER) galaxy NGC 1052. We combined observations of NGC 1052 taken with VLBA, GMVA, and EHT over one week in the spring of 2017. For the first time, NGC 1052 was detected with the EHT, providing a size of the central region in-between both jet bases of 250 RS (Schwarzschild radii) perpendicular to the jet axes. This size estimate supports previous studies of the jets expansion profile which suggest two breaks of the profile at around 300 RS and 10000 RS distances to the core. Furthermore, we estimated the magnetic field to be 1.25 Gauss at a distance of 22 {\mu}as from the central engine by fitting a synchrotron-self absorption spectrum to the innermost emission feature, which shows a spectral turn-over at about 130 GHz. Assuming a purely poloidal magnetic field, this implies an upper limit on the magnetic field strength at the event horizon of 26000 Gauss, which is consistent with previous measurements. The complex, low-brightness, double-sided jet structure in NGC 1052 makes it a challenge to detect the source at millimeter (mm) wavelengths. However, our first EHT observations have demonstrated that detection is possible up to at least 230 GHz. This study offers a glimpse through the dense surrounding torus and into the innermost central region, where the jets are formed. This has enabled us to finally resolve this region and provide improved constraints on its expansion and magnetic field strength., Comment: 22 pages, 10 figures, published in A&A

Giant impacts dominate the late stages of accretion of rocky planets. They contribute to the heating, melting, and sometimes vaporizing of the bodies involved in the impacts. Due to fractionation during melting and vaporization, planet-building impacts can significantly change the composition and geochemical signatures of rocky objects. Using first-principles molecular dynamics simulations, we analyze the shock behavior of complex realistic silicate systems, representative of both rocky bodies. We introduce a novel criterion for vapor formation that uses entropy calculations to determine the minimum impact velocity required to pass the threshold for vapor production. We derive impact velocity criteria for vapor formation (7.1 km per s for chondritic bodies) and show that this threshold is reached in 61 and 89 percent of impacts in dynamical simulations of the late stages of accretion with classical and annulus starting configuration (respectively) for analogs of Earth. These outcomes should be nuanced by factors such as the impact angle and the mass of the impacting bodies, which further influence the vaporization dynamics and the resultant material distribution. Our findings indicate that vaporization was common during accretion and likely played a crucial role in shaping the early environments and material properties of terrestrial planets., Comment: 17 pages, 4 figures

Solving hard optimization problems is one of the most promising application domains for quantum computers due to the ubiquity of such problems in industry and the availability of broadly applicable quantum speedups. However, the ability of near-term quantum computers to tackle industrial-scale optimization problems is limited by their size and the overheads of quantum error correction. Quantum Random Access Optimization (QRAO) has been proposed to reduce the space requirements of quantum optimization. However, to date QRAO has only been implemented using variational algorithms, which suffer from the need to train instance-specific variational parameters, making them difficult to scale. We propose and benchmark a non-variational approach to QRAO based on the Quantum Alternating Operator Ansatz (QAOA) for the MaxCut problem. We show that instance-independent ``fixed'' parameters achieve good performance, removing the need for variational parameter optimization. Additionally, we evaluate different design choices, such as various mixers and initial states, as well as QAOA operator implementations when customizing for QRAO, and identify a strategy that performs well in practice. Our results pave the way for the practical execution of QRAO on early fault-tolerant quantum computers., Comment: 8 pages, 8 figures

Randomized experiments are the preferred approach for evaluating the effects of interventions, but they are costly and often yield estimates with substantial uncertainty. On the other hand, in silico experiments leveraging foundation models offer a cost-effective alternative that can potentially attain higher statistical precision. However, the benefits of in silico experiments come with a significant risk: statistical inferences are not valid if the models fail to accurately predict experimental responses to interventions. In this paper, we propose a novel approach that integrates the predictions from multiple foundation models with experimental data while preserving valid statistical inference. Our estimator is consistent and asymptotically normal, with asymptotic variance no larger than the standard estimator based on experimental data alone. Importantly, these statistical properties hold even when model predictions are arbitrarily biased. Empirical results across several randomized experiments show that our estimator offers substantial precision gains, equivalent to a reduction of up to 20% in the sample size needed to match the same precision as the standard estimator based on experimental data alone.

The garnet compound Yb$_3$Ga$_5$O$_{12}$ is a fascinating material that is considered highly suitable for low-temperature refrigeration, via the magnetocaloric effect, in addition to enabling the exploration of quantum states with long-range dipolar interactions. It has previously been theorized that the magnetocaloric effect can be enhanced, in Yb$_3$Ga$_5$O$_{12}$ , via magnetic soft mode excitations which in the hyperkagome structure would be derived from an emergent magnetic structure formed from nanosized 10-spin loops. We study the magnetic field dependence of bands of magnetic soft mode excitations in the effective spin $S = 1/2$ hyperkagome compound Yb$_3$Ga$_5$O$_{12}$ using single crystal inelastic neutron scattering. We probe the magnetically short ranged ordered state, in which we determine magnetic nanoscale structures coexisting with a fluctuating state, and the magnetically saturated state. We determine that Yb$_3$Ga$_5$O$_{12}$ can be described as a quantum dipolar magnet with perturbative weak near-neighbor and inter-hyperkagome exchange interaction. The magnetic excitations, under the application of a magnetic field, reveal highly robust soft modes with distinctive signatures of the quantum nature of the Yb3+ spins. Our results enhance our understanding of soft modes in topological frustrated magnets that drive both the unusual physics of quantum dipolar systems and future refrigerant material design.

We use JWST Near-Infrared Spectrograph (NIRSpec) observations from the the Cosmic Evolution Early Release survey (CEERS), GLASS-JWST ERS (GLASS), and JWST Advanced Deep Extragalactic Survey (JADES) to measure rest-frame optical emission-line ratios of 90 galaxies at z>4. The stacked spectra of galaxies with and without a broad-line feature reveal a difference in the [OIII]$\lambda$ 4363 and H$\gamma$ ratios. This motivated our investigation of the [OIII]/H$\gamma$ vs [NeIII]/[OII] diagram. We define two AGN/SF classification lines based on 1869 SDSS galaxies at z$\sim$0. After applying a redshift correction to the AGN/SF lines we find 76.8% of BLAGN continue to land in the AGN region of the diagnostic largely due to the [NeIII]/[OII] ratio. However, 40.2% of non-BLAGN land in the AGN region as well, this could be due to star forming galaxies having harder ionization of there are narrow line AGN which are not accounted for. This indicates the potential of the [NeIII]/[OII] ratio to continue classifying galaxies to z$\sim$6. We further inspect galaxies without broad emission lines in each region of [OIII]/H\gamma vs [NeIII]/[OII] diagram and found that they have slightly stronger CIII]$\lambda$1908 fluxes and equivalent width when landing in the BLAGN region. However, the cause of this higher ionization is unclear. Additionally, we find that BLAGN are characterized by a higher ionization (at constant electron temperature) compared to non-broad line galaxies., Comment: 14 pages, 8 figures

