Search

Following the success in NLP, the best vision models are now in the billion parameter ranges. Adapting these large models to a target distribution has become computationally and economically prohibitive. Addressing this challenge, we introduce UpStep, an Unsupervised Parameter-efficient Source-free post-pretraining approach, designed to efficiently adapt a base model from a source domain to a target domain: i) we design a self-supervised training scheme to adapt a pretrained model on an unlabeled target domain in a setting where source domain data is unavailable. Such source-free setting comes with the risk of catastrophic forgetting, hence, ii) we propose center vector regularization (CVR), a set of auxiliary operations that minimize catastrophic forgetting and additionally reduces the computational cost by skipping backpropagation in 50\% of the training iterations. Finally iii) we perform this adaptation process in a parameter-efficient way by adapting the pretrained model through low-rank adaptation methods, resulting in a fraction of parameters to optimize. We utilize various general backbone architectures, both supervised and unsupervised, trained on Imagenet as our base model and adapt them to a diverse set of eight target domains demonstrating the adaptability and generalizability of our proposed approach.

We report the measurement, using angle-resolved photoemission spectroscopy, of the metallic electronic structure of the hole-doped thermoelectric oxide CuRh$_{0.9}$Mg$_{0.1}$O$_2$. The material is found to have a ``pudding mold'' type band structure, with a nearly flat band edge located near the Fermi level, which is thought to be the origin of the thermoelectric behavior of this material. The experimental data match the density functional theory of the undoped parent compound, simply corrected by a rigid shift of the bands. Transport calculations based on the observed band structure yield a Seebeck coefficient of $\sim 200 \,\mu$V/K for the undoped parent material, consistent with experimental measurements. Our results show that CuRhO$_2$ is a textbook example of how pure band-structural effects can result in a large thermoelectric figure of merit, demonstrating that flat band edges in oxides are a realistic route for the efficient conversion of thermal energy., Comment: Accepted to Physical Review Materials

Significant obstacles exist in scientific domains including genetics, climate modeling, and astronomy due to the management, preprocess, and training on complicated data for deep learning. Even while several large-scale solutions offer distributed execution environments, open-source alternatives that integrate scalable runtime tools, deep learning and data frameworks on high-performance computing platforms remain crucial for accessibility and flexibility. In this paper, we introduce Deep Radical-Cylon(RC), a heterogeneous runtime system that combines data engineering, deep learning frameworks, and workflow engines across several HPC environments, including cloud and supercomputing infrastructures. Deep RC supports heterogeneous systems with accelerators, allows the usage of communication libraries like MPI, GLOO and NCCL across multi-node setups, and facilitates parallel and distributed deep learning pipelines by utilizing Radical Pilot as a task execution framework. By attaining an end-to-end pipeline including preprocessing, model training, and postprocessing with 11 neural forecasting models (PyTorch) and hydrology models (TensorFlow) under identical resource conditions, the system reduces 3.28 and 75.9 seconds, respectively. The design of Deep RC guarantees the smooth integration of scalable data frameworks, such as Cylon, with deep learning processes, exhibiting strong performance on cloud platforms and scientific HPC systems. By offering a flexible, high-performance solution for resource-intensive applications, this method closes the gap between data preprocessing, model training, and postprocessing., Comment: 13 pages, 9 figures, 4 tables

We investigate the properties of neutron stars with antikaon condensation in the framework of the Relativistic Mean-Field (RMF) model with a $\sigma$-cut potential. The well-known RMF models, TM1 and TM1e, are used to analyze the structure and composition of neutron stars. The antikaon condensation part of the equation of state (EoS) is constrained from the experimental data of K$^{-}$ atomic and kaon-nucleon scattering. The $\sigma$-cut potential, which is known to make the EoS stiffer at high densities, is modulated by a free parameter $f_{s}$. Our present analysis suggests that one can obtain neutron star configurations heavier than 2$M_{\odot}$ with antikaon condensates in most cases for $f_{s}$ = 0.6. The antikaon phase transition is a second-order for $f_{s}$ = 0.6 for both TM1 and TM1e parameter sets. The calculated global properties of neutron stars with antikaon condensates i.e., mass and radius seem to be in resonable agreement with other theoretical and observational data., Comment: 10 pages, 10 figures Accepted in Physical Review C (PRC) on 25-02-2025

Large Language Models (LLMs) have demonstrated remarkable capabilities in code generation from natural language prompts, revolutionizing software development workflows. As we advance towards agent-based development paradigms, these models form the cornerstone of next-generation software development lifecycles. However, current benchmarks for evaluating multilingual code generation capabilities are predominantly English-centric, limiting their applicability across the global developer community. To address this limitation, we present IndicEval-XL, a comprehensive benchmark for code generation that incorporates 6 major Indic languages, collectively spoken by approximately 14\% of the world's population. Our benchmark bridges these languages with 12 programming languages, creating a robust evaluation framework. This work is particularly significant given India's representation of one-eighth of the global population and the crucial role Indic languages play in Indian society. IndicEval-XL represents a significant step toward expanding the linguistic diversity in code generation systems and evaluation frameworks. By developing resources that support multiple languages, we aim to make AI-powered development tools more inclusive and accessible to developers of various linguistic backgrounds. To facilitate further research and development in this direction, we make our dataset and evaluation benchmark publicly available at https://github.com/telekom/IndicEval-XL

Modern applications are built as large, distributed systems spanning numerous modules, teams, and data centers. Despite robust engineering and recovery strategies, failures and performance issues remain inevitable, risking significant disruptions and affecting end users. Rapid and accurate root cause identification is therefore vital to ensure system reliability and maintain key service metrics. We have developed a novel causality-based Root Cause Identification (RCI) algorithm that emphasizes causation over correlation. This algorithm has been integrated into IBM Instana-bridging research to practice at scale-and is now in production use by enterprise customers. By leveraging "causal AI," Instana stands apart from typical Application Performance Management (APM) tools, pinpointing issues in near real-time. This paper highlights Instana's advanced failure diagnosis capabilities, discussing both the theoretical underpinnings and practical implementations of the RCI algorithm. Real-world examples illustrate how our causality-based approach enhances reliability and performance in today's complex system landscapes.

