Search

Personalized outfit recommendation remains a complex challenge, demanding both fashion compatibility understanding and trend awareness. This paper presents a novel framework that harnesses the expressive power of large language models (LLMs) for this task, mitigating their "black box" and static nature through fine-tuning and direct feedback integration. We bridge the item visual-textual gap in items descriptions by employing image captioning with a Multimodal Large Language Model (MLLM). This enables the LLM to extract style and color characteristics from human-curated fashion images, forming the basis for personalized recommendations. The LLM is efficiently fine-tuned on the open-source Polyvore dataset of curated fashion images, optimizing its ability to recommend stylish outfits. A direct preference mechanism using negative examples is employed to enhance the LLM's decision-making process. This creates a self-enhancing AI feedback loop that continuously refines recommendations in line with seasonal fashion trends. Our framework is evaluated on the Polyvore dataset, demonstrating its effectiveness in two key tasks: fill-in-the-blank, and complementary item retrieval. These evaluations underline the framework's ability to generate stylish, trend-aligned outfit suggestions, continuously improving through direct feedback. The evaluation results demonstrated that our proposed framework significantly outperforms the base LLM, creating more cohesive outfits. The improved performance in these tasks underscores the proposed framework's potential to enhance the shopping experience with accurate suggestions, proving its effectiveness over the vanilla LLM based outfit generation., Comment: CIKM 2024

Foundation models pre-trained using self-supervised and weakly-supervised learning have shown powerful transfer learning capabilities on various downstream tasks, including language understanding, text generation, and image recognition. Recently, the Earth observation (EO) field has produced several foundation models pre-trained directly on multispectral satellite imagery (e.g., Sentinel-2) for applications like precision agriculture, wildfire and drought monitoring, and natural disaster response. However, few studies have investigated the ability of these models to generalize to new geographic locations, and potential concerns of geospatial bias -- models trained on data-rich developed countries not transferring well to data-scarce developing countries -- remain. We investigate the ability of popular EO foundation models to transfer to new geographic regions in the agricultural domain, where differences in farming practices and class imbalance make transfer learning particularly challenging. We first select six crop classification datasets across five continents, normalizing for dataset size and harmonizing classes to focus on four major cereal grains: maize, soybean, rice, and wheat. We then compare three popular foundation models, pre-trained on SSL4EO-S12, SatlasPretrain, and ImageNet, using in-distribution (ID) and out-of-distribution (OOD) evaluation. Experiments show that pre-trained weights designed explicitly for Sentinel-2, such as SSL4EO-S12, outperform general pre-trained weights like ImageNet. Furthermore, the benefits of pre-training on OOD data are the most significant when only 10--100 ID training samples are used. Transfer learning and pre-training with OOD and limited ID data show promising applications, as many developing regions have scarce crop type labels. All harmonized datasets and experimental code are open-source and available for download.

E-commerce platforms have a vast catalog of items to cater to their customers' shopping interests. Most of these platforms assist their customers in the shopping process by offering optimized recommendation carousels, designed to help customers quickly locate their desired items. Many models have been proposed in academic literature to generate and enhance the ranking and recall set of items in these carousels. Conventionally, the accompanying carousel title text (header) of these carousels remains static. In most instances, a generic text such as "Items similar to your current viewing" is utilized. Fixed variations such as the inclusion of specific attributes "Other items from a similar seller" or "Items from a similar brand" in addition to "frequently bought together" or "considered together" are observed as well. This work proposes a novel approach to customize the header generation process of these carousels. Our work leverages user-generated reviews that lay focus on specific attributes (aspects) of an item that were favorably perceived by users during their interaction with the given item. We extract these aspects from reviews and train a graph neural network-based model under the framework of a conditional ranking task. We refer to our innovative methodology as Dynamic Text Snippets (DTS) which generates multiple header texts for an anchor item and its recall set. Our approach demonstrates the potential of utilizing user-generated reviews and presents a unique paradigm for exploring increasingly context-aware recommendation systems., Comment: 7 pages, 3 figures, PAIS 2024 (ECAI)

Heterophilous graphs, where dissimilar nodes tend to connect, pose a challenge for graph neural networks (GNNs) as their superior performance typically comes from aggregating homophilous information. Increasing the GNN depth can expand the scope (i.e., receptive field), potentially finding homophily from the higher-order neighborhoods. However, uniformly expanding the scope results in subpar performance since real-world graphs often exhibit homophily disparity between nodes. An ideal way is personalized scopes, allowing nodes to have varying scope sizes. Existing methods typically add node-adaptive weights for each hop. Although expressive, they inevitably suffer from severe overfitting. To address this issue, we formalize personalized scoping as a separate scope classification problem that overcomes GNN overfitting in node classification. Specifically, we predict the optimal GNN depth for each node. Our theoretical and empirical analysis suggests that accurately predicting the depth can significantly enhance generalization. We further propose Adaptive Scope (AS), a lightweight MLP-based approach that only participates in GNN inference. AS encodes structural patterns and predicts the depth to select the best model for each node's prediction. Experimental results show that AS is highly flexible with various GNN architectures across a wide range of datasets while significantly improving accuracy.

