Search

We study a generalized Kardar-Parisi-Zhang (KPZ) equation [D. Jana et al, Phys. Rev. E 109, L032104 (2024)], that sets the paradigm for universality in roughening of growing nonequilibrium surfaces without any conservation laws, but with competing local and nonlocal nonlinear effects. We show that such a generalized KPZ equation in two dimensions can describe a strong coupling rough or a crumpled surface, in addition to a weak coupling phase. The conformation fluctuations of such a rough surface are given by nonuniversal exponents, with orientational long-ranged order and positional short-ranged order, whereas the crumpled phase has positional and orientational short range order. Experimental and theoretical implications of these results are discussed.

We prove that for every odd $q\geq 3$, any $q$-query binary, possibly non-linear locally decodable code ($q$-LDC) $E:\{\pm1\}^k \rightarrow \{\pm1\}^n$ must satisfy $k \leq \tilde{O}(n^{1-2/q})$. For even $q$, this bound was established in a sequence of prior works. For $q=3$, the above bound was achieved in a recent work of Alrabiah, Guruswami, Kothari and Manohar using an argument that crucially exploits known exponential lower bounds for $2$-LDCs. Their strategy hits an inherent bottleneck for $q \geq 5$. Our key insight is identifying a general sufficient condition on the hypergraph of local decoding sets called $t$-approximate strong regularity. This condition demands that 1) the number of hyperedges containing any given subset of vertices of size $t$ (i.e., its co-degree) be equal to the same but arbitrary value $d_t$ up to a multiplicative constant slack, and 2) all other co-degrees be upper-bounded relative to $d_t$. This condition significantly generalizes related proposals in prior works that demand absolute upper bounds on all co-degrees. We give an argument based on spectral bounds on Kikuchi Matrices that lower bounds the blocklength of any LDC whose local decoding sets satisfy $t$-approximate strong regularity for any $t \leq q$. Crucially, unlike prior works, our argument works despite having no non-trivial absolute upper bound on the co-degrees of any set of vertices. To apply our argument to arbitrary $q$-LDCs, we give a new, greedy, approximate strong regularity decomposition that shows that arbitrary, dense enough hypergraphs can be partitioned (up to a small error) into approximately strongly regular pieces satisfying the required relative bounds on the co-degrees.

The paper considers the controller synthesis problem for general MIMO systems with unknown dynamics, aiming to fulfill the temporal reach-avoid-stay task, where the unsafe regions are time-dependent, and the target must be reached within a specified time frame. The primary aim of the paper is to construct the spatiotemporal tube (STT) using a sampling-based approach and thereby devise a closed-form approximation-free control strategy to ensure that system trajectory reaches the target set while avoiding time-dependent unsafe sets. The proposed scheme utilizes a novel method involving STTs to provide controllers that guarantee both system safety and reachability. In our sampling-based framework, we translate the requirements of STTs into a Robust optimization program (ROP). To address the infeasibility of ROP caused by infinite constraints, we utilize the sampling-based Scenario optimization program (SOP). Subsequently, we solve the SOP to generate the tube and closed-form controller for an unknown system, ensuring the temporal reach-avoid-stay specification. Finally, the effectiveness of the proposed approach is demonstrated through three case studies: an omnidirectional robot, a SCARA manipulator, and a magnetic levitation system.

Large Language Models (LLMs) have revolutionized natural language processing (NLP), excelling in tasks like text generation and summarization. However, their increasing adoption in mission-critical applications raises concerns about hardware-based threats, particularly bit-flip attacks (BFAs). BFAs, enabled by fault injection methods such as Rowhammer, target model parameters in memory, compromising both integrity and performance. Identifying critical parameters for BFAs in the vast parameter space of LLMs poses significant challenges. While prior research suggests transformer-based architectures are inherently more robust to BFAs compared to traditional deep neural networks, we challenge this assumption. For the first time, we demonstrate that as few as three bit-flips can cause catastrophic performance degradation in an LLM with billions of parameters. Current BFA techniques are inadequate for exploiting this vulnerability due to the difficulty of efficiently identifying critical parameters within the immense parameter space. To address this, we propose AttentionBreaker, a novel framework tailored for LLMs that enables efficient traversal of the parameter space to identify critical parameters. Additionally, we introduce GenBFA, an evolutionary optimization strategy designed to refine the search further, isolating the most critical bits for an efficient and effective attack. Empirical results reveal the profound vulnerability of LLMs to AttentionBreaker. For example, merely three bit-flips (4.129 x 10^-9% of total parameters) in the LLaMA3-8B-Instruct 8-bit quantized (W8) model result in a complete performance collapse: accuracy on MMLU tasks drops from 67.3% to 0%, and Wikitext perplexity skyrockets from 12.6 to 4.72 x 10^5. These findings underscore the effectiveness of AttentionBreaker in uncovering and exploiting critical vulnerabilities within LLM architectures.