On-farm sensor data have allowed farmers to implement field management techniques and intensively track the corresponding responses. These data combined with historical records open the door for real-time field management improvements with the help of current advancements in computing power. However, despite these advances, the statistical design of experiments is rarely used to evaluate the performance of field management techniques accurately. Traditionally, randomized block design is prevalent in statistical designs of field trials, but in practice it is limited in dealing with large variations in soil classes, management practices, and crop varieties. More specifically, although this experimental design is suited for most trial types, it is not the optimal choice when multiple factors are tested over multifarious natural variations in farms, due to the economic constraints caused by the sheer number of variables involved. Experimental refinement is required to better estimate the effects of the primary factor in the presence of auxiliary factors. In this way, farmers can better understand the characteristics and limitations of the primary factor. This work presents a framework for automating the analysis of local field variations by fusing soil classification data and lidar topography data with historical yield. This framework will be leveraged to automate the designing of field experiments based on multiple topographic features, Comment: This was presented at the ASABE 2023 AIM meeting with id: 2301158

We present RMA-Mamba, a novel architecture that advances the capabilities of vision state space models through a specialized reverse mamba attention module (RMA). The key innovation lies in RMA-Mamba's ability to capture long-range dependencies while maintaining precise local feature representation through its hierarchical processing pipeline. By integrating Vision Mamba (VMamba)'s efficient sequence modeling with RMA's targeted feature refinement, our architecture achieves superior feature learning across multiple scales. This dual-mechanism approach enables robust handling of complex morphological patterns while maintaining computational efficiency. We demonstrate RMA-Mamba's effectiveness in the challenging domain of pathological liver segmentation (from both CT and MRI), where traditional segmentation approaches often fail due to tissue variations. When evaluated on a newly introduced cirrhotic liver dataset (CirrMRI600+) of T2-weighted MRI scans, RMA-Mamba achieves the state-of-the-art performance with a Dice coefficient of 92.08%, mean IoU of 87.36%, and recall of 92.96%. The architecture's generalizability is further validated on the cancerous liver segmentation from CT scans (LiTS: Liver Tumor Segmentation dataset), yielding a Dice score of 92.9% and mIoU of 88.99%. The source code of the proposed RMA-Mamba is available at https://github.com/JunZengz/RMAMamba., Comment: 16 pages, 3 figures

We present an end-to-end deep learning framework for automated liver cirrhosis stage estimation from multi-sequence MRI. Cirrhosis is the severe scarring (fibrosis) of the liver and a common endpoint of various chronic liver diseases. Early diagnosis is vital to prevent complications such as decompensation and cancer, which significantly decreases life expectancy. However, diagnosing cirrhosis in its early stages is challenging, and patients often present with life-threatening complications. Our approach integrates multi-scale feature learning with sequence-specific attention mechanisms to capture subtle tissue variations across cirrhosis progression stages. Using CirrMRI600+, a large-scale publicly available dataset of 628 high-resolution MRI scans from 339 patients, we demonstrate state-of-the-art performance in three-stage cirrhosis classification. Our best model achieves 72.8% accuracy on T1W and 63.8% on T2W sequences, significantly outperforming traditional radiomics-based approaches. Through extensive ablation studies, we show that our architecture effectively learns stage-specific imaging biomarkers. We establish new benchmarks for automated cirrhosis staging and provide insights for developing clinically applicable deep learning systems. The source code will be available at https://github.com/JunZengz/CirrhosisStage., Comment: 8 pages, 1 figure

Coherence refers to correlations between field vibrations at two separate points in degrees of freedom such as space, time, and polarisation. In the context of space, coherence theory has been formulated between two transverse positions which can be described either in the cartesian coordinates or in the cylindrical coordinates. When expressed in cylindrical coordinates, spatial coherence is described in terms of azimuthal and radial coordinates. The description of spatial coherence in radial degree of freedom has been formulated only recently in JOSA A 40, 411 (2023). In the present article, we demonstrate an efficient experimental technique for measuring radial coherence, and we report measurement of radial coherence of two different types of radially partially coherent optical fields., Comment: 9 pages, 3 figures

The high-dimensional basis of orbital angular momentum (OAM) has several added and unique advantages for photonics quantum technologies compared to the polarization basis, which is only two-dimensional. However, one of the major roadblocks in implementing OAM-based applications with their full potentials is the absence of an ideal OAM-mode detector. Despite the plethora of efforts in the last three decades, currently, there is no OAM detector that can detect a broad OAM-mode spectrum, has uniform detection-efficiency over all the modes, measures the true spectrum, and works for an arbitrary quantum state without the need for any prior information. In this article, we experimentally demonstrate just such an OAM detector. We report detection of pure and mixed OAM states with fidelities more than 98% and with measurement times of only a few minutes for dimensionalities up to 100. We expect our work to substantially boost the OAM-based photonics quantum technology efforts., Comment: 40 pages (Main text 26 pages, Supplementary Material 14 pages), 13 figures

Modern neutrino-nucleus cross section predictions need to incorporate sophisticated nuclear models to achieve greater predictive precision. However, the computational complexity of these advanced models often limits their practicality for experimental analyses. To address this challenge, we introduce a new Monte Carlo method utilizing Normalizing Flows to generate surrogate cross sections that closely approximate those of the original model while significantly reducing computational overhead. As a case study, we built a Monte Carlo event generator for the neutrino-nucleus cross section model developed by the Ghent group. This model employs a Hartree-Fock procedure to establish a quantum mechanical framework in which both the bound and scattering nucleon states are solutions to the mean-field nuclear potential. The surrogate cross sections generated by our method demonstrate excellent accuracy with a relative effective sample size of more than $98.4 \%$, providing a computationally efficient alternative to traditional Monte Carlo sampling methods for differential cross sections., Comment: Submitted to Physics Review D