Horospherical Schubert varieties are determined. It is shown that the stabilizer of an arbitrary point in a Schubert variety is a strongly solvable algebraic group. The connectedness of this stabilizer subgroup is discussed. Moreover, a new family of spherical varieties, called doubly spherical varieties, is introduced. It is shown that every nearly toric Schubert variety is doubly spherical., Comment: 22 pages

There has been substantial recent concern that pricing algorithms might learn to ``collude.'' Supra-competitive prices can emerge as a Nash equilibrium of repeated pricing games, in which sellers play strategies which threaten to punish their competitors who refuse to support high prices, and these strategies can be automatically learned. In fact, a standard economic intuition is that supra-competitive prices emerge from either the use of threats, or a failure of one party to optimize their payoff. Is this intuition correct? Would preventing threats in algorithmic decision-making prevent supra-competitive prices when sellers are optimizing for their own revenue? No. We show that supra-competitive prices can emerge even when both players are using algorithms which do not encode threats, and which optimize for their own revenue. We study sequential pricing games in which a first mover deploys an algorithm and then a second mover optimizes within the resulting environment. We show that if the first mover deploys any algorithm with a no-regret guarantee, and then the second mover even approximately optimizes within this now static environment, monopoly-like prices arise. The result holds for any no-regret learning algorithm deployed by the first mover and for any pricing policy of the second mover that obtains them profit at least as high as a random pricing would -- and hence the result applies even when the second mover is optimizing only within a space of non-responsive pricing distributions which are incapable of encoding threats. In fact, there exists a set of strategies, neither of which explicitly encode threats that form a Nash equilibrium of the simultaneous pricing game in algorithm space, and lead to near monopoly prices. This suggests that the definition of ``algorithmic collusion'' may need to be expanded, to include strategies without explicitly encoded threats.

In this paper, we extend the results of Grantcharov and Robitaille in 2021 on mixed tensor products and Capelli determinants to the superalgebra setting. Specifically, we construct a family of superalgebra homomorphisms $\varphi_R : U(\mathfrak{gl}(m+1|n)) \rightarrow \mathcal{D}'(m|n) \otimes U(\mathfrak{gl}(m|n))$ for a certain space of differential operators $\mathcal{D}'(m|n)$ indexed by a central element $R$ of $\mathcal{D}'(m|n) \otimes U(\mathfrak{gl}(m|n))$. We then use this homomorphism to determine the image of Gelfand generators of the center of $U(\mathfrak{gl}(m+1|n))$. We achieve this by first relating $\varphi_R$ to the corresponding Harish-Chandra homomorphisms and then proving a super-analog of Newton's formula for $\mathfrak{gl}(m)$ relating Capelli generators and Gelfand generators. We also use the homomorphism $\varphi_R$ to obtain representations of $U(\mathfrak{gl}(m+1|n))$ from those of $U(\mathfrak{gl}(m|n))$, and find conditions under which these inflations are simple. Finally, we show that for a distinguished central element $R_1$ in $\mathcal{D}'(m|n)\otimes U(\mathfrak{gl}(m|n))$, the kernel of $\varphi_{R_1}$ is the ideal of $U(\mathfrak{gl}(m+1|n))$ generated by the first Gelfand invariant $G_1$.

We investigate the redshift evolution of the concentration-mass relationship of dark matter haloes in state-of-the-art cosmological hydrodynamic simulations and their dark-matter-only counterparts. By combining the IllustrisTNG suite and the novel MillenniumTNG simulation, our analysis encompasses a wide range of box size ($50 - 740 \: \rm cMpc$) and mass resolution ($8.5 \times 10^4 - 3.1 \times 10^7 \: \rm M_{\odot}$ per baryonic mass element). This enables us to study the impact of baryons on the concentration-mass relationship in the redshift interval $0

Attention mechanisms are critically important in the advancement of synthetic aperture radar (SAR) automatic target recognition (ATR) systems. Traditional SAR ATR models often struggle with the noisy nature of the SAR data, frequently learning from background noise rather than the most relevant image features. Attention mechanisms address this limitation by focusing on crucial image components, such as the shadows and small parts of a vehicle, which are crucial for accurate target classification. By dynamically prioritizing these significant features, attention-based models can efficiently characterize the entire image with a few pixels, thus enhancing recognition performance. This capability allows for the discrimination of targets from background clutter, leading to more practical and robust SAR ATR models. We show that attention modules increase top-1 accuracy, improve input robustness, and are qualitatively more explainable on the MSTAR dataset.

The Predicted Mean Vote (PMV) index is a widely accepted method in the building automation sector because it can precisely estimate indoor thermal comfort levels depending on a variety of environmental parameters. This study suggests an experimental setup for automated real-time optimization of heating, ventilation and air-conditioning (HVAC) operations in closed spaces utilizing PMV-based modelling and Multilayer perceptron (MLP) based prediction, with an experimental setup which includes ESP32 and BME280 sensor. The main objective of this paper is to employ the MLP algorithm and predicted mean vote model to dynamically predict comfort differences considering the fluctuations in environmental conditions such as temperature and humidity. The proposed method is implemented across various settings, encompassing both an anechoic chamber and laboratory environments. The findings indicate that the proposed approach effectively enhances the management of HVAC systems in confined areas, leading to higher energy efficiency and enhanced indoor thermal comfort. The effectiveness of the suggested solution is supported by experimental validation, which shows that significant energy savings are achieved while maintaining comfort levels for occupants within.

Quantum error correction provides a path to reach practical quantum computing by combining multiple physical qubits into a logical qubit, where the logical error rate is suppressed exponentially as more qubits are added. However, this exponential suppression only occurs if the physical error rate is below a critical threshold. In this work, we present two surface code memories operating below this threshold: a distance-7 code and a distance-5 code integrated with a real-time decoder. The logical error rate of our larger quantum memory is suppressed by a factor of $\Lambda$ = 2.14 $\pm$ 0.02 when increasing the code distance by two, culminating in a 101-qubit distance-7 code with 0.143% $\pm$ 0.003% error per cycle of error correction. This logical memory is also beyond break-even, exceeding its best physical qubit's lifetime by a factor of 2.4 $\pm$ 0.3. We maintain below-threshold performance when decoding in real time, achieving an average decoder latency of 63 $\mu$s at distance-5 up to a million cycles, with a cycle time of 1.1 $\mu$s. To probe the limits of our error-correction performance, we run repetition codes up to distance-29 and find that logical performance is limited by rare correlated error events occurring approximately once every hour, or 3 $\times$ 10$^9$ cycles. Our results present device performance that, if scaled, could realize the operational requirements of large scale fault-tolerant quantum algorithms., Comment: 10 pages, 4 figures, Supplementary Information

The median $M$ of a graph $G$ is the set of vertices with a minimum total distance to all other vertices in the graph. In this paper, we determine the median of Sierpi\'{n}ski triangle graphs. Sierpi\'{n}ski triangle graphs, also known as Sierpi\'{n}ski gasket graphs of order $n$ are graphs formed by contracting all non-clique edges from the Sierpi\'{n}ski graphs of order ($n+1$).