We study quantitative aspects of the following fact: For every graph $F$, there exists a graph $G$ with the property that any $2$-coloring of the triangles of $G$ yields an induced copy of $F$, in which all triangles are monochromatic. We define the Ramsey number $R_{\text{ind}}^{\Delta}(F)$ as the smallest size of such a graph $G$. Although this fact has several proofs, all of them provide tower-type bounds. We study the number $R_{\text{ind}}^{\Delta}(F)$ for some particular classes of graphs $F$., Comment: 23 pages, comments are welcome

We explore the influence of interactions between dark matter (DM) and dark energy (DE) on the cosmic evolution of the Universe within a viscous dark energy (VDE) framework. Moving beyond traditional interacting dark energy (IDE) models, we propose a generalized IDE model adaptable to diverse IDE scenarios via IDE coupling parameters. In order to investigate deviations from $\Lambda$CDM across cosmic epochs by highlighting how viscous and the interactions between DM and DE impact cosmic density and expansion rates, we consider a model agnostic form of VDE. Eventually we perform a Bayesian analysis using the Union 2.1 Supernova Ia dataset and Markov Chain Monte Carlo (MCMC) sampling to obtain optimal values of model parameters. This comprehensive analysis provides insights about the interplay between viscous and IDE in shaping the Universe's expansion history., Comment: 13 pages, 7 figures

Transformer model has gained prominence as a popular deep neural network architecture for neural language processing (NLP) and computer vision (CV) applications. However, the extensive use of nonlinear operations, like softmax, poses a performance bottleneck during transformer inference and comprises up to 40% of the total latency. Hence, we propose innovations at the circuit, architecture, and algorithm levels to accelerate the transformer. At the circuit level, we propose topkima-combining top-k activation selection with in-memory ADC (IMA) to implement a low-energy and low-latency softmax without any sorting latency. Only the k largest activations are sent to the softmax calculation block, reducing the huge computational cost of softmax. Using a modified training scheme with top-k only in the forward pass, experimental results demonstrate only a 0.4% to 1.2% reduction in accuracy across ViT, distilBERT, and BERT-base models when evaluated on CIFAR-10, CIFAR-100, and SQuAD datasets with k=5. At the architecture level, an improved scale-free technique is introduced to reduce the computational cost of attention. The combined system, dubbed Topkima-Former, enhances 1.8x-84x speedup and 1.3x-35x energy efficiency (EE) over prior In-memory computing (IMC) accelerators. Compared to a conventional softmax macro and a digital top-k (Dtopk) softmax macro, our proposed tokima softmax macro achieves about 15x and 8x faster speed respectively., Comment: 7 pages

We investigate the reflected entropy for bipartite mixed state configurations in a $T\bar{T}$ deformed boundary conformal field theory in $2$ dimensions (BCFT$_2$). The bulk dual is described by asymptotically AdS$_3$ geometries with the cut off surface pushed deeper into the bulk and truncated by an end of the world brane. We obtain the reflected entropy up to a linear order in the radial cut-off for static and time dependent configurations involving an eternal black hole, from the island and defect extremal surface (DES) prescriptions in the context of the deformed AdS/BCFT. We observe agreement of the leading order correction for all cases between the two prescriptions. We also obtain the analogous of the Page curves for the reflected entropy and investigate the modification due to the $T\bar{T}$ deformation., Comment: 46 pages, 28 figures

Causal effect estimation is a critical task in statistical learning that aims to find the causal effect on subjects by identifying causal links between a number of predictor (or, explanatory) variables and the outcome of a treatment. In a regressional framework, we assign a treatment and outcome model to estimate the average causal effect. Additionally, for high dimensional regression problems, variable selection methods are also used to find a subset of predictor variables that maximises the predictive performance of the underlying model for better estimation of the causal effect. In this paper, we propose a different approach. We focus on the variable selection aspects of high dimensional causal estimation problem. We suggest a cautious Bayesian group LASSO (least absolute shrinkage and selection operator) framework for variable selection using prior sensitivity analysis. We argue that in some cases, abstaining from selecting (or, rejecting) a predictor is beneficial and we should gather more information to obtain a more decisive result. We also show that for problems with very limited information, expert elicited variable selection can give us a more stable causal effect estimation as it avoids overfitting. Lastly, we carry a comparative study with synthetic dataset and show the applicability of our method in real-life situations.

Sequential learning in a multi-agent resource constrained matching market has received significant interest in the past few years. We study decentralized learning in two-sided matching markets where the demand side (aka players or agents) competes for a `large' supply side (aka arms) with potentially time-varying preferences, to obtain a stable match. Despite a long line of work in the recent past, existing learning algorithms such as Explore-Then-Commit or Upper-Confidence-Bound remain inefficient for this problem. In particular, the per-agent regret achieved by these algorithms scales linearly with the number of arms, $K$. Motivated by the linear contextual bandit framework, we assume that for each agent an arm-mean can be represented by a linear function of a known feature vector and an unknown (agent-specific) parameter. Moreover, our setup captures the essence of a dynamic (non-stationary) matching market where the preferences over arms change over time. Our proposed algorithms achieve instance-dependent logarithmic regret, scaling independently of the number of arms, $K$.