Large Language Models (LLMs) demonstrate remarkable capabilities in question answering (QA), but metrics for assessing their reliance on memorization versus retrieval remain underdeveloped. Moreover, while finetuned models are state-of-the-art on closed-domain tasks, general-purpose models like GPT-4o exhibit strong zero-shot performance. This raises questions about the trade-offs between memorization, generalization, and retrieval. In this work, we analyze the extent to which multimodal retrieval-augmented VLMs memorize training data compared to baseline VLMs. Using the WebQA benchmark, we contrast finetuned models with baseline VLMs on multihop retrieval and question answering, examining the impact of finetuning on data memorization. To quantify memorization in end-to-end retrieval and QA systems, we propose several proxy metrics by investigating instances where QA succeeds despite retrieval failing. Our results reveal the extent to which finetuned models rely on memorization. In contrast, retrieval-augmented VLMs have lower memorization scores, at the cost of accuracy (72% vs 52% on WebQA test set). As such, our measures pose a challenge for future work to reconcile memorization and generalization in both Open-Domain QA and joint Retrieval-QA tasks.

The robustness of large language models (LLMs) against knowledge conflicts in unimodal question answering systems has been well studied. However, the effect of conflicts in information sources on vision language models (VLMs) in multimodal settings has not yet been explored. In this work, we propose \segsub, a framework that applies targeted perturbations to image sources to study and improve the robustness of VLMs against three different types of knowledge conflicts, namely parametric, source, and counterfactual conflicts. Contrary to prior findings that showed that LLMs are sensitive to parametric conflicts arising from textual perturbations, we find VLMs are largely robust to image perturbation. On the other hand, VLMs perform poorly on counterfactual examples (<30% accuracy) and fail to reason over source conflicts (<1% accuracy). We also find a link between hallucinations and image context, with GPT-4o prone to hallucination when presented with highly contextualized counterfactual examples. While challenges persist with source conflicts, finetuning models significantly improves reasoning over counterfactual samples. Our findings highlight the need for VLM training methodologies that enhance their reasoning capabilities, particularly in addressing complex knowledge conflicts between multimodal sources.

The Ultraviolet (UV) Type Ia Supernova CubeSat (UVIa) is a CubeSat/SmallSat mission concept that stands to test critical space-borne UV technology for future missions like the Habitable Worlds Observatory (HWO) while elucidating long-standing questions about the explosion mechanisms of Type Ia supernovae (SNe Ia). UVIa will observe whether any SNe Ia emit excess UV light shortly after explosion to test progenitor/explosion models and provide follow-up over many days to characterize their UV and optical flux variations over time, assembling a comprehensive multi-band UV and optical low-redshift anchor sample for upcoming high-redshift SNe Ia surveys (e.g., Euclid, Vera Rubin Observatory, Nancy Roman Space Telescope). UVIa's mission profile requires it to perform rapid and frequent visits to newly discovered SNe Ia, simultaneously observing each SNe Ia in two UV bands (FUV: 1500-1800A and NUV: 1800-2400A) and one optical band (u-band: 3000-4200A). In this study, we describe the UVIa mission concept science motivation, mission design, and key technology development., Comment: submitted to JATIS under the call for papers "Ultraviolet Science & Instrumentation: On the Way to Habitable Worlds Observatory and Beyond"

Large Language Models (LLMs) have exhibited remarkable capabilities across diverse domains, prompting investigations into their potential as generic reasoning engines. While recent studies have explored inference-time computation to enhance model performance on complex problems, current research lacks a formal framework to characterize the complexity of reasoning tasks. This study introduces the Predicate-Enumeration-Aggregation (PEA) framework, a formal approach to describe and solve a class of important reasoning tasks termed computational reasoning problems. The PEA framework decomposes these problems into predicate and enumeration components, using LLMs to synthesize programs based on specified predicates, enumeration, and aggregation rules. These synthesized programs are then executed to obtain solutions to the computational tasks. We demonstrate the framework's efficacy on benchmark tasks including Boolean satisfiability problems, game of $24$, and planning problems. Empirical evaluation reveals that PEA substantially enhances the performance of underlying models on benchmark computational problems, yielding an average accuracy improvement of approximately $50\%$, coupled with increased efficiency.

We revisit the Jaynes-Cummings model as an autonomous thermodynamic machine, where a qubit is driven by a cavity containing initially a large coherent field. Our analysis reveals a transition between the expected behavior of ideal-work source of the cavity at short times, and a long-time dynamics where the cavity autonomously measures the qubit and exerts a result-dependent drive. This autonomous feedback then purifies the qubit irrespective of its initial state. We show that the cavity functions thermodynamically as an autonomous Maxwell demon, trading mutual information for cooling power.

Let $K$ be a totally real number field and $ \mathcal{O}_K$ be the ring of integers of $K$. This manuscript examines the asymptotic solutions of the Fermat equation of signature $(r, r, p)$, specifically $x^r+y^r=dz^p$ over $K$, where $r,p \geq5$ are rational primes and $d\in \mathcal{O}_K \setminus \{0\}$. For a certain class of fields $K$, we first prove that the equation $x^r+y^r=dz^p$ has no asymptotic solution $(a,b,c) \in \mathcal{O}_K^3$ with $2 |c$. Then, we study the asymptotic solutions $(a,b,c) \in \mathcal{O}_K^3$ to the equation $x^5+y^5=dz^p$ with $2 \nmid c$. We use the modular method to prove these results., Comment: 15 pages