With recent rapid growth in online shopping, AI-powered Engagement Surfaces (ES) have become ubiquitous across retail services. These engagement surfaces perform an increasing range of functions, including recommending new products for purchase, reminding customers of their orders and providing delivery notifications. Understanding the causal effect of engagement surfaces on value driven for customers and businesses remains an open scientific question. In this paper, we develop a dynamic causal model at scale to disentangle value attributable to an ES, and to assess its effectiveness. We demonstrate the application of this model to inform business decision-making by understanding returns on investment in the ES, and identifying product lines and features where the ES adds the most value., Comment: 10 pages, 5 figures. Accepted at Applied Data Science track of KDD 2024, Barcelona, Spain

We investigate vortex dynamics in three magnetic regions, viz., Quiet Sun, Weak Plage, and Strong Plage, using realistic three-dimensional simulations from a comprehensive radiation-MHD code, MURaM. We find that the spatial extents and spatial distribution of vortices vary for different setups even though the photospheric turbulence responsible for generating vortices has similar profiles for all three regions. We investigate kinetic and magnetic swirling strength and find them consistent with the Alfv\'en wave propagation. Using a flux tube expansion model and linear magnetohydrodynamics (MHD) wave theory, we find that the deviation in kinetic swirling strength from the theoretically expected value is the highest for the Strong Plage, least for the Weak Plage, and intermediate for the Quiet Sun at chromospheric heights. It suggests that Weak Plage is the most favoured region for chromospheric swirls, though they are of smaller spatial extents than in Quiet Sun. We also conjecture that vortex interactions within a single flux tube in Strong Plage lead to an energy cascade from larger to smaller vortices that further result in much lower values of kinetic swirling strength than other regions. Fourier spectra of horizontal magnetic fields at 1 Mm height also show the steep cascade from large to smaller scales for Strong Plage. These findings indicate the potential of vortex-induced torsional Alfv\'en waves to travel higher in the atmosphere without damping for weaker magnetic regions such as the Quiet Sun, whereas vortices would result in dissipation and heating due to the vortex interactions in narrow flux tubes for the strongly magnetized regions such as Strong Plage., Comment: accepted for publication in MNRAS

Quantum interconnects facilitate entanglement distribution between non-local computational nodes. For superconducting processors, microwave photons are a natural means to mediate this distribution. However, many existing architectures limit node connectivity and directionality. In this work, we construct a chiral quantum interconnect between two nominally identical modules in separate microwave packages. We leverage quantum interference to emit and absorb microwave photons on demand and in a chosen direction between these modules. We optimize the protocol using model-free reinforcement learning to maximize absorption efficiency. By halting the emission process halfway through its duration, we generate remote entanglement between modules in the form of a four-qubit W state with 62.4 +/- 1.6% (leftward photon propagation) and 62.1 +/- 1.2% (rightward) fidelity, limited mainly by propagation loss. This quantum network architecture enables all-to-all connectivity between non-local processors for modular and extensible quantum computation., Comment: 25 pages, 9 figures, 5 tables

We report the first experimental determination of the interspecies $s$-wave scattering length\,($a_{34}$) between the $2\,^3S_1\,(F=3/2,m_F=3/2)$ state of $^3$He$^*$ and the $2\,^3S_1\,(m_J=1)$ state of $^4$He$^*$. We determine $a_{34}$ by inducing oscillations in a trapped Bose-Einstein condensate of $^4$He$^*$ and measuring the damping rate of these oscillations due to the presence of $^3$He$^*$ atoms. The deduced value of $a_{34}=29\pm3$\,nm is in good agreement with theoretical predictions. The knowledge of this scattering length is important for many fundamental experiments between these helium isotopes.

Self-rewetting fluids (SRFs) (e.g., aqueous solutions of long-chain alcohols) show anomalous nonlinear (quadratic) variations of surface tension with temperature involving a positive gradient, leading to different thermocapillary convection compared to normal fluids (NFs). Moreover, surface-active materials or surfactants can significantly alter interfacial dynamics by their adsorption on fluid interfaces. The coupled effects of temperature- and surfactant-induced Marangoni stresses on migration bubbles in SRFs remain unexplored. We use a robust lattice Boltzmann (LB) method based on central moments to simulate the two-fluid motions, capture interfaces, and compute the transport of energy and surfactant concentration fields, and systematically study the surfactant-laden bubble dynamics in SRFs. When compared to motion of bubbles in normal fluids, in which they continuously migrate without a stationary behavior, our results show that they exhibit dramatically different characteristics in SRFs. Not only is the bubble motion directed from the hot to the cold sides in SRFs, which is opposite to that in NFs, but, more importantly, the bubble attains an equilibrium position. In the absence of surfactants, such an equilibrium position arises at the reference temperature occurring at the center of the domain. The addition of surfactants moves the equilibrium location further upstream, which is controlled by the magnitude of the Gibbs elasticity parameter that determines the magnitude of the surface tension variation with surfactants. The quadratic sensitivity coefficient of surface tension on temperature associated with the SRF modulates this behavior. The lower this quantity, the greater is the role of surfactants modifying the equilibrium position of the bubble in SRF. These findings provide new means to potentially manipulate the bubble dynamics, and especially to tune its equilibrium states., Comment: 50 pages, 26 figures. arXiv admin note: text overlap with arXiv:2210.07501