Segmentation of brain tumors is a critical step in treatment planning, yet manual segmentation is both time-consuming and subjective, relying heavily on the expertise of radiologists. In Sub-Saharan Africa, this challenge is magnified by overburdened medical systems and limited access to advanced imaging modalities and expert radiologists. Automating brain tumor segmentation using deep learning offers a promising solution. Convolutional Neural Networks (CNNs), especially the U-Net architecture, have shown significant potential. However, a major challenge remains: achieving generalizability across different datasets. This study addresses this gap by developing a deep learning ensemble that integrates UNet3D, V-Net, and MSA-VNet models for the semantic segmentation of gliomas. By initially training on the BraTS-GLI dataset and fine-tuning with the BraTS-SSA dataset, we enhance model performance. Our ensemble approach significantly outperforms individual models, achieving DICE scores of 0.8358 for Tumor Core, 0.8521 for Whole Tumor, and 0.8167 for Enhancing Tumor. These results underscore the potential of ensemble methods in improving the accuracy and reliability of automated brain tumor segmentation, particularly in resource-limited settings.

Surveying prior literature to establish a foundation for new knowledge is essential for scholarly progress. However, survey articles are resource-intensive and challenging to create, and can quickly become outdated as new research is published, risking information staleness and inaccuracy. Keeping survey articles current with the latest evidence is therefore desirable, though there is a limited understanding of why, when, and how these surveys should be updated. Toward this end, through a series of in-depth retrospective interviews with 11 researchers, we present an empirical examination of the work practices in authoring and updating survey articles in computing research. We find that while computing researchers acknowledge the value in maintaining an updated survey, continuous updating remains unmanageable and misaligned with academic incentives. Our findings suggest key leverage points within current workflows that present opportunities for enabling technologies to facilitate more efficient and effective updates., Comment: 10 pages; To appear at CHI 2025

Air pollution (AP) poses a great threat to human health, and people are paying more attention than ever to its prediction. Accurate prediction of AP helps people to plan for their outdoor activities and aids protecting human health. In this paper, long-short term memory (LSTM) recurrent neural networks (RNNs) have been used to predict the future concentration of air pollutants (APS) in Macau. Additionally, meteorological data and data on the concentration of APS have been utilized. Moreover, in Macau, some air quality monitoring stations (AQMSs) have less observed data in quantity, and, at the same time, some AQMSs recorded less observed data of certain types of APS. Therefore, the transfer learning and pre-trained neural networks have been employed to assist AQMSs with less observed data to build a neural network with high prediction accuracy. The experimental sample covers a period longer than 12-year and includes daily measurements from several APS as well as other more classical meteorological values. Records from five stations, four out of them are AQMSs and the remaining one is an automatic weather station, have been prepared from the aforesaid period and eventually underwent to computational intelligence techniques to build and extract a prediction knowledge-based system. As shown by experimentation, LSTM RNNs initialized with transfer learning methods have higher prediction accuracy; it incurred shorter training time than randomly initialized recurrent neural networks., Comment: Forecasting, environment monitoring, transfer learning, recurrent neural network, airborne particle matter

This paper examines the systemic risks posed by incremental advancements in artificial intelligence, developing the concept of `gradual disempowerment', in contrast to the abrupt takeover scenarios commonly discussed in AI safety. We analyze how even incremental improvements in AI capabilities can undermine human influence over large-scale systems that society depends on, including the economy, culture, and nation-states. As AI increasingly replaces human labor and cognition in these domains, it can weaken both explicit human control mechanisms (like voting and consumer choice) and the implicit alignments with human interests that often arise from societal systems' reliance on human participation to function. Furthermore, to the extent that these systems incentivise outcomes that do not line up with human preferences, AIs may optimize for those outcomes more aggressively. These effects may be mutually reinforcing across different domains: economic power shapes cultural narratives and political decisions, while cultural shifts alter economic and political behavior. We argue that this dynamic could lead to an effectively irreversible loss of human influence over crucial societal systems, precipitating an existential catastrophe through the permanent disempowerment of humanity. This suggests the need for both technical research and governance approaches that specifically address the risk of incremental erosion of human influence across interconnected societal systems., Comment: 19 pages, 2 figures

Two-dimensional (2D) transition metal carbides and nitrides, known as MXenes, possess unique physical and chemical properties, enabling diverse applications in fields ranging from energy storage to communication, catalysis, sensing, healthcare, and beyond. The transition metal and nonmetallic atoms in MXenes can exhibit distinct coordination environments, potentially leading to a wide variety of 2D phases. Despite extensive research and significant advancements, a fundamental understanding of MXenes' phase diversity and its relationship with their hierarchical precursors, including intermediate MAX phases and parent bulk phases, remains limited. Using high-throughput modeling based on first-principles density functional theory, we unveil a wide range of MXenes and comprehensively evaluate their relative stabilities across a large chemical space. The key lies in considering both octahedral and trigonal prismatic coordination environments characteristic of various bulk phases. Through this comprehensive structural library of MXenes, we uncover a close alignment between the phase stability of MXenes and that of their hierarchical 3D counterparts. Building on this, we demonstrate a new design strategy where the atomic coordination environments in parent bulk phases can serve as reliable predictors for the design of MXenes, reducing reliance on intermediate MAX phases. Our study significantly expands the landscape of MXenes, at least doubling the number of possible structures., Comment: 17 pages, 5 illustrations (3 figures and 2 tables)

This article deals with a generalization of the superadiabatic projectors method. In a general framework, the well-known superadiabatic projectors are constructed and accurately described in the case of rank one, when a remarkable factorization occurs. We apply these ideas to spectral theory and we explain how our abstract results allow to recover or improve recent results about the semiclassical magnetic Laplacian.

We report measurements of anisotropic triple-$q$ charge density wave (CDW) fluctuations in the transition metal dichalcogenide 1$T$-TiSe$_2$ over a large volume of reciprocal space with X-ray diffuse scattering. Above the transition temperature, $T_{\text{CDW}}$, the out-of-plane diffuse scattering is characterized by rod-like structures which indicate that the CDW fluctuations in neighboring layers are largely decoupled. In addition, the in-plane diffuse scattering is marked by ellipses which reveal that the in-plane fluctuations are anisotropic. Our analysis of the diffuse scattering line shapes and orientations suggests that the three charge density wave components contain independent phase fluctuations. At $T_{\text{CDW}}$, long range coherence is established in both the in-plane and out-of-plane directions, consistent with the large observed value of the CDW gap compared to $T_{\text{CDW}}$, and the predicted presence of a hierarchy of energy scales.