Federated Learning (FL) enables collaborative model training while keeping client data private. However, exposing individual client updates makes FL vulnerable to reconstruction attacks. Secure aggregation mitigates such privacy risks but prevents the server from verifying the validity of each client update, creating a privacy-robustness tradeoff. Recent efforts attempt to address this tradeoff by enforcing checks on client updates using zero-knowledge proofs, but they support limited predicates and often depend on public validation data. We propose SLVR, a general framework that securely leverages clients' private data through secure multi-party computation. By utilizing clients' data, SLVR not only eliminates the need for public validation data, but also enables a wider range of checks for robustness, including cross-client accuracy validation. It also adapts naturally to distribution shifts in client data as it can securely refresh its validation data up-to-date. Our empirical evaluations show that SLVR improves robustness against model poisoning attacks, particularly outperforming existing methods by up to 50% under adaptive attacks. Additionally, SLVR demonstrates effective adaptability and stable convergence under various distribution shift scenarios., Comment: 29 pages

Understanding the mechanism of structural phase transitions in rare-earth elements is a fundamental challenge in condensed matter physics, with significant implications for materials science applications. In this study, we present a systematic investigation on the phase transitions of yttrium under low-pressure conditions ($<$30 GPa) focusing on the hcp, Sm-type, and dhcp phases. A comparative analysis of the generalized gradient approximation (GGA) and meta-GGA functionals reveals that the PBE-GGA functional significantly underestimates the phase transition pressures, whereas the r$^2$SCAN functional provides accurate predictions of phase transition pressures which are in excellent agreement with experimental data. The results confirm that the phase transitions in yttrium are driven by vibrational instabilities, as evidenced by the emergence of soft acoustic modes in the phonon dispersion curves for the hcp and Sm-type phase. Elastic properties calculations further confirm mechanical softening at the phase boundaries, particularly in the hcp phase, suggesting a strong correlation between elastic instability and structural transitions. These findings suggest that the emergence of soft modes in the phonon dispersion curves might be a key factor driving the structural phase transition in the rare earth materials.

The polar magnetic field plays a crucial role in the solar dynamo model and contributes to predicting future solar cycles. However, continuous and direct measurements of this polar field have been available only since 1976, with data provided by the Wilcox Solar Observatory (WSO). Recent findings suggest that the Ca ii K Polar Network Index (PNI) can serve as a promising proxy for estimating the polar field of the Sun. In this study, we aim to reconstruct the polar field for the pre-1976 period by leveraging Ca ii K data from the Kodaikanal Solar Observatory (KoSO; 1904-2007) and modern Ca ii K observations from the Rome Precision Solar Photometric Telescope (Rome-PSPT; 2000-2022). We employ an automatic adaptive threshold technique to detect polar networks and calculate PNI values. Then, we calibrate these PNI values with the WSO polar field to reconstruct the polar field over 119 years., Comment: IAUG South Africa Proceeding

Adversarial Machine Learning (AML) addresses vulnerabilities in AI systems where adversaries manipulate inputs or training data to degrade performance. This article provides a comprehensive analysis of evasion and poisoning attacks, formalizes defense mechanisms with mathematical rigor, and discusses the challenges of implementing robust solutions in adaptive threat models. Additionally, it highlights open challenges in certified robustness, scalability, and real-world deployment.

Realizing the vision of using AI agents to automate critical IT tasks depends on the ability to measure and understand effectiveness of proposed solutions. We introduce ITBench, a framework that offers a systematic methodology for benchmarking AI agents to address real-world IT automation tasks. Our initial release targets three key areas: Site Reliability Engineering (SRE), Compliance and Security Operations (CISO), and Financial Operations (FinOps). The design enables AI researchers to understand the challenges and opportunities of AI agents for IT automation with push-button workflows and interpretable metrics. ITBench includes an initial set of 94 real-world scenarios, which can be easily extended by community contributions. Our results show that agents powered by state-of-the-art models resolve only 13.8% of SRE scenarios, 25.2% of CISO scenarios, and 0% of FinOps scenarios. We expect ITBench to be a key enabler of AI-driven IT automation that is correct, safe, and fast.

Pathologic diagnosis is a critical phase in deciding the optimal treatment procedure for dealing with colorectal cancer (CRC). Colonic polyps, precursors to CRC, can pathologically be classified into two major types: adenomatous and hyperplastic. For precise classification and early diagnosis of such polyps, the medical procedure of colonoscopy has been widely adopted paired with various imaging techniques, including narrow band imaging and white light imaging. However, the existing classification techniques mainly rely on a single imaging modality and show limited performance due to data scarcity. Recently, generative artificial intelligence has been gaining prominence in overcoming such issues. Additionally, various generation-controlling mechanisms using text prompts and images have been introduced to obtain visually appealing and desired outcomes. However, such mechanisms require class labels to make the model respond efficiently to the provided control input. In the colonoscopy domain, such controlling mechanisms are rarely explored; specifically, the text prompt is a completely uninvestigated area. Moreover, the unavailability of expensive class-wise labels for diverse sets of images limits such explorations. Therefore, we develop a novel model, PathoPolyp-Diff, that generates text-controlled synthetic images with diverse characteristics in terms of pathology, imaging modalities, and quality. We introduce cross-class label learning to make the model learn features from other classes, reducing the burdensome task of data annotation. The experimental results report an improvement of up to 7.91% in balanced accuracy using a publicly available dataset. Moreover, cross-class label learning achieves a statistically significant improvement of up to 18.33% in balanced accuracy during video-level analysis. The code is available at https://github.com/Vanshali/PathoPolyp-Diff.

This work investigates the theoretical boundaries of creating publicly-detectable schemes to enable the provenance of watermarked imagery. Metadata-based approaches like C2PA provide unforgeability and public-detectability. ML techniques offer robust retrieval and watermarking. However, no existing scheme combines robustness, unforgeability, and public-detectability. In this work, we formally define such a scheme and establish its existence. Although theoretically possible, we find that at present, it is intractable to build certain components of our scheme without a leap in deep learning capabilities. We analyze these limitations and propose research directions that need to be addressed before we can practically realize robust and publicly-verifiable provenance.