We discuss the cold dark matter plus massive neutrinos simulations of the MillenniumTNG (MTNG) project, which aim to improve understanding of how well ongoing and future large-scale galaxy surveys will measure neutrino masses. Our largest simulations, $3000\,{\rm Mpc}$ on a side, use $10240^3$ particles of mass $m_{p} = 6.66\times 10^{8}\,h^{-1}{\rm M}_\odot$ to represent cold dark matter, and $2560^3$ to represent a population of neutrinos with summed mass $M_\nu = 100\,{\rm meV}$. Smaller volume runs with $\sim 630\,{\rm Mpc}$ also include cases with $M_\nu = 0\,\textrm{and}\, 300\,{\rm meV}$. All simulations are carried out twice using the paired-and-fixed technique for cosmic variance reduction. We evolve the neutrino component using the particle-based $\delta f$ importance sampling method, which greatly reduces shot noise in the neutrino density field. In addition, we modify the GADGET-4 code to account both for the influence of relativistic and mildly relativistic components on the expansion rate and for non-Newtonian effects on the largest represented simulation scales. This allows us to quantify accurately the impact of neutrinos on basic statistical measures of nonlinear structure formation, such as the matter power spectrum and the halo mass function. We use semi-analytic models of galaxy formation to predict the galaxy population and its clustering properties as a function of summed neutrino mass, finding significant ($\sim 10\%$) impacts on the cosmic star formation rate history, the galaxy mass function, and the clustering strength. This offers the prospect of identifying combinations of summary statistics that are optimally sensitive to the neutrino mass., Comment: 19 pages, 15 figures

Attention-based transformers have been remarkably successful at modeling generative processes across various domains and modalities. In this paper, we study the behavior of transformers on data drawn from \kth Markov processes, where the conditional distribution of the next symbol in a sequence depends on the previous $k$ symbols observed. We observe a surprising phenomenon empirically which contradicts previous findings: when trained for sufficiently long, a transformer with a fixed depth and $1$ head per layer is able to achieve low test loss on sequences drawn from \kth Markov sources, even as $k$ grows. Furthermore, this low test loss is achieved by the transformer's ability to represent and learn the in-context conditional empirical distribution. On the theoretical side, our main result is that a transformer with a single head and three layers can represent the in-context conditional empirical distribution for \kth Markov sources, concurring with our empirical observations. Along the way, we prove that \textit{attention-only} transformers with $O(\log_2(k))$ layers can represent the in-context conditional empirical distribution by composing induction heads to track the previous $k$ symbols in the sequence. These results provide more insight into our current understanding of the mechanisms by which transformers learn to capture context, by understanding their behavior on Markov sources., Comment: 29 pages, 10 figures

In recent years, Large Language Models (LLM) have demonstrated impressive capabilities in the field of natural language processing (NLP). This paper explores the application of LLMs in negotiation transcript analysis by the Vanderbilt AI Negotiation Lab. Starting in September 2022, we applied multiple strategies using LLMs from zero shot learning to fine tuning models to in-context learning). The final strategy we developed is explained, along with how to access and use the model. This study provides a sense of both the opportunities and roadblocks for the implementation of LLMs in real life applications and offers a model for how LLMs can be applied to coding in other fields.

Farmers face various challenges when it comes to identifying diseases in rice leaves during their early stages of growth, which is a major reason for poor produce. Therefore, early and accurate disease identification is important in agriculture to avoid crop loss and improve cultivation. In this research, we propose a novel hybrid deep learning (DL) classifier designed by extending the Squeeze-and-Excitation network architecture with a channel attention mechanism and the Swish ReLU activation function. The channel attention mechanism in our proposed model identifies the most important feature channels required for classification during feature extraction and selection. The dying ReLU problem is mitigated by utilizing the Swish ReLU activation function, and the Squeeze-andExcitation blocks improve information propagation and cross-channel interaction. Upon evaluation, our model achieved a high F1-score of 99.76% and an accuracy of 99.74%, surpassing the performance of existing models. These outcomes demonstrate the potential of state-of-the-art DL techniques in agriculture, contributing to the advancement of more efficient and reliable disease detection systems., Comment: 17 pages, 4 tables, 10 figures

Large Language Models such as GPTs (Generative Pre-trained Transformers) exhibit remarkable capabilities across a broad spectrum of applications. Nevertheless, due to their intrinsic complexity, these models present substantial challenges in interpreting their internal decision-making processes. This lack of transparency poses critical challenges when it comes to their adaptation by financial institutions, where concerns and accountability regarding bias, fairness, and reliability are of paramount importance. Mechanistic interpretability aims at reverse engineering complex AI models such as transformers. In this paper, we are pioneering the use of mechanistic interpretability to shed some light on the inner workings of large language models for use in financial services applications. We offer several examples of how algorithmic tasks can be designed for compliance monitoring purposes. In particular, we investigate GPT-2 Small's attention pattern when prompted to identify potential violation of Fair Lending laws. Using direct logit attribution, we study the contributions of each layer and its corresponding attention heads to the logit difference in the residual stream. Finally, we design clean and corrupted prompts and use activation patching as a causal intervention method to localize our task completion components further. We observe that the (positive) heads $10.2$ (head $2$, layer $10$), $10.7$, and $11.3$, as well as the (negative) heads $9.6$ and $10.6$ play a significant role in the task completion.