We prove upper bounds which are independent of the dimension of the ambient space, on the number of realizable zero-nonzero patterns as well as sign conditions (when the field of coefficients is ordered) of a finite set of polynomials $\mathcal{P}$ restricted to some algebraic subset $V$.Our bounds (which are tight) depend on the number and the degrees of the polynomials in $\mathcal{P}$, as well as the degree (of the embedding) of $V$ and the dimension of $V$, but are independent of the dimension of the space in which $V$ is embedded. This last feature of our bounds is useful in situations where the ambient dimension could be much larger than $\dim V$. We give several applications of our results. We generalize existing results on bounding the speeds of algebraically defined classes of graphs, as well as lower bounds in terms of the number of connected components for testing membership in semi-algebraic sets using algebraic computation trees. Motivated by quantum complexity theory we introduce a notion of relative rank (additive as well as multiplicative) in finite dimensional vector spaces and algebras relative to a fixed algebraic subset in the vector space or algebra -- which generalizes the classical definition of ranks of tensors. We prove a very general lower bound on the maximum relative rank of finite subsets relative to algebraic subsets of bounded degree and dimension which is independent of the dimension of the vector space or algebra. We show how our lower bound implies a quantum analog of a classical lower bound result of Shannon for Boolean circuits -- that almost all Boolean functions require (classical) circuits of size at least $\Omega(2^n/n)$., Comment: 46 pages. Comments welcome. This paper supersedes the paper by the same authors titled "Quantum Analog of Shannon's Lower Bound Theorem'' (arXiv:2308.13091 [quant-ph])

The majority of atomic nuclei have deformed shapes and nearly all these shapes are symmetric with respect to reflection. There are only a few reflection asymmetric pear-shaped nuclei that have been found in actinide and lanthanide regions, which have static octupole deformation. These nuclei possess an intrinsic electric dipole moment due to the shift between the center of charge and the center of mass. This manifests in the enhancement of the electric dipole transition rates. In this article, we report on the measurement of the lifetimes of the high spin levels of the two alternate parity bands in $^{100}$Ru through the Doppler Shift Attenuation Method. The estimated electric dipole transition rates have been compared with the calculated transition rates using the triaxial projected shell model without octupole deformation, and are found to be an order of magnitude enhanced. Thus, the observation of seven inter-leaved electric dipole transitions with enhanced rates establish $^{100}$Ru as possibly the first octupole deformed nucleus reported in the A $\approx$ 100 mass region.

Balance functions have been regarded in the past as a method of investigating the late-stage hadronization found in the presence of a strongly-coupled medium. They are also used to constrain mechanisms of particle production in large and small collision systems. Measurements of charge balance functions for inclusive and identified particle pairs are reported as a function of charged particle multiplicity in proton--proton collisions simulated with the PYTHIA8 and the EPOS4 models. The charge balance functions of inclusive, pion, kaon, and proton pairs exhibit amplitudes and shapes that depend on particle species and differ significantly in the two models due to the different particle production mechanisms implemented in PYTHIA and EPOS. The shapes and amplitudes also evolve with multiplicity in both models. In addition, the evolution of the longitudinal rms width and that of balance functions integrals with multiplicity (and average transverse momentum) feature significant differences in the two models.

Millimeter-wave refracting telescopes targeting the degree-scale structure of the cosmic microwave background (CMB) have recently grown to diffraction-limited apertures of over 0.5 meters. These instruments are entirely housed in vacuum cryostats to support their sub-kelvin bolometric detectors and to minimize radiative loading from thermal emission due to absorption loss in their transmissive optical elements. The large vacuum window is the only optical element in the system at ambient temperature, and therefore minimizing loss in the window is crucial for maximizing detector sensitivity. This motivates the use of low-loss polymer materials and a window as thin as practicable. However, the window must simultaneously meet the requirement to keep sufficient vacuum, and therefore must limit gas permeation and remain mechanically robust against catastrophic failure under pressure. We report on the development of extremely thin composite polyethylene window technology that meets these goals. Two windows have been deployed for two full observing seasons on the BICEP3 and BA150 CMB telescopes at the South Pole. On BICEP3, the window has demonstrated a 6% improvement in detector sensitivity., Comment: 20 pages, 12 figures, 4 tables

The concept of Vapnik-Chervonenkis (VC) density is pivotal across various mathematical fields, including model theory, discrete geometry, and probability theory. In this paper, we introduce a topological generalization of VC-density. Let $Y$ be a topological space and $\mathcal{X},\mathcal{Z}^{(0)},\ldots,\mathcal{Z}^{(q-1)}$ be families of subspaces of $Y$. We define a two parameter family of numbers, $\mathrm{vcd}^{p,q}_{\mathcal{X},\overline{\mathcal{Z}}}$, which we refer to as the degree $p$, order $q