Continuous Latent Space (CLS) and Discrete Latent Space (DLS) models, like AttnUNet and VQUNet, have excelled in medical image segmentation. In contrast, Synergistic Continuous and Discrete Latent Space (CDLS) models show promise in handling fine and coarse-grained information. However, they struggle with modeling long-range dependencies. CLS or CDLS-based models, such as TransUNet or SynergyNet are adept at capturing long-range dependencies. Since they rely heavily on feature pooling or aggregation using self-attention, they may capture dependencies among redundant regions. This hinders comprehension of anatomical structure content, poses challenges in modeling intra-class and inter-class dependencies, increases false negatives and compromises generalization. Addressing these issues, we propose L2GNet, which learns global dependencies by relating discrete codes obtained from DLS using optimal transport and aligning codes on a trainable reference. L2GNet achieves discriminative on-the-fly representation learning without an additional weight matrix in self-attention models, making it computationally efficient for medical applications. Extensive experiments on multi-organ segmentation and cardiac datasets demonstrate L2GNet's superiority over state-of-the-art methods, including the CDLS method SynergyNet, offering an novel approach to enhance deep learning models' performance in medical image analysis.

For a new parameterization of the modified effective chiral model, developed primarily to regulate the density content of the symmetry energy and its higher order terms, equations of state (EoSs) for hyperon-rich matter ($H$) and delta baryon matter ($\Delta$) were obtained. The models were used to investigate the emission of gravitational waves (GWs) through $f$-mode oscillations in the corresponding neutron stars. We obtained the stellar structure, $f$-mode frequency and tidal deformability $\Lambda$ for our models. We report that the $\Delta$ EoS is stiffer compared to the $H$ EoS. We also analyzed the velocity of sound in these media. The corresponding mass--radius relationships were obtained and compared with various observations. We studied the dependence of $f$-mode frequencies on the stellar mass, redshift and tidal deformability. We employed the well known Cowling approximation to obtain the $f$-mode frequencies for $l=2,\,3$ and $4$ modes of oscillation. We found that the $f$-mode frequencies of the $H$ and $\Delta$ EoSs were almost the same in the lower mass region, while we observed a substantial difference between them in the high-mass region. We also obtained an empirical relation for the EoSs considered. The various attributes obtained for our models showed close agreement with various observational constraints from pulsars and GW events.

Generic non-equilibrium many-body systems display a linear growth of bipartite entanglement entropy in time, followed by a volume law saturation. In stark contrast, the Page curve dynamics of black hole physics shows that the entropy peaks at the Page time $t_{\text{Page}}$ and then decreases to zero. Here, we investigate such Page-like behavior of the von Neumann entropy in a model of strongly correlated spinless fermions in a typical system-environment setup, and characterize the properties of the Page curve dynamics in the presence of interactions using numerically exact matrix product states methods. The two phases of growth, namely the linear growth and the bending down, are shown to be separated by a non-analyticity in the min-entropy before $t_{\text{Page}}$, which separates two different quantum phases, realized as the respective ground states of the corresponding entanglement (or equivalently, modular) Hamiltonian. We confirm and generalize, by introducing interactions, the findings of \href{https://journals.aps.org/prb/abstract/10.1103/PhysRevB.109.224308}{Phys. Rev. B 109, 224308 (2024)} for a free spinless fermionic chain where the corresponding entanglement Hamiltonian undergoes a quantum phase transition at the point of non-analyticity. However, in the presence of interactions, a scaling analysis gives a non-zero critical time for the non-analyticity in the thermodynamic limit only for weak to intermediate interaction strengths, while the dynamics leading to the non-analyticity becomes \textit{instantaneous} for interactions large enough. We present a physical picture explaining these findings., Comment: 11+7 pages, 10+11 figures, 1+1 tables

The Sun's polar magnetic field is pivotal in understanding solar dynamo processes and forecasting future solar cycles. However, direct measurements of the polar field is only available since the 1970s. The chromospheric Ca II K polar network index (PNI; the fractional area of the chromospheric network regions above a certain latitude) has recently emerged as a reliable proxy for polar magnetic fields. In this study, we derive PNI estimates from newly calibrated, rotation-corrected Ca II K observations from the Kodaikanal Solar Observatory (1904-2007) and modern data from the Rome Precision Solar Photometric Telescope (2000-2022). We use both of those Ca II K archives to identify polar network regions with an automatic adaptive threshold segmentation technique and calculate the PNI. The PNI obtained from both the archives shows a significant correlation with the measured polar field from WSO (Pearson correlation coefficient r > 0.93) and the derived polar field based on an Advective Flux Transport Model (r > 0.91). The PNI series also shows a significant correlation with faculae counts derived from Mount Wilson Observatory observations (r > 0.87) for both KoSO and Rome-PSPT data. Finally, we use the PNI series from both archives to reconstruct the polar magnetic field over a 119-year-long period, which includes last 11 solar cycles (Cycle 14-24). We also obtain a relationship between the amplitude of solar cycles (in 13-month smoothed sunspot number) and the strength of the reconstructed polar field at the preceding solar cycle minimum to validate the prediction of the ongoing solar cycle, Cycle 25., Comment: 14 pages including 9 Figures and 2 Tables; accepted in ApJ

We consider the family of elliptic curves $E_{a,b}:y^2=x^3+a(x-b)^2$ with $a,b \in \mathbb{Z}$. These elliptic curves have a rational $3$-isogeny, say $\varphi$. We give an upper and a lower bound on the rank of the $\varphi$-Selmer group of $E_{a,b}$ over $K:=\mathbb{Q}(\zeta_3)$ in terms of the $3$-part of the ideal class group of certain quadratic extension of $K$. Using our bounds on the Selmer groups, we construct infinitely many curves in this family with arbitrary large $3$-Selmer rank over $K$ and no non-trivial $K$-rational point of order $3$. We also show that for a positive proportion of natural numbers $n$, the curve $E_{n,n}/\mathbb{Q}$ has root number $-1$ and $3$-Selmer rank $=1$.