Alzheimer's disease (AD) represents the primary form of neurodegeneration, impacting millions of individuals each year and causing progressive cognitive decline. Accurately diagnosing and classifying AD using neuroimaging data presents ongoing challenges in medicine, necessitating advanced interventions that will enhance treatment measures. In this research, we introduce a dual attention enhanced deep learning (DL) framework for classifying AD from neuroimaging data. Combined spatial and self-attention mechanisms play a vital role in emphasizing focus on neurofibrillary tangles and amyloid plaques from the MRI images, which are difficult to discern with regular imaging techniques. Results demonstrate that our model yielded remarkable performance in comparison to existing state of the art (SOTA) convolutional neural networks (CNNs), with an accuracy of 99.1%. Moreover, it recorded remarkable metrics, with an F1-Score of 99.31%, a precision of 99.24%, and a recall of 99.5%. These results highlight the promise of cutting edge DL methods in medical diagnostics, contributing to highly reliable and more efficient healthcare solutions., Comment: 18 pages, 10 figures, 6 tables

Breast cancer poses a profound threat to lives globally, claiming numerous lives each year. Therefore, timely detection is crucial for early intervention and improved chances of survival. Accurately diagnosing and classifying breast tumors using ultrasound images is a persistent challenge in medicine, demanding cutting-edge solutions for improved treatment strategies. This research introduces multiattention-enhanced deep learning (DL) frameworks designed for the classification and segmentation of breast cancer tumors from ultrasound images. A spatial channel attention mechanism is proposed for segmenting tumors from ultrasound images, utilizing a novel LinkNet DL framework with an InceptionResNet backbone. Following this, the paper proposes a deep convolutional neural network with an integrated multi-attention framework (DCNNIMAF) to classify the segmented tumor as benign, malignant, or normal. From experimental results, it is observed that the segmentation model has recorded an accuracy of 98.1%, with a minimal loss of 0.6%. It has also achieved high Intersection over Union (IoU) and Dice Coefficient scores of 96.9% and 97.2%, respectively. Similarly, the classification model has attained an accuracy of 99.2%, with a low loss of 0.31%. Furthermore, the classification framework has achieved outstanding F1-Score, precision, and recall values of 99.1%, 99.3%, and 99.1%, respectively. By offering a robust framework for early detection and accurate classification of breast cancer, this proposed work significantly advances the field of medical image analysis, potentially improving diagnostic precision and patient outcomes., Comment: 29 pages, 15 figures, 6 tables

The physics of superconductivity in magic-angle twisted bilayer graphene (MATBG) is a topic of keen interest in moir\'e systems research, and it may provide insight into the pairing mechanism of other strongly correlated materials such as high-$T_{\mathrm{c}}$ superconductors. Here, we use DC-transport and microwave circuit quantum electrodynamics (cQED) to measure directly the superfluid stiffness of superconducting MATBG via its kinetic inductance. We find the superfluid stiffness to be much larger than expected from conventional single-band Fermi liquid theory; rather, it aligns well with theory involving quantum geometric effects that are dominant at the magic angle. The temperature dependence of the superfluid stiffness exhibits a power-law behavior, which contraindicates an isotropic BCS model; instead, the extracted power-law exponents indicate an anisotropic superconducting gap, whether interpreted using the conventional anisotropic BCS model or a quantum geometric theory of flat-band superconductivity. Moreover, the quadratic dependence of the stiffness on both DC and microwave current is consistent with Ginzburg-Landau theory. Taken together, these findings strongly suggest a connection between quantum geometry, superfluid stiffness, and unconventional superconductivity in MATBG. Finally, the combined DC-microwave measurement platform used here is applicable to the investigation of other atomically thin superconductors.

We show that a Bose-Einstein Condensate (BEC) launched with non-zero initial momentum into a periodically kicked optical lattice creates an asymmetrically localized momentum distribution in a moving frame with a small initial current. This asymmetric localization is investigated under two scenarios; (a) when the BEC is in motion in the laboratory frame and, (b) when the optical lattice is in motion in the laboratory frame. The asymmetric features are shown to arise from the early-time dynamics induced by the broken parity symmetry and, asymptotically, freeze as the dynamical localization stabilizes. The micromotion of BEC is measured using the early-time asymmetry. In this context, micromotion refers to the extremely low initial velocity of the BEC along the lattice direction. This originates from the jitter when the hybrid trap potential is turned off. By employing BEC in a kicked and moving optical lattice, the asymmetry in early-time dynamics is measured to precisely characterize and quantify the micromotion phenomena in the quantum system. Micromotion measurement has applications in quantifying systematic shifts and uncertainties in light-pulse interferometers.

Although Deligne's theorem classifies all symmetric tensor categories (STCs) with moderate growth over algebraically closed fields of characteristic zero, the classification does not extend to positive characteristic. At the forefront of the study of STCs is the search for an analog to Deligne's theorem in positive characteristic, and it has become increasingly apparent that the Verlinde categories are to play a significant role. Moreover, these categories are largely unstudied, but have already shown very interesting phenomena as both a generalization of and a departure from superalgebra and supergeometry. In this paper, we study $\mathrm{Ver}_4^+$, the simplest non-trivial Verlinde category in characteristic $2$. In particular, we classify all isomorphism classes of non-degenerate symmetric bilinear forms and non-degenerate quadratic forms and study the associated Witt semi-ring that arises from the addition and multiplication operations on bilinear forms.

Deriving meaningful representations from complex, high-dimensional data in unsupervised settings is crucial across diverse machine learning applications. This paper introduces a framework for multi-scale graph network embedding based on spectral graph wavelets that employs a contrastive learning approach. A significant feature of the proposed embedding is its capacity to establish a correspondence between the embedding space and the input feature space which aids in deriving feature importance of the original features. We theoretically justify our approach and demonstrate that, in Paley-Wiener spaces on combinatorial graphs, the spectral graph wavelets operator offers greater flexibility and better control over smoothness properties compared to the Laplacian operator. We validate the effectiveness of our proposed graph embedding on a variety of public datasets through a range of downstream tasks, including clustering and unsupervised feature importance.

Understanding how interacting particles approach thermal equilibrium is a major challenge of quantum simulators. Unlocking the full potential of such systems toward this goal requires flexible initial state preparation, precise time evolution, and extensive probes for final state characterization. We present a quantum simulator comprising 69 superconducting qubits which supports both universal quantum gates and high-fidelity analog evolution, with performance beyond the reach of classical simulation in cross-entropy benchmarking experiments. Emulating a two-dimensional (2D) XY quantum magnet, we leverage a wide range of measurement techniques to study quantum states after ramps from an antiferromagnetic initial state. We observe signatures of the classical Kosterlitz-Thouless phase transition, as well as strong deviations from Kibble-Zurek scaling predictions attributed to the interplay between quantum and classical coarsening of the correlated domains. This interpretation is corroborated by injecting variable energy density into the initial state, which enables studying the effects of the eigenstate thermalization hypothesis (ETH) in targeted parts of the eigenspectrum. Finally, we digitally prepare the system in pairwise-entangled dimer states and image the transport of energy and vorticity during thermalization. These results establish the efficacy of superconducting analog-digital quantum processors for preparing states across many-body spectra and unveiling their thermalization dynamics.