Foundation models (FMs) have shown transformative potential in radiology by performing diverse, complex tasks across imaging modalities. Here, we developed CT-FM, a large-scale 3D image-based pre-trained model designed explicitly for various radiological tasks. CT-FM was pre-trained using 148,000 computed tomography (CT) scans from the Imaging Data Commons through label-agnostic contrastive learning. We evaluated CT-FM across four categories of tasks, namely, whole-body and tumor segmentation, head CT triage, medical image retrieval, and semantic understanding, showing superior performance against state-of-the-art models. Beyond quantitative success, CT-FM demonstrated the ability to cluster regions anatomically and identify similar anatomical and structural concepts across scans. Furthermore, it remained robust across test-retest settings and indicated reasonable salient regions attached to its embeddings. This study demonstrates the value of large-scale medical imaging foundation models and by open-sourcing the model weights, code, and data, aims to support more adaptable, reliable, and interpretable AI solutions in radiology., Comment: 6 figures, followed by 9 Extended Data Figures and a Supplementary Information document

The $r$-neighbourhood complexity of a graph $G$ is the function counting, for a given integer $k$, the largest possible number, over all vertex-subsets $A$ of size $k$, of subsets of $A$ realized as the intersection between the $r$-neighbourhood of some vertex and $A$. A refinement of this notion is the $r$-profile complexity, that counts the maximum number of distinct distance-vectors from any vertex to the vertices of $A$, ignoring distances larger than $r$. Typically, in structured graph classes such as graphs of bounded VC-dimension or chordal graphs, these functions are bounded, leading to insights into their structural properties and efficient algorithms. We improve existing bounds on the $r$-profile complexity (and thus on the $r$-neighbourhood complexity) for graphs in several structured graph classes. We show that the $r$-profile complexity of graphs excluding $K_h$ as a minor is in $O_h(r^{3h-3}k)$. For graphs of treewidth at most $t$ we give a bound in $O_t(r^{t+1}k)$, which is tight up to a function of $t$ as a factor. These bounds improve results and answer a question of Joret and Rambaud [Combinatorica, 2024]. For outerplanar graphs, we can improve our treewidth bound by a factor of $r$ and conjecture that a similar improvement holds for graphs with bounded simple treewidth. For graphs of treelength at most $\ell$, we give the upper bound in $O(k(r^2(\ell+1)^k))$. Our bounds also imply relations between the order, diameter and metric dimension of graphs in these classes, improving results from [Beaudou et al., SIDMA 2017].

With the development of deep learning, Dynamic Portfolio Optimization (DPO) problem has received a lot of attention in recent years, not only in the field of finance but also in the field of deep learning. Some advanced research in recent years has proposed the application of Deep Reinforcement Learning (DRL) to the DPO problem, which demonstrated to be more advantageous than supervised learning in solving the DPO problem. However, there are still certain unsolved issues: 1) DRL algorithms usually have the problems of slow learning speed and high sample complexity, which is especially problematic when dealing with complex financial data. 2) researchers use DRL simply for the purpose of obtaining high returns, but pay little attention to the problem of risk control and trading strategy, which will affect the stability of model returns. In order to address these issues, in this study we revamped the intrinsic structure of the model based on the Deep Deterministic Policy Gradient (DDPG) and proposed the Augmented DDPG model. Besides, we also proposed an innovative risk control strategy based on Quantum Price Levels (QPLs) derived from Quantum Finance Theory (QFT). Our experimental results revealed that our model has better profitability as well as risk control ability with less sample complexity in the DPO problem compared to the baseline models., Comment: 8 pages

Using optical and near-infrared images of the Cassiopeia A (Cas A) supernova remnant covering the time period 1951 to 2022, together with optical spectra of selected filaments, we present an investigation of Cas A's reverse shock velocity and the effects it has on the remnant's metal-rich ejecta. We find the sequence of optical ejecta brightening and the appearance of new optical ejecta indicating the advancement of the remnant's reverse shock in the remnant's main shell has velocities typically between 1000 and 2000 km/s, which is ~1000 km/s less than recent measurements made in X-rays. We further find the reverse shock appears to move much more slowly and is nearly even stationary in the sky frame along the remnant's western limb. However, we do not find the reverse shock to move inward at velocities as large as ~2000 km/s as has been reported. Optical ejecta in Cas A's main emission shell have proper motions indicating outward tangential motions ~3500 - 6000 km/s, with the smaller values preferentially along the remnant's southern regions which we speculate may be partially the cause of the remnant's faint and more slowly evolving southern sections. Following interaction with the reverse shock, ejecta knots exhibit extended mass ablated trails 0.2" - 0.5" in length leading to extended emission indicating reverse shock induced decelerated velocities as large as 1000 km/s. Such ablated material is most prominently seen in higher ionization line emissions, whereas denser parts of ejecta knots show surprisingly little deceleration., Comment: 36 pages, 20 figures

One of the main unknowns in galaxy evolution is how gas flows into and out of galaxies in the circumgalactic medium (CGM). Studies observing the CGM in absorption using multiple or extended background objects suggest a high degree of variation on relatively small ($\lesssim 1$ kpc) spatial scales. Similarly, high-resolution simulations generally exhibit small-scale substructure in the gas around galaxies. We examine the small-scale structure of the $z = 1$ CGM using simulations from the FOGGIE (Figuring Out Gas & Galaxies in Enzo) project. We select gaseous substructures ("clumps") by their local overdensity and investigate their physical properties, including temperature, metallicity, and kinematics with respect to the galaxy and the nearby surroundings. FOGGIE resolves clumps down to sphericalized radii $R \sim 0.25$ kpc at $z = 1$. The distribution of clumps peaks at $\sim 10^5$ $\rm M_{\odot}$ and $10^{4}$ K, consistent with relatively condensed, cool gas with a slight preference for inflow-like velocities. Many clumps show internal temperature and density variations, and thus internally varying ionization levels for key diagnostic ions such as HI, MgII, and OVI. The average metallicity in clumps is about a factor 1.5--2$\times$ lower in metallicity than nearby gas, suggesting that the metals are not well-mixed between structured and diffuse CGM, which may have implications for observational metallicity estimations of dense CGM clouds. We estimate the survivability of CGM clumps and find that structures larger than 0.5 kpc are generally long-lived. Finally, we qualitatively compare the simulated cloud properties to Milky Way high-velocity clouds., Comment: submitted to ApJ

We investigate the entanglement of a ladder of spins formed by two sublattices, a ''cluster'' chain and the ''environment'', consisting of independent spins, both coupled by an exchange interaction and evolving under a unitary discrete time dynamics. The automaton is defined by the composition of the two body spin swap gate (between sublattices) and the three body cluster interaction. We observe that, depending on the set of coupling constants, the cluster subsystem evolves towards states corresponding to different entanglement phases. In the weak coupling regime the subsystem remains near the topological cluster state. Increasing the coupling strength leads to random states which transform from almost pure to fully mixed, according to the effective number of the environment active degrees of freedom.