The rapid growth of encryption has significantly enhanced privacy and security while posing challenges for network traffic classification. Recent approaches address these challenges by transforming network traffic into text or image formats to leverage deep-learning models originally designed for natural language processing, and computer vision. However, these transformations often contradict network protocol specifications, introduce noisy features, and result in resource-intensive processes. To overcome these limitations, we propose NetMatrix, a minimalistic tabular representation of network traffic that eliminates noisy attributes and focuses on meaningful features leveraging RFCs (Request for Comments) definitions. By combining NetMatrix with a vanilla XGBoost classifier, we implement a lightweight approach, LiM ("Less is More") that achieves classification performance on par with state-of-the-art methods such as ET-BERT and YaTC. Compared to selected baselines, experimental evaluations demonstrate that LiM improves resource consumption by orders of magnitude. Overall, this study underscores the effectiveness of simplicity in traffic representation and machine learning model selection, paving the way towards resource-efficient network traffic classification., Comment: Accepted to The Web Conference (WWW) 2025 Short Paper Track

Persistent homology applied to the activity of grid cells in the Medial Entorhinal Cortex suggests that this activity lies on a toroidal manifold. By analyzing real data and a simple model, we show that neural oscillations play a key role in the appearance of this toroidal topology. To quantitatively monitor how changes in spike trains influence the topology of the data, we first define a robust measure for the degree of toroidality of a dataset. Using this measure, we find that small perturbations ($\sim$ 100 ms) of spike times have little influence on both the toroidality and the hexagonality of the ratemaps. Jittering spikes by $\sim$ 100-500 ms, however, destroys the toroidal topology, while still having little impact on grid scores. These critical jittering time scales fall in the range of the periods of oscillations between the theta and eta bands. We thus hypothesized that these oscillatory modulations of neuronal spiking play a key role in the appearance and robustness of toroidal topology and the hexagonal spatial selectivity is not sufficient. We confirmed this hypothesis using a simple model for the activity of grid cells, consisting of an ensemble of independent rate-modulated Poisson processes. When these rates were modulated by oscillations, the network behaved similarly to the real data in exhibiting toroidal topology, even when the position of the fields were perturbed. In the absence of oscillations, this similarity was substantially lower. Furthermore, we find that the experimentally recorded spike trains indeed exhibit temporal modulations at the eta and theta bands, and that the ratio of the power in the eta band to that of the theta band, $A_{\eta}/A_{\theta}$, correlates with the critical jittering time at which the toroidal topology disappears., Comment: 41 pages, 27 figures, published at doi.org/10.1371/journal.pcbi.1012776

The first few hours of a supernova contain significant information about the progenitor system. The most modern wide-field surveys that scan the sky repeatedly every few days can discover all kinds of transients in those early epochs. At such times, some progenitor footprints may be visible, elucidating critical explosion parameters and helping to distinguish between leading explosion models. A dedicated spectroscopic classification programme using the optical spectrograph OSIRIS mounted to the Gran Telescopio de Canarias was set up to try to obtain observations of supernova at those early epochs. With the time awarded, we obtained spectra for 10 SN candidates, which we present here. Half of them were thermonuclear SNe, while the other half were core-collapse SNe. Most (70\%) were observed within the first six days of the estimated explosion, with two being captured within the first 48 hours. We present a characterization of the spectra, together with other public ancillary photometry from ZTF and ATLAS. This programme shows the need for an accompanying rapid-response spectroscopic programme to existing and future deep photometric wide-field surveys located at the right longitude to be able to trigger observations in a few hours after the discovery of the supernova candidate. Both the future La Silla Southern Supernova Survey (LS4) and the Legacy Survey of Space and Time (LSST) both located in Chile will be providing discovery and follow up of most of the transients in the southern hemisphere. This paper demonstrates that with a rapid spectroscopic programme and stringent triggering criteria, obtaining a sample of SN with spectra within a day of the explosion is possible., Comment: 20 pages, 16 figures. Submitted to A&A

This paper investigates the existence and uniqueness of mild solutions, as well as the approximate controllability, of a class of fractional evolution equations with nonlocal conditions in Hilbert spaces. Sufficient conditions for approximate controllability are established through a novel approach to the approximate solvability of semilinear operator equations. The methodology utilizes Green's function and constructs a control function based on the Gramian controllability operator. The analysis is based on Schauder's fixed point theorem and the theory of fractional order solution operators and resolvent operators. To demonstrate the feasibility of the proposed theoretical results, an illustrative example is provided., Comment: 22 pages

This paper introduces a scalable Anomaly Detection Service with a generalizable API tailored for industrial time-series data, designed to assist Site Reliability Engineers (SREs) in managing cloud infrastructure. The service enables efficient anomaly detection in complex data streams, supporting proactive identification and resolution of issues. Furthermore, it presents an innovative approach to anomaly modeling in cloud infrastructure by utilizing Large Language Models (LLMs) to understand key components, their failure modes, and behaviors. A suite of algorithms for detecting anomalies is offered in univariate and multivariate time series data, including regression-based, mixture-model-based, and semi-supervised approaches. We provide insights into the usage patterns of the service, with over 500 users and 200,000 API calls in a year. The service has been successfully applied in various industrial settings, including IoT-based AI applications. We have also evaluated our system on public anomaly benchmarks to show its effectiveness. By leveraging it, SREs can proactively identify potential issues before they escalate, reducing downtime and improving response times to incidents, ultimately enhancing the overall customer experience. We plan to extend the system to include time series foundation models, enabling zero-shot anomaly detection capabilities., Comment: Accepted at the AAAI-2025 Deployable AI Workshop

This study presents the Adaptive Minimum-Variance Portfolio (AMVP) framework and the Adaptive Minimum-Risk Rate (AMRR) metric, innovative tools designed to optimize portfolios dynamically in volatile and nonstationary financial markets. Unlike traditional minimum-variance approaches, the AMVP framework incorporates real-time adaptability through advanced econometric models, including ARFIMA-FIGARCH processes and non-Gaussian innovations. Empirical applications on cryptocurrency and equity markets demonstrate the proposed framework's superior performance in risk reduction and portfolio stability, particularly during periods of structural market breaks and heightened volatility. The findings highlight the practical implications of using the AMVP and AMRR methodologies to address modern investment challenges, offering actionable insights for portfolio managers navigating uncertain and rapidly changing market conditions.