We introduce the RIGEL model, a novel framework to self-consistently model the effects of stellar feedback in the multiphase ISM of dwarf galaxies with radiative transfer (RT) on a star-by-star basis. The RIGEL model integrates detailed implementations of feedback from individual massive stars into the RHD code, AREPO-RT. It forms individual massive stars from the resolved multiphase ISM by sampling the IMF and tracks their evolution individually. The lifetimes, photon production rates, mass-loss rates, and wind velocities of these stars are determined by their initial masses and metallicities based on a library that incorporates a variety of stellar models. The RT equations are solved in seven spectral bins accounting for the IR to HeII ionizing bands, using an M1 RT scheme. The thermochemistry model tracks the non-equilibrium H, He chemistry and the equilibrium abundance of CI, CII, OI, OII, and CO to capture the thermodynamics of all ISM phases. We evaluated the performance of the RIGEL model using $1\,{\rm M}_\odot$ resolution simulations of isolated dwarf galaxies. We found that the SFR and ISRF show strong positive correlations to the metallicity of the galaxy. Photoionization and photoheating can reduce the SFR by an order of magnitude by removing the available cold-dense gas fuel for star formation. The ISRF also changes the thermal structure of the ISM. Radiative feedback occurs immediately after the birth of massive stars and rapidly disperses the molecular clouds within 1 Myr. As a consequence, radiative feedback reduces the age spread of star clusters to less than 2 Myr, prohibits the formation of massive star clusters, and shapes the cluster initial mass function to a steep power-law form with a slope of $\sim-2$. The mass-loading factor of the fiducial galaxy has a median of $\sim50$, while turning off radiative feedback reduces this factor by an order of magnitude., Comment: 27 pages, 18 figures; A&A in press; abstract slightly abridged; comments welcome

Sparse Matricized Tensor Times Khatri-Rao Product (spMTTKRP) is the bottleneck kernel of sparse tensor decomposition. In this work, we propose a GPU-based algorithm design to address the key challenges in accelerating spMTTKRP computation, including (1) eliminating global atomic operations across GPU thread blocks, (2) avoiding the intermediate values being communicated between GPU thread blocks and GPU global memory, and (3) ensuring a balanced distribution of workloads across GPU thread blocks. Our approach also supports dynamic tensor remapping, enabling the above optimizations in all the modes of the input tensor. Our approach achieves a geometric mean speedup of 1.5x, 2.0x, and 21.7x in total execution time across widely used datasets compared with the state-of-the-art GPU implementations. Our work is the only GPU implementation that can support tensors with modes greater than 4 since the state-of-the-art works have implementation constraints for tensors with a large number of modes., Comment: In 21st ACM International Conference on Computing Frontiers (CF '24), May 7-9, 2024, Ischia, Italy

Quantum resonance in the paradigmatic kicked rotor system is a purely quantum effect that ignores the state of underlying classical chaos. In this work, it is shown that quantum resonance leads to superlinear entanglement production. In $N$-interacting kicked rotors set to be at quantum resonance, entanglement growth is super-linear until a crossover timescale $t^*$, beyond which growth slows down to a logarithmic form with superimposed oscillations. By mapping positional interaction to momentum space and analytically assessing the linear entropy, we unravel the mechanism driving these two distinct growth profiles. The analytical results agree with the numerical simulations performed for two- and three-interacting kicked rotors. The late time entanglement oscillation is sensitive to changes in scaled Planck's constant with a high quality factor suitable for high precision measurements. These results are amenable to an experimental realization on atom optics setup., Comment: 6 pages, 3 figures

Fluxonium qubits are recognized for their high coherence times and high operation fidelities, attributed to their unique design incorporating over 100 Josephson junctions per superconducting loop. However, this complexity poses significant fabrication challenges, particularly in achieving high yield and junction uniformity with traditional methods. Here, we introduce an overlap process for Josephson junction fabrication that achieves nearly 100% yield and maintains uniformity across a 2-inch wafer with less than 5% variation for the phase slip junction and less than 2% for the junction array. Our compact junction array design facilitates fluxonium qubits with energy relaxation times exceeding 1 millisecond at the flux frustration point, demonstrating consistency with state-of-the-art dielectric loss tangents and flux noise across multiple devices. This work suggests the scalability of high coherence fluxonium processors using CMOS-compatible processes, marking a significant step towards practical quantum computing.

Biomedical data is growing exponentially, and managing it is increasingly challenging. While Findable, Accessible, Interoperable and Reusable (FAIR) data principles provide guidance, their adoption has proven difficult, especially in larger enterprises like pharmaceutical companies. In this manuscript, we describe how we leverage an Ontology-Based Data Management (OBDM) strategy for digital transformation in Novo Nordisk Research & Early Development. Here, we include both our technical blueprint and our approach for organizational change management. We further discuss how such an OBDM ecosystem plays a pivotal role in the organizations digital aspirations for data federation and discovery fuelled by artificial intelligence. Our aim for this paper is to share the lessons learned in order to foster dialogue with parties navigating similar waters while collectively advancing the efforts in the fields of data management, semantics and data driven drug discovery., Comment: 14 pages, 2 figures