In this arxiv-post I present my solutions (published or not) to Problems that appeared in Amer. Math. Monthly, Math. Magazine, Elemente der Mathematik and CRUX, that were mostly done in collaboration with Rudolf Rupp. Some of them (including a few own proposals which were published) were also done in cooperation with Rainer Br\"uck, Bikash Chakraborty, Pamela Gorkin, Gerd Herzog, J\'er\^ome No\"el, Peter Pflug and Amol Sasane., Comment: This submission will be updated on a regular basis

Network coding enhances performance in network communications and distributed storage by increasing throughput and robustness while reducing latency. Batched Sparse (BATS) codes are a class of capacity-achieving network codes, but their practical applications are hindered by their structure, computational intensity, and power demands of finite field operations. Most literature focuses on algorithmic-level techniques to improve coding efficiency. Optimization with an algorithm/hardware co-designing approach has long been neglected. Leveraging the unique structure of BATS codes, we first present CS-BATS, a hardware-friendly variant. Next we propose a simple but effective bounded-value generator, to reduce the size of a finite field multiplier by up to 70%. Finally, we report on a scalable and resource-efficient FPGA-based network coding accelerator that achieves a throughput of 27 Gbps, a speedup of more than 300 over software.

The 2017 observing campaign of the Event Horizon Telescope (EHT) delivered the first very long baseline interferometry (VLBI) images at the observing frequency of 230 GHz, leading to a number of unique studies on black holes and relativistic jets from active galactic nuclei (AGN). In total, eighteen sources were observed: the main science targets, Sgr A* and M87 along with various calibrators. We investigated the morphology of the sixteen AGN in the EHT 2017 data set, focusing on the properties of the VLBI cores: size, flux density, and brightness temperature. We studied their dependence on the observing frequency in order to compare it with the Blandford-K\"onigl (BK) jet model. We modeled the source structure of seven AGN in the EHT 2017 data set using linearly polarized circular Gaussian components and collected results for the other nine AGN from dedicated EHT publications, complemented by lower frequency data in the 2-86 GHz range. Then, we studied the dependences of the VLBI core flux density, size, and brightness temperature on the frequency measured in the AGN host frame. We compared the observations with the BK jet model and estimated the magnetic field strength dependence on the distance from the central black hole. Our results indicate a deviation from the standard BK model, particularly in the decrease of the brightness temperature with the observing frequency. Either bulk acceleration of the jet material, energy transfer from the magnetic field to the particles, or both are required to explain the observations.

We present the Cosmic Evolution Early Release Science (CEERS) Survey, a 77.2 hour Director's Discretionary Early Release Science Program. CEERS demonstrates, tests, and validates efficient extragalactic surveys using coordinated, overlapping parallel observations with the JWST instrument suite, including NIRCam and MIRI imaging, NIRSpec low (R~100) and medium (R~1000) resolution spectroscopy, and NIRCam slitless grism (R~1500) spectroscopy. CEERS targets the Hubble Space Telescope-observed region of the Extended Groth Strip (EGS) field, supported by a rich set of multiwavelength data. CEERS facilitated immediate community science in both of the extragalactic core JWST science drivers ``First Light" and ``Galaxy Assembly," including: 1) The discovery and characterization of large samples of galaxies at z >~ 10 from ~90 arcmin^2 of NIRCam imaging, constraining their abundance and physical nature; 2) Deep spectra of >1000 galaxies, including dozens of galaxies at 63; and 4) Characterizing galaxy mid-IR emission with MIRI to study dust-obscured star-formation and supermassive black hole growth at z~1-3. As a legacy product for the community, the CEERS team has provided several data releases, accompanied by detailed notes on the data reduction procedures and notebooks to aid in reproducibility. In addition to an overview of the survey and quality of the data, we provide science highlights from the first two years with CEERS data., Comment: 38 pages, 13 figures, 6 tables

In the neighborhood growth dynamics on a Hamming rectangle $[0,m-1]\times[0,n-1]\subseteq \mathbb{Z}_+^2$, the decision to add a point is made by counting the currently occupied points on the horizontal and the vertical line through it, and checking whether the pair of counts lies outside a fixed Young diagram. After the initially occupied set is chosen, the synchronous rule is iterated. The Young domination number with a fixed latency $L$ is the smallest cardinality of an initial set that covers the rectangle by $L$ steps, for $L=0,1,\ldots$ We compute this number for some special cases, including $k$-domination for any $k$ when $m=n$, and devise approximation algorithms in the general case. These results have implications in extremal graph theory, via an equivalence between the case $L = 1$ and bipartite Tur\'an numbers for families of double stars. Our approach is based on a variety of techniques including duality, algebraic formulations, explicit constructions, and dynamic programming.

In Sub-Saharan Africa (SSA), the utilization of lower-quality Magnetic Resonance Imaging (MRI) technology raises questions about the applicability of machine learning methods for clinical tasks. This study aims to provide a robust deep learning-based brain tumor segmentation (BraTS) method tailored for the SSA population using a threefold approach. Firstly, the impact of domain shift from the SSA training data on model efficacy was examined, revealing no significant effect. Secondly, a comparative analysis of 3D and 2D full-resolution models using the nnU-Net framework indicates similar performance of both the models trained for 300 epochs achieving a five-fold cross-validation score of 0.93. Lastly, addressing the performance gap observed in SSA validation as opposed to the relatively larger BraTS glioma (GLI) validation set, two strategies are proposed: fine-tuning SSA cases using the GLI+SSA best-pretrained 2D fullres model at 300 epochs, and introducing a novel neural style transfer-based data augmentation technique for the SSA cases. This investigation underscores the potential of enhancing brain tumor prediction within SSA's unique healthcare landscape.