White matter bundle segmentation is crucial for studying brain structural connectivity, neurosurgical planning, and neurological disorders. White Matter Segmentation remains challenging due to structural similarity in streamlines, subject variability, symmetry in 2 hemispheres, etc. To address these challenges, we propose TractoGPT, a GPT-based architecture trained on streamline, cluster, and fusion data representations separately. TractoGPT is a fully-automatic method that generalizes across datasets and retains shape information of the white matter bundles. Experiments also show that TractoGPT outperforms state-of-the-art methods on average DICE, Overlap and Overreach scores. We use TractoInferno and 105HCP datasets and validate generalization across dataset., Comment: Accepted as a conference paper at 23rd IEEE International Symposium on Biomedical Imaging 2025. IEEE holds the copyright for this publication

This study examines the investment landscape of Pakistan as an emerging and frontier market, focusing on implications for international investors, particularly those in the United States, through exchange-traded funds (ETFs) with exposure to Pakistan. The analysis encompasses 30 ETFs with varying degrees of exposure to Pakistan, covering the period from January 1, 2016, to February 2024. This research highlights the potential benefits and risks associated with investing in these ETFs, emphasizing the importance of thorough risk assessments and portfolio performance comparisons. By providing descriptive statistics and performance metrics based on historical optimization, this paper aims to equip investors with the necessary insights to make informed decisions when optimizing their portfolios with Pakistan-exposed ETFs. The second part of the paper introduces and assesses dynamic optimization methodologies. This section is designed to explore the adaptability and performance metrics of dynamic optimization techniques in comparison with conventional historical optimization methods. By integrating dynamic optimization into the investigation, this research aims to offer insights into the efficacy of these contrasting methodologies in the context of Pakistan-exposed ETFs. The findings underscore the significance of Pakistan's market dynamics within the broader context of emerging markets, offering a pathway for diversification and potential growth in investment strategies.

The dynamics of infection spread in populations has received popular attention since the outbreak of Covid-19 and many statistical models have been developed. One of the interesting areas of research is short-time dynamics in confined, indoor environments. We have modeled this using a simple Monte Carlo scheme. Our model is generally applicable for the peer-to-peer transmission case, when the infection spread occurs only between an infected subject and a healthy subject with a certain probability, i.e., airborne and surface transmission is neglected. The probability of infection spread is incorporated using a simple exponential decay with distance between the subjects. Simulations are performed for the cases of (1) constant subject population and (2) variable subject population due to inflow/outflow. We specifically focus on the large fluctuations in the dynamics due to finite number of subjects. Results of our study may be useful to determine social-distancing guidelines in indoor contexts.

Large language model (LLM) serving is becoming an increasingly important workload for cloud providers. Based on performance SLO requirements, LLM inference requests can be divided into (a) interactive requests that have tight SLOs in the order of seconds, and (b) batch requests that have relaxed SLO in the order of minutes to hours. These SLOs can degrade based on the arrival rates, multiplexing, and configuration parameters, thus necessitating the use of resource autoscaling on serving instances and their batch sizes. However, previous autoscalers for LLM serving do not consider request SLOs leading to unnecessary scaling and resource under-utilization. To address these limitations, we introduce Chiron, an autoscaler that uses the idea of hierarchical backpressure estimated using queue size, utilization, and SLOs. Our experiments show that Chiron achieves up to 90% higher SLO attainment and improves GPU efficiency by up to 70% compared to existing solutions.

We present a novel approach to understand vehicular traffic jams by studying a simple model, Elementary Cellular Automaton Rule 184 (ECA 184). Using key traffic observables, such as the total delay and relaxation time, as well as microscopic measures like delays and lifetimes of individual jams, we show how these quantities can fully characterize the system's behavior, revealing features analogous to those of a continuous phase transition. We exploit specific properties of ECA 184 to develop an efficient algorithm for calculating these observables numerically and introduce an auxiliary quantity, termed ''elementary jams'', which allows us to determine these observables. We discuss the implications of our results, highlighting connections to a potential field-theoretic description of traffic and suggest future application of these methods to more complex models. Supporting code can be found on: https://github.com/Aryaman-Jha/ECA_184_Critical, Comment: 9 pages, 19 figures

Imitation learning from human demonstrations enables robots to perform complex manipulation tasks and has recently witnessed huge success. However, these techniques often struggle to adapt behavior to new preferences or changes in the environment. To address these limitations, we propose Fine-tuning Diffusion Policy with Human Preference (FDPP). FDPP learns a reward function through preference-based learning. This reward is then used to fine-tune the pre-trained policy with reinforcement learning (RL), resulting in alignment of pre-trained policy with new human preferences while still solving the original task. Our experiments across various robotic tasks and preferences demonstrate that FDPP effectively customizes policy behavior without compromising performance. Additionally, we show that incorporating Kullback-Leibler (KL) regularization during fine-tuning prevents over-fitting and helps maintain the competencies of the initial policy.