Blockchain applications often rely on lightweight clients to access and verify on-chain data efficiently without the need to run a resource-intensive full node. These light clients must maintain robust security to protect the blockchain's integrity for users of applications built upon it, achieving this with minimal resources and without significant latency. Moreover, different applications have varying security needs. This work focuses on addressing these two key requirements in the context of Proof-of-Stake (PoS) blockchains and identifying the fundamental cost-latency trade-offs to achieve tailored, optimal security for each light client. The key security guarantee of PoS blockchains is economic (implied by the "stake"). In this paper we formalize this cryptoeconomic security to light clients, ensuring that the cost of corrupting the data provided to light clients must outweigh the potential profit, thereby economically deterring malicious actors. We further introduce "insured" cryptoeconomic security to light clients, providing unconditional protection via the attribution of adversarial actions and the consequent slashing of stakes. The divisible and fungible nature of stake facilitates programmable security, allowing for customization of the security level and insurance amount according to the specific needs of different applications. We implemented the protocols in less than 1000 lines of Solidity and TypeScript code and evaluated their gas cost, latency, and the computational overhead. For example, for a transaction with value of \$32k, the light client can choose between zero cost with a latency of 5 hours or instant confirmation with an insurance cost of \$7.45. Thus, the client can select the optimal point on the latency-cost trade-off spectrum that best aligns with its needs. Light clients require negligible storage and face minimal computational costs,...

This study investigates the robustness of graph embedding methods for community detection in the face of network perturbations, specifically edge deletions. Graph embedding techniques, which represent nodes as low-dimensional vectors, are widely used for various graph machine learning tasks due to their ability to capture structural properties of networks effectively. However, the impact of perturbations on the performance of these methods remains relatively understudied. The research considers state-of-the-art graph embedding methods from two families: matrix factorization (e.g., LE, LLE, HOPE, M-NMF) and random walk-based (e.g., DeepWalk, LINE, node2vec). Through experiments conducted on both synthetic and real-world networks, the study reveals varying degrees of robustness within each family of graph embedding methods. The robustness is found to be influenced by factors such as network size, initial community partition strength, and the type of perturbation. Notably, node2vec and LLE consistently demonstrate higher robustness for community detection across different scenarios, including networks with degree and community size heterogeneity. These findings highlight the importance of selecting an appropriate graph embedding method based on the specific characteristics of the network and the task at hand, particularly in scenarios where robustness to perturbations is crucial., Comment: 17 pages, 26 figures, 3 tables. Comments are welcome

Radiative transfer (RT) is a crucial ingredient for self-consistent modelling of numerous astrophysical phenomena across cosmic history. However, on-the-fly integration into radiation-hydrodynamics (RHD) simulations is computationally demanding, particularly due to the stringent time-stepping conditions and increased dimensionality inherent in multi-frequency collisionless Boltzmann physics. The emergence of exascale supercomputers, equipped with extensive CPU cores and GPU accelerators, offers new opportunities for enhancing RHD simulations. We present a novel optimization of AREPO-RT explicitly tailored for such high-performance computing environments. We implement a novel node-to-node communication strategy that utilizes shared memory to substitute intra-node communication with direct memory access. Furthermore, combining multiple inter-node messages into a single message substantially enhances network bandwidth utilization and performance for large-scale simulations on modern supercomputers. The single-message node-to-node approach also improves performance on smaller-scale machines with less optimized networks. Furthermore, by transitioning all RT-related calculations to GPUs, we achieve a significant computational speedup of around 15 for standard benchmarks compared to the original CPU implementation. As a case study, we perform cosmological RHD simulations of the Epoch of Reionization, employing a similar setup as the THESAN project. In this context, RT becomes sub-dominant such that even without modifying the core AREPO codebase, there is an overall threefold improvement in efficiency. The advancements presented here have broad implications, potentially transforming the complexity and scalability of future simulations for a wide variety of astrophysical studies., Comment: 20 pages, 21 Figures, submitted to MNRAS

The rapid evolution of text-to-image diffusion models has opened the door of generative AI, enabling the translation of textual descriptions into visually compelling images with remarkable quality. However, a persistent challenge within this domain is the optimization of prompts to effectively convey abstract concepts into concrete objects. For example, text encoders can hardly express "peace", while can easily illustrate olive branches and white doves. This paper introduces a novel approach named Prompt Optimizer for Abstract Concepts (POAC) specifically designed to enhance the performance of text-to-image diffusion models in interpreting and generating images from abstract concepts. We propose a Prompt Language Model (PLM), which is initialized from a pre-trained language model, and then fine-tuned with a curated dataset of abstract concept prompts. The dataset is created with GPT-4 to extend the abstract concept to a scene and concrete objects. Our framework employs a Reinforcement Learning (RL)-based optimization strategy, focusing on the alignment between the generated images by a stable diffusion model and optimized prompts. Through extensive experiments, we demonstrate that our proposed POAC significantly improves the accuracy and aesthetic quality of generated images, particularly in the description of abstract concepts and alignment with optimized prompts. We also present a comprehensive analysis of our model's performance across diffusion models under different settings, showcasing its versatility and effectiveness in enhancing abstract concept representation., Comment: WWW 2024 Companion

While there has been a large body of research attempting to circumvent tokenization for language modeling (Clark et al., 2022; Xue et al., 2022), the current consensus is that it is a necessary initial step for designing state-of-the-art performant language models. In this paper, we investigate tokenization from a theoretical point of view by studying the behavior of transformers on simple data generating processes. When trained on data drawn from certain simple $k^{\text{th}}$-order Markov processes for $k > 1$, transformers exhibit a surprising phenomenon - in the absence of tokenization, they empirically fail to learn the right distribution and predict characters according to a unigram model (Makkuva et al., 2024). With the addition of tokenization, however, we empirically observe that transformers break through this barrier and are able to model the probabilities of sequences drawn from the source near-optimally, achieving small cross-entropy loss. With this observation as starting point, we study the end-to-end cross-entropy loss achieved by transformers with and without tokenization. With the appropriate tokenization, we show that even the simplest unigram models (over tokens) learnt by transformers are able to model the probability of sequences drawn from $k^{\text{th}}$-order Markov sources near optimally. Our analysis provides a justification for the use of tokenization in practice through studying the behavior of transformers on Markovian data., Comment: 58 pages, 10 figures