Operator-splitting methods are widely used to solve differential equations, especially those that arise from multi-scale or multi-physics models, because a monolithic (single-method) approach may be inefficient or even infeasible. The most common operator-splitting methods are the first-order Lie--Trotter (or Godunov) and the second-order Strang (Strang--Marchuk) splitting methods. High-order splitting methods with real coefficients require backward-in-time integration in each operator and hence may be adversely impacted by instability for certain operators such as diffusion. However, besides the method coefficients, there are many other ancillary aspects to an overall operator-splitting method that are important but often overlooked. For example, the operator ordering and the choice of sub-integration methods can significantly affect the stability and efficiency of an operator-splitting method. In this paper, we investigate some design principles for the construction of operator-splitting methods, including minimization of local error measure, choice of sub-integration method, maximization of linear stability, and minimization of overall computational cost. We propose a new four-stage, third-order, 2-split operator-splitting method with seven sub-integrations per step and optimized linear stability for a benchmark problem from cardiac electrophysiology. We then propose a general principle to further improve stability and efficiency of such operator-splitting methods by using low-order, explicit sub-integrators for unstable sub-integrations. We demonstrate an almost 30\% improvement in the performance of methods derived from these design principles compared to the best-known third-order methods., Comment: 20 pages, 5 figures

Health outcomes depend on complex environmental and sociodemographic factors whose effects change over location and time. Only recently has fine-grained spatial and temporal data become available to study these effects, namely the MEDSAT dataset of English health, environmental, and sociodemographic information. Leveraging this new resource, we use a variety of variable importance techniques to robustly identify the most informative predictors across multiple health outcomes. We then develop an interpretable machine learning framework based on Generalized Additive Models (GAMs) and Multiscale Geographically Weighted Regression (MGWR) to analyze both local and global spatial dependencies of each variable on various health outcomes. Our findings identify NO2 as a global predictor for asthma, hypertension, and anxiety, alongside other outcome-specific predictors related to occupation, marriage, and vegetation. Regional analyses reveal local variations with air pollution and solar radiation, with notable shifts during COVID. This comprehensive approach provides actionable insights for addressing health disparities, and advocates for the integration of interpretable machine learning in public health., Comment: AAAI

In the corona, plasma is accelerated to hundreds of kilometers per second, and heated to temperatures hundreds of times hotter than the Sun's surface, before it escapes to form the solar wind. Decades of space-based experiments have shown that the energization process does not stop after it escapes. Instead, the solar wind continues to accelerate and it cools far more slowly than a freely-expanding adiabatic gas. Recent work suggests that fast solar wind requires additional momentum beyond what can be provided by the observed thermal pressure gradients alone whereas it is sufficient for the slowest wind. The additional acceleration for fast wind can be provided through an Alfv\'en wave pressure gradient. Beyond this fast-slow categorization, however, a subset of slow solar wind exhibits high Alfv\'enicity that suggest Alfv\'en waves could play a larger role in its acceleration compared to conventional slow wind outflows. Through a well-timed conjunction between Solar Orbiter and Parker Solar Probe, we trace the energetics of slow wind to compare with a neighboring Alfv\'enic slow solar wind stream. An analysis that integrates remote and heliospheric properties and modeling of the two distinct solar wind streams finds Alfv\'enic slow solar wind behaves like fast wind, where a wave pressure gradient is required to reconcile its full acceleration, while non-Alfv\'enic slow wind can be driven by its non-adiabatic electron and proton thermal pressure gradients. Derived coronal conditions of the source region indicate good model compatibility but extended coronal observations are required to effectively trace solar wind energetics below Parker's orbit., Comment: 20 pages, 9 figures, accepted for publication to ApJ on January 3, 2025

Continuous gravitational waves (CWs) emission from neutron stars carries information about their internal structure and equation of state, and it can provide tests of General Relativity. We present a search for CWs from a set of 45 known pulsars in the first part of the fourth LIGO--Virgo--KAGRA observing run, known as O4a. We conducted a targeted search for each pulsar using three independent analysis methods considering the single-harmonic and the dual-harmonic emission models. We find no evidence of a CW signal in O4a data for both models and set upper limits on the signal amplitude and on the ellipticity, which quantifies the asymmetry in the neutron star mass distribution. For the single-harmonic emission model, 29 targets have the upper limit on the amplitude below the theoretical spin-down limit. The lowest upper limit on the amplitude is $6.4\!\times\!10^{-27}$ for the young energetic pulsar J0537-6910, while the lowest constraint on the ellipticity is $8.8\!\times\!10^{-9}$ for the bright nearby millisecond pulsar J0437-4715. Additionally, for a subset of 16 targets we performed a narrowband search that is more robust regarding the emission model, with no evidence of a signal. We also found no evidence of non-standard polarizations as predicted by the Brans-Dicke theory., Comment: main paper: 12 pages, 6 figures, 4 tables

Partial differential equation (PDE) is an important math tool in science and engineering. This paper experimentally demonstrates an optical neural PDE solver by leveraging the back-propagation-free on-photonic-chip training of physics-informed neural networks.

Applying Large language models (LLMs) within specific domains requires substantial adaptation to account for the unique terminologies, nuances, and context-specific challenges inherent to those areas. Here, we introduce DropMicroFluidAgents (DMFAs), an advanced language-driven framework leveraging state-of-the-art pre-trained LLMs. DMFAs employs LLM agents to perform two key functions: (1) delivering focused guidance, answers, and suggestions specific to droplet microfluidics and (2) generating machine learning models to optimise and automate the design of droplet microfluidic devices, including the creation of code-based computer-aided design (CAD) scripts to enable rapid and precise design execution. Experimental evaluations demonstrated that the integration of DMFAs with the LLAMA3.1 model yielded the highest accuracy of 76.15%, underscoring the significant performance enhancement provided by agent integration. This effect was particularly pronounced when DMFAs were paired with the GEMMA2 model, resulting in a 34.47% improvement in accuracy compared to the standalone GEMMA2 configuration. This study demonstrates the effective use of LLM agents in droplet microfluidics research as powerful tools for automating workflows, synthesising knowledge, optimising designs, and interacting with external systems. These capabilities enable their application across education and industrial support, driving greater efficiency in scientific discovery and innovation.

Despite the remarkable coherence of Large Language Models (LLMs), existing evaluation methods often suffer from fluency bias and rely heavily on multiple-choice formats, making it difficult to assess factual accuracy and complex reasoning effectively. LLMs thus frequently generate factually inaccurate responses, especially in complex reasoning tasks, highlighting two prominent challenges: (1) the inadequacy of existing methods to evaluate reasoning and factual accuracy effectively, and (2) the reliance on human evaluators for nuanced judgment, as illustrated by Williams and Huckle (2024)[1], who found manual grading indispensable despite automated grading advancements. To address evaluation gaps in open-ended reasoning tasks, we introduce the EQUATOR Evaluator (Evaluation of Question Answering Thoroughness in Open-ended Reasoning). This framework combines deterministic scoring with a focus on factual accuracy and robust reasoning assessment. Using a vector database, EQUATOR pairs open-ended questions with human-evaluated answers, enabling more precise and scalable evaluations. In practice, EQUATOR significantly reduces reliance on human evaluators for scoring and improves scalability compared to Williams and Huckle's (2004)[1] methods. Our results demonstrate that this framework significantly outperforms traditional multiple-choice evaluations while maintaining high accuracy standards. Additionally, we introduce an automated evaluation process leveraging smaller, locally hosted LLMs. We used LLaMA 3.2B, running on the Ollama binaries to streamline our assessments. This work establishes a new paradigm for evaluating LLM performance, emphasizing factual accuracy and reasoning ability, and provides a robust methodological foundation for future research.