Thanks to work of Rouse, Sutherland, and Zureick-Brown, it is known exactly which subgroups of GL$_2(\mathbf{Z}_3)$ can occur as the image of the $3$-adic Galois representation attached to a non-CM elliptic curve over $\mathbf{Q}$, with a single exception: the normaliser of the non-split Cartan subgroup of level 27. In this paper, we complete the classification of 3-adic Galois images by showing that the normaliser of the non-split Cartan subgroup of level 27 cannot occur as a 3-adic Galois image of a non-CM elliptic curve. Our proof proceeds via computing the $\mathbf{Q}(\zeta_3)$-rational points on a certain smooth plane quartic curve $X'_H$ (arising as a quotient of the modular curve $X_{ns}^+(27)$) defined over $\mathbf{Q}(\zeta_3)$ whose Jacobian has Mordell--Weil rank 6. To this end, we describe how to carry out the quadratic Chabauty method for a modular curve $X$ defined over a number field $F$, which, when applicable, determines a finite subset of $X(F\otimes\mathbf{Q}_p)$ in certain situations of larger Mordell--Weil rank than previously considered. Together with an analysis of local heights above 3, we apply this quadratic Chabauty method to determine $X'_H(\mathbf{Q}(\zeta_3))$. This allows us to compute the set $X_{ns}^+(27)(\mathbf{Q})$, finishing the classification of 3-adic images of Galois.

The global rotational profile of the solar atmosphere and its variation at different layers, although crucial for a comprehensive understanding of the dynamics of the solar magnetic field, has been a subject to contradictory results throughout the past century. In this study, we thereby unify the results for different parts of the multi-thermal Solar atmosphere by utilizing 13 years of data in 7 wavelength channels of the Atmospheric Imaging Assembly (AIA) atop the Solar Dynamic Observatory (SDO). Using the method of image correlation, we find that the solar atmosphere exhibits a rotational profile that is up to 4.18% and 1.92% faster at the equator and comparatively less differential than that of the photosphere, as derived from Doppler measurements and sunspots, respectively and exhibits variation at different respective heights. Additionally, we find results suggestive of the role played by the rooting of different magnetic field structures on a comparison with helioseismology data., Comment: 4 pages, 2 figures, Proceedings for XXXIInd IAU General Assembly Focused Meeting (FM)-8

As autonomous vehicle (AV) technology advances towards maturity, it becomes imperative to examine the security vulnerabilities within these cyber-physical systems. While conventional cyber-security concerns are often at the forefront of discussions, it is essential to get deeper into the various layers of vulnerability that are often overlooked within mainstream frameworks. Our goal is to spotlight imminent challenges faced by AV operators and explore emerging technologies for comprehensive solutions. This research outlines the diverse security layers, spanning physical, cyber, coding, and communication aspects, in the context of AVs. Furthermore, we provide insights into potential solutions for each potential attack vector, ensuring that autonomous vehicles remain secure and resilient in an evolving threat landscape.

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

X-ray emissions due to the two-electron one-photon (TEOP) process in the neon projectile and aluminum target have been successfully observed for the beam energy window of 1.8-2.1 MeV. Experimental TEOP transition energies have been compared with theoretical predictions of flexible atomic structure code (FAC) and General-purpose Relativistic Atomic Structure (GRASP) package. Present results have been verified with reported theoretical and experimental values. Transition rates of the TEOP transitions have also been studied using the said codes. The observed lines have been assigned when the measured transition energies are in good agreement with the theoretical values. Such assignments have further been validated with the good agreements between the experimental and theoretical transition rates. Note that only the TEOP lines in projectile ions are seen with 1.8 MeV energy. In contrast, the TEOP lines in target ions are also observed well with 2.1 MeV energy. Thus, this study sheds useful light on the excitation mechanism of the TEOP processes in the low energy regimes., Comment: 6 pages, 2 figures, 3 tables. arXiv admin note: text overlap with arXiv:2201.02566

The superheated emulsion detector consisting of the droplets of tetra-fluoroethane (C2HC$_2$H$_2$F$_4$2F4) has been fabricated at the laboratory and installed at the 555m deep underground laboratory, JUSL during July to Dec 2022. The 500ml detector ran for an effective period of 48.6 days at a threshold of 5.87 keV with an exposure of 2.47 kg-days. The acoustic signals produced due to the bubble nucleation were collected by the acoustic sensor and FPGAbased data acquisition system. The data shows a minimum sensitivity of SI-nucleon for carbon at WIMP mass of 22.81 GeV/c$^2$ and SD (p) for fluorine at 30.67 GeV/c$^2$. The threshold of WIMP mass is 5.16 GeV/c$^2$ for F and 4.44 GeV/c$^2$ for C at the operating threshold of 5.87 keV. The first result of the dark matter direct search experiment named InDEx with tetra-fluoro-ethane active liquid from JUSL underground laboratory is reported in this article.

The pervasiveness of proprietary language models has raised critical privacy concerns, necessitating advancements in private inference (PI), where computations are performed directly on encrypted data without revealing users' sensitive information. While PI offers a promising solution, its practical deployment is hindered by substantial communication and latency overheads, primarily stemming from nonlinear operations. To address this, we introduce an information-theoretic framework to characterize the role of nonlinearities in decoder-only language models, laying a principled foundation for optimizing transformer-architectures tailored to the demands of PI. By leveraging Shannon's entropy as a quantitative measure, we uncover the previously unexplored dual significance of nonlinearities: beyond ensuring training stability, they are crucial for maintaining attention head diversity. Specifically, we find that their removal triggers two critical failure modes: {\em entropy collapse} in deeper layers that destabilizes training, and {\em entropic overload} in earlier layers that leads to under-utilization of Multi-Head Attention's (MHA) representational capacity. We propose an entropy-guided attention mechanism paired with a novel entropy regularization technique to mitigate entropic overload. Additionally, we explore PI-friendly alternatives to layer normalization for preventing entropy collapse and stabilizing the training of LLMs with reduced-nonlinearities. Our study bridges the gap between information theory and architectural design, establishing entropy dynamics as a principled guide for developing efficient PI architectures. The code and implementation are available at https://github.com/Nandan91/entropy-guided-attention-llm, Comment: Accepted to the 6th AAAI Workshop on Privacy-Preserving Artificial Intelligence (PPAI), 2025. arXiv admin note: substantial text overlap with arXiv:2410.13060