Graph neural networks (GNNs) have recently empowered various novel computer vision (CV) tasks. In GNN-based CV tasks, a combination of CNN layers and GNN layers or only GNN layers are employed. This paper introduces GCV-Turbo, a domain-specific accelerator on FPGA for end-to-end acceleration of GNN-based CV tasks. GCV-Turbo consists of two key components: (1) a \emph{novel} hardware architecture optimized for the computation kernels in both CNNs and GNNs using the same set of computation resources. (2) a PyTorch-compatible compiler that takes a user-defined model as input, performs end-to-end optimization for the computation graph of a given GNN-based CV task, and produces optimized code for hardware execution. The hardware architecture and the compiler work synergistically to support a variety of GNN-based CV tasks. We implement GCV-Turbo on a state-of-the-art FPGA and evaluate its performance across six representative GNN-based CV tasks with diverse input data modalities (e.g., image, human skeleton, point cloud). Compared with state-of-the-art CPU (GPU) implementations, GCV-Turbo achieves an average latency reduction of $68.4\times$ ($4.1\times$) on these six GNN-based CV tasks. Moreover, GCV-Turbo supports the execution of the standalone CNNs or GNNs, achieving performance comparable to that of state-of-the-art CNN (GNN) accelerators for widely used CNN-only (GNN-only) models.

Convolutional Neural Networks (CNNs) have demonstrated remarkable ability throughout the field of computer vision. However, CNN inference requires a large number of arithmetic operations, making them expensive to deploy in hardware. Current approaches alleviate this issue by developing hardware-supported, algorithmic processes to simplify spatial convolution functions. However, these methods still heavily rely on matrix multiplication, leading to significant computational overhead. To bridge the gap between hardware, algorithmic acceleration, and approximate matrix multiplication, we propose TabConv, a novel, table-based approximation for convolution to significantly reduce arithmetic operations during inference. Additionally, we introduce a priority masking technique based on cosine similarity to select layers for table-based approximation, thereby maintaining the model performance. We evaluate our approach on popular CNNs: ResNet-18, ResNet-34, and NetworkInNetwork (NIN). TabConv preserves over 93% of the original model's performance while reducing arithmetic operations by 36.5%, 25.8%, and 99.4% for ResNet-18 on CIFAR-10, CIFAR-100, and MNIST, respectively, 35.6% and 99.3% for ResNet-34 on CIFAR-10 and MNIST, and 98.9% for NIN on MNIST, achieving low-computation inference., Comment: 8 pages, Accepted at CF '24

Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR) is a key technique used in military applications like remote-sensing image recognition. Vision Transformers (ViTs) are the current state-of-the-art in various computer vision applications, outperforming their CNN counterparts. However, using ViTs for SAR ATR applications is challenging due to (1) standard ViTs require extensive training data to generalize well due to their low locality; the standard SAR datasets, however, have a limited number of labeled training data which reduces the learning capability of ViTs; (2) ViTs have a high parameter count and are computation intensive which makes their deployment on resource-constrained SAR platforms difficult. In this work, we develop a lightweight ViT model that can be trained directly on small datasets without any pre-training by utilizing the Shifted Patch Tokenization (SPT) and Locality Self-Attention (LSA) modules. We directly train this model on SAR datasets which have limited training samples to evaluate its effectiveness for SAR ATR applications. We evaluate our proposed model, that we call VTR (ViT for SAR ATR), on three widely used SAR datasets: MSTAR, SynthWakeSAR, and GBSAR. Further, we propose a novel FPGA accelerator for VTR, in order to enable deployment for real-time SAR ATR applications., Comment: SPIE DCS 2024

Deep Learning (DL) Models for Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR), while delivering improved performance, have been shown to be quite vulnerable to adversarial attacks. Existing works improve robustness by training models on adversarial samples. However, by focusing mostly on attacks that manipulate images randomly, they neglect the real-world feasibility of such attacks. In this paper, we propose FACTUAL, a novel Contrastive Learning framework for Adversarial Training and robust SAR classification. FACTUAL consists of two components: (1) Differing from existing works, a novel perturbation scheme that incorporates realistic physical adversarial attacks (such as OTSA) to build a supervised adversarial pre-training network. This network utilizes class labels for clustering clean and perturbed images together into a more informative feature space. (2) A linear classifier cascaded after the encoder to use the computed representations to predict the target labels. By pre-training and fine-tuning our model on both clean and adversarial samples, we show that our model achieves high prediction accuracy on both cases. Our model achieves 99.7% accuracy on clean samples, and 89.6% on perturbed samples, both outperforming previous state-of-the-art methods., Comment: 2024 IEEE Radar Conference

The Steinberg tensor product theorem is a fundamental result in the modular representation theory of reductive algebraic groups. It describes any finite-dimensional simple module of highest weight $\lambda$ over such a group as the tensor product of Frobenius twists of simple modules with highest weights the weights appearing in a $p$-adic decomposition of $\lambda$, thereby reducing the character problem to a a finite collection of weights. In recent years this theorem has been extended to various quasi-reductive supergroup schemes. In this paper, we prove the analogous result for the general linear group scheme $GL(X)$ for any object $X$ in the Verlinde category $\mathrm{Ver}_p$.

Kac's ten-dimensional simple Jordan superalgebra over a field of characteristic 5 is obtained from a process of semisimplification, via tensor categories, from the exceptional simple Jordan algebra (or Albert algebra), together with a suitable order 5 automorphism. This explains McCrimmon's 'bizarre result' asserting that, in characteristic 5, Kac's superalgebra is a sort of 'degree 3 Jordan superalgebra'. As an outcome, the exceptional simple Lie superalgebra el(5;5), specific of characteristic 5, is obtained from the simple Lie algebra of type $E_8$ and an order 5 automorphism. In the process, precise recipes to obtain superalgebras from algebras in the category of representations of the cyclic group $C_p$, over a field of characteristic $p>2$, are given., Comment: 22 pages