Tensor Robust Principal Component Analysis (TRPCA) holds a crucial position in machine learning and computer vision. It aims to recover underlying low-rank structures and characterizing the sparse structures of noise. Current approaches often encounter difficulties in accurately capturing the low-rank properties of tensors and balancing the trade-off between low-rank and sparse components, especially in a mixed-noise scenario. To address these challenges, we introduce a Bayesian framework for TRPCA, which integrates a low-rank tensor nuclear norm prior and a generalized sparsity-inducing prior. By embedding the proposed priors within the Bayesian framework, our method can automatically determine the optimal tensor nuclear norm and achieve a balance between the nuclear norm and sparse components. Furthermore, our method can be efficiently extended to the weighted tensor nuclear norm model. Experiments conducted on synthetic and real-world datasets demonstrate the effectiveness and superiority of our method compared to state-of-the-art approaches.

Although the search for quantum many-body phenomena in magnetic materials has a strong focus on highly frustrated systems, even unfrustrated quantum magnets show a multitude of unconventional phenomena in their spin excitation spectra. YbBr$_3$ is an excellent realization of the $S = 1/2$ antiferromagnetic Heisenberg model on the honeycomb lattice, and we have performed detailed spectroscopic experiments with both unpolarized and polarized neutrons at all applied magnetic fields up to saturation. We observe extensive excitation continua, which cause strong renormalization and the decay of single magnons at higher fields, while coherent features include field-induced ``shadows'' of the single magnons and the spectacular emergence of a roton-like excitation. To guide and interpret our experiments, we performed systematic calculations by the method of cylinder matrix-product states that provide quantitative agreement with the neutron scattering data and a qualitative benchmark for the spectral signatures of strong quantum fluctuations even in the absence of magnetic frustration., Comment: 7+8 pages, 4+7 figures

We analyze the spectrum of the Laplace operator, subject to homogeneous complex magnetic fields in the plane. For real magnetic fields, it is well-known that the spectrum consists of isolated eigenvalues of infinite multiplicities (Landau levels). We demonstrate that when the magnetic field has a nonzero imaginary component, the spectrum expands to cover the entire complex plane. Additionally, we show that the Landau levels (appropriately rotated and now embedded in the complex plane) persists, unless the magnetic field is purely imaginary in which case they disappear and the spectrum becomes purely continuous., Comment: Joyeux Noel !

Quantum coherence in superconducting circuits has increased steadily over the last decades as a result of a growing understanding of the various loss mechanisms. Recently, tantalum (Ta) emerged as a promising material to address microscopic sources of loss found on niobium (Nb) or aluminum (Al) surfaces. However, the effects of film and interface microstructure on low-temperature microwave loss are still not well understood. Here we present a systematic study of the structural and electrical properties of Ta and Nb films sputtered on c-plane sapphire at varying growth temperatures. As growth temperature is increased, our results show that the onset of epitaxial growth of $\alpha$-phase Ta correlates with lower Ta surface roughness, higher critical temperature, and higher residual resistivity ratio, but surprisingly also correlates with a significant increase in loss at microwave frequency. We determine that the source of loss is located at the Ta/sapphire interface and show that it can be fully mitigated by either growing a thin, epitaxial Nb inter-layer between the Ta film and the substrate or by intentionally treating the sapphire surface with \textit{in-situ} argon plasma before Ta growth. In addition to elucidating this interfacial microwave loss, this work provides adequate process details that should allow for the reproducible growth of low-loss Ta film across fabrication facilities.

Model immunization is an emerging direction that aims to mitigate the potential risk of misuse associated with open-sourced models and advancing adaptation methods. The idea is to make the released models' weights difficult to fine-tune on certain harmful applications, hence the name ``immunized''. Recent work on model immunization focuses on the single-concept setting. However, models need to be immunized against multiple concepts in real-world situations. To address this gap, we propose an immunization algorithm that, simultaneously, learns a single ``difficult initialization'' for adaptation methods over a set of concepts. We achieve this by incorporating a differentiable merging layer that combines a set of model weights adapted over multiple concepts. In our experiments, we demonstrate the effectiveness of multi-concept immunization by generalizing prior work's experiment setup of re-learning and personalization adaptation to multiple concepts., Comment: AAAI 2025

We propose and implement a comprehensive quantum compilation toolkit for solving the maximum independent set (MIS) problem on quantum hardware based on Rydberg atom arrays. Our end-to-end pipeline involves three core components to efficiently map generic MIS instances onto Rydberg arrays with unit-disk connectivity, with modules for graph reduction, hardware compatibility checks, and graph embedding. The first module (reducer) provides hardware-agnostic and deterministic reduction logic that iteratively reduces the problem size via lazy clique removals. We find that real-world networks can typically be reduced by orders of magnitude on sub-second time scales, thus significantly cutting down the eventual load for quantum devices. Moreover, we show that reduction techniques may be an important tool in the ongoing search for potential quantum speedups, given their ability to identify hard problem instances. In particular, for Rydberg-native MIS instances, we observe signatures of an easy-hard-easy transition and quantify a critical degree indicating the onset of a hard problem regime. The second module (compatibility checker) implements a hardware compatibility checker that quickly determines whether or not a given input graph may be compatible with the restrictions imposed by Rydberg quantum hardware. The third module (embedder) describes hardware-efficient graph embedding routines to generate (approximate) encodings with controllable overhead and optimized ancilla placements. We exemplify our pipeline with experiments run on the QuEra Aquila device available on Amazon Braket. In aggregate, our work provides a set of tools that extends the class of problems that can be tackled with near-term Rydberg atom arrays., Comment: Manuscript: 20 pages, 16 figures. Appendix: 1 page, 3 figures

Consider a graph $G$ with a long path $P$. When is it the case that $G$ also contains a long induced path? This question has been investigated in general as well as within a number of different graph classes since the 80s. We have recently observed in a companion paper (Long induced paths in sparse graphs and graphs with forbidden patterns, arXiv:2411.08685, 2024) that most existing results can recovered in a simple way by considering forbidden ordered patterns of edges along the path $P$. In particular we proved that if we forbid some fixed ordered matching along a path of order $n$ in a graph $G$, then $G$ must contain an induced path of order $(\log n)^{\Omega(1)}$. Moreover, we completely characterized the forbidden ordered patterns forcing the existence of an induced path of polynomial size. The purpose of the present paper is to completely characterize the ordered patterns $H$ such that forbidding $H$ along a path $P$ of order $n$ implies the existence of an induced path of order $(\log n)^{\Omega(1)}$. These patterns are star forests with some specific ordering, which we called constellations. As a direct consequence of our result, we show that if a graph $G$ has a path of length $n$ and does not contain $K_t$ as a topological minor, then $G$ contains an induced path of order $(\log n)^{\Omega(1/t \log^2 t)}$. The previously best known bound was $(\log n)^{f(t)}$ for some unspecified function $f$ depending on the Topological Minor Structure Theorem of Grohe and Marx (2015)., Comment: Some material from arXiv:2304.09679 has been reused (such as definitions and remarks), so text overlap is to be expected

A remarkable characteristic of quantum computing is the potential for reliable computation despite faulty qubits. This can be achieved through quantum error correction, which is typically implemented by repeatedly applying static syndrome checks, permitting correction of logical information. Recently, the development of time-dynamic approaches to error correction has uncovered new codes and new code implementations. In this work, we experimentally demonstrate three time-dynamic implementations of the surface code, each offering a unique solution to hardware design challenges and introducing flexibility in surface code realization. First, we embed the surface code on a hexagonal lattice, reducing the necessary couplings per qubit from four to three. Second, we walk a surface code, swapping the role of data and measure qubits each round, achieving error correction with built-in removal of accumulated non-computational errors. Finally, we realize the surface code using iSWAP gates instead of the traditional CNOT, extending the set of viable gates for error correction without additional overhead. We measure the error suppression factor when scaling from distance-3 to distance-5 codes of $\Lambda_{35,\text{hex}} = 2.15(2)$, $\Lambda_{35,\text{walk}} = 1.69(6)$, and $\Lambda_{35,\text{iSWAP}} = 1.56(2)$, achieving state-of-the-art error suppression for each. With detailed error budgeting, we explore their performance trade-offs and implications for hardware design. This work demonstrates that dynamic circuit approaches satisfy the demands for fault-tolerance and opens new alternative avenues for scalable hardware design., Comment: 11 pages, 5 figures, Supplementary Information

Quantum error correction is essential for bridging the gap between the error rates of physical devices and the extremely low logical error rates required for quantum algorithms. Recent error-correction demonstrations on superconducting processors have focused primarily on the surface code, which offers a high error threshold but poses limitations for logical operations. In contrast, the color code enables much more efficient logic, although it requires more complex stabilizer measurements and decoding techniques. Measuring these stabilizers in planar architectures such as superconducting qubits is challenging, and so far, realizations of color codes have not addressed performance scaling with code size on any platform. Here, we present a comprehensive demonstration of the color code on a superconducting processor, achieving logical error suppression and performing logical operations. Scaling the code distance from three to five suppresses logical errors by a factor of $\Lambda_{3/5}$ = 1.56(4). Simulations indicate this performance is below the threshold of the color code, and furthermore that the color code may be more efficient than the surface code with modest device improvements. Using logical randomized benchmarking, we find that transversal Clifford gates add an error of only 0.0027(3), which is substantially less than the error of an idling error correction cycle. We inject magic states, a key resource for universal computation, achieving fidelities exceeding 99% with post-selection (retaining about 75% of the data). Finally, we successfully teleport logical states between distance-three color codes using lattice surgery, with teleported state fidelities between 86.5(1)% and 90.7(1)%. This work establishes the color code as a compelling research direction to realize fault-tolerant quantum computation on superconducting processors in the near future.

Automating brain tumor segmentation using deep learning methods is an ongoing challenge in medical imaging. Multiple lingering issues exist including domain-shift and applications in low-resource settings which brings a unique set of challenges including scarcity of data. As a step towards solving these specific problems, we propose Convolutional adapter-inspired Parameter-efficient Fine-tuning (PEFT) of MedNeXt architecture. To validate our idea, we show our method performs comparable to full fine-tuning with the added benefit of reduced training compute using BraTS-2021 as pre-training dataset and BraTS-Africa as the fine-tuning dataset. BraTS-Africa consists of a small dataset (60 train / 35 validation) from the Sub-Saharan African population with marked shift in the MRI quality compared to BraTS-2021 (1251 train samples). We first show that models trained on BraTS-2021 dataset do not generalize well to BraTS-Africa as shown by 20% reduction in mean dice on BraTS-Africa validation samples. Then, we show that PEFT can leverage both the BraTS-2021 and BraTS-Africa dataset to obtain mean dice of 0.8 compared to 0.72 when trained only on BraTS-Africa. Finally, We show that PEFT (0.80 mean dice) results in comparable performance to full fine-tuning (0.77 mean dice) which may show PEFT to be better on average but the boxplots show that full finetuning results is much lesser variance in performance. Nevertheless, on disaggregation of the dice metrics, we find that the model has tendency to oversegment as shown by high specificity (0.99) compared to relatively low sensitivity(0.75). The source code is available at https://github.com/CAMERA-MRI/SPARK2024/tree/main/PEFT_MedNeXt, Comment: Accepted to "The International Brain Tumor Segmentation (BraTS) challenge organized at MICCAI 2024 conference"

The development of 6G wireless technologies is rapidly advancing, with the 3rd Generation Partnership Project (3GPP) entering the pre-standardization phase and aiming to deliver the first specifications by 2028. This paper explores the OpenAirInterface (OAI) project, an open-source initiative that plays a crucial role in the evolution of 5G and the future 6G networks. OAI provides a comprehensive implementation of 3GPP and O-RAN compliant networks, including Radio Access Network (RAN), Core Network (CN), and software-defined User Equipment (UE) components. The paper details the history and evolution of OAI, its licensing model, and the various projects under its umbrella, such as RAN, the CN, as well as the Operations, Administration and Maintenance (OAM) projects. It also highlights the development methodology, Continuous Integration/Continuous Delivery (CI/CD) processes, and end-to-end systems powered by OAI. Furthermore, the paper discusses the potential of OAI for 6G research, focusing on spectrum, reflective intelligent surfaces, and Artificial Intelligence (AI)/Machine Learning (ML) integration. The open-source approach of OAI is emphasized as essential for tackling the challenges of 6G, fostering community collaboration, and driving innovation in next-generation wireless technologies., Comment: This work has been submitted to the Elsesvier Computer Networks Journal for possible publication

We investigate primordial non-Gaussianity (NG) arising from the explicit $U(1)$ symmetry-breaking interactions during inflation involving a nearly massless axial component of a complex scalar field $P$. We analyze the induced NG parameter $f_{\mathrm{NL}}$ under scenarios where the axial field functions as either a curvaton or cold dark matter (CDM). In the curvaton framework, there is a conventional contribution to the local NG of $f_{\rm NL} \simeq -O(1)$. Additional positive local NG can result from either the self-interactions of axial field fluctuations, their interactions with a light radial partner, or kinetic mixing with the inflaton via $U(1)$ symmetry-breaking terms. We identify parameter regions where the interactions lead to cancellations, suppressing the overall local NG to $|f^{\rm loc}_{\mathrm{NL}}| \lesssim O(0.1)$, while leaving the trispectrum largely unaffected. In the CDM scenario, these interactions enhance the NG in the isocurvature fluctuations. Moreover, interactions between the axial field and another light scalar, such as a curvaton, can generate $O(1)$ curvature NG signals and significant mixed curvature-isocurvature NGs that are within the reach of future experiments with $\sigma(f^{\rm loc}_{{\rm NL}})\sim1$. We also explore the role of a heavy radial field in generating oscillating correlation signals, noting that such signals can dominate the shape of the mixed adiabatic-isocurvature bispectrum. In certain cases, an oscillatory isocurvature bispectrum signal may be observable in the future, aiding in distinguishing between certain types of the $U(1)$-breaking self-interactions of the axial field., Comment: 46 pages of main text, 5 appendices, 6 figures

As Generative AI continues to become more accessible, the case for robust detection of generated images in order to combat misinformation is stronger than ever. Invisible watermarking methods act as identifiers of generated content, embedding image- and latent-space messages that are robust to many forms of perturbations. The majority of current research investigates full-image attacks against images with a single watermarking method applied. We introduce novel improvements to watermarking robustness as well as minimizing degradation on image quality during attack. Firstly, we examine the application of both image-space and latent-space watermarking methods on a single image, where we propose a custom watermark remover network which preserves one of the watermarking modalities while completely removing the other during decoding. Then, we investigate localized blurring attacks (LBA) on watermarked images based on the GradCAM heatmap acquired from the watermark decoder in order to reduce the amount of degradation to the target image. Our evaluation suggests that 1) implementing the watermark remover model to preserve one of the watermark modalities when decoding the other modality slightly improves on the baseline performance, and that 2) LBA degrades the image significantly less compared to uniform blurring of the entire image. Code is available at: https://github.com/tomputer-g/IDL_WAR, Comment: YouTube link for the presentation: https://www.youtube.com/watch?v=0vwFG1HSrUE

Atmospheric energy transport is central to the cooling of primordial magma oceans. Theoretical studies of atmospheres on lava planets have assumed that convection is the only process involved in setting the atmospheric temperature structure. This significantly influences the ability for a magma ocean to cool. It has been suggested that convective stability in these atmospheres could preclude permanent magma oceans. We develop a new 1D radiative-convective model in order to investigate when the atmospheres overlying magma oceans are convectively stable. Using a coupled interior-atmosphere framework, we simulate the early evolution of two terrestrial-mass exoplanets: TRAPPIST-1 c and HD 63433 d. Our simulations suggest that the atmosphere of HD 63433 d exhibits deep isothermal layers which are convectively stable. However, it is able to maintain a permanent magma ocean and an atmosphere depleted in H2O. It is possible to maintain permanent magma oceans underneath atmospheres without convection. Absorption features of CO2 and SO2 within synthetic emission spectra are associated with mantle redox state, meaning that future observations of HD 63433 d may provide constraints on the geochemical properties of a magma ocean analogous with the early Earth. Simulations of TRAPPIST-1 c indicate that it is expected to have solidified within 100 Myr, outgassing a thick atmosphere in the process. Cool isothermal stratospheres generated by low molecular-weight atmospheres can mimic the emission of an atmosphere-less body. Future work should consider how atmospheric escape and chemistry modulates the lifetime of magma oceans, and the role of tidal heating in sustaining atmospheric convection, Comment: Accepted for publication in MNRAS

Advances in artificial intelligence (AI) present significant risks and opportunities, requiring improved governance to mitigate societal harms and promote equitable benefits. Current incentive structures and regulatory delays may hinder responsible AI development and deployment, particularly in light of the transformative potential of large language models (LLMs). To address these challenges, we propose developing the following three contributions: (1) a large multimodal text and economic-timeseries foundation model that integrates economic and natural language policy data for enhanced forecasting and decision-making, (2) algorithmic mechanisms for eliciting diverse and representative perspectives, enabling the creation of data-driven public policy recommendations, and (3) an AI-driven web platform for supporting transparent, inclusive, and data-driven policymaking.

Federated learning (FL) has emerged as a prominent machine learning paradigm in edge computing environments, enabling edge devices to collaboratively optimize a global model without sharing their private data. However, existing FL frameworks suffer from efficacy deterioration due to the system heterogeneity inherent in edge computing, especially in the presence of domain shifts across local data. In this paper, we propose a heterogeneous FL framework DapperFL, to enhance model performance across multiple domains. In DapperFL, we introduce a dedicated Model Fusion Pruning (MFP) module to produce personalized compact local models for clients to address the system heterogeneity challenges. The MFP module prunes local models with fused knowledge obtained from both local and remaining domains, ensuring robustness to domain shifts. Additionally, we design a Domain Adaptive Regularization (DAR) module to further improve the overall performance of DapperFL. The DAR module employs regularization generated by the pruned model, aiming to learn robust representations across domains. Furthermore, we introduce a specific aggregation algorithm for aggregating heterogeneous local models with tailored architectures and weights. We implement DapperFL on a realworld FL platform with heterogeneous clients. Experimental results on benchmark datasets with multiple domains demonstrate that DapperFL outperforms several state-of-the-art FL frameworks by up to 2.28%, while significantly achieving model volume reductions ranging from 20% to 80%. Our code is available at: https://github.com/jyzgh/DapperFL., Comment: Oral accepted by NeurIPS 2024

The commonsense reasoning capabilities of vision-language models (VLMs), especially in abductive reasoning and defeasible reasoning, remain poorly understood. Most benchmarks focus on typical visual scenarios, making it difficult to discern whether model performance stems from keen perception and reasoning skills, or reliance on pure statistical recall. We argue that by focusing on atypical events in videos, clearer insights can be gained on the core capabilities of VLMs. Explaining and understanding such out-of-distribution events requires models to extend beyond basic pattern recognition and regurgitation of their prior knowledge. To this end, we introduce BlackSwanSuite, a benchmark for evaluating VLMs' ability to reason about unexpected events through abductive and defeasible tasks. Our tasks artificially limit the amount of visual information provided to models while questioning them about hidden unexpected events, or provide new visual information that could change an existing hypothesis about the event. We curate a comprehensive benchmark suite comprising over 3,800 MCQ, 4,900 generative and 6,700 yes/no tasks, spanning 1,655 videos. After extensively evaluating various state-of-the-art VLMs, including GPT-4o and Gemini 1.5 Pro, as well as open-source VLMs such as LLaVA-Video, we find significant performance gaps of up to 32% from humans on these tasks. Our findings reveal key limitations in current VLMs, emphasizing the need for enhanced model architectures and training strategies., Comment: For data, visit https://blackswan.cs.ubc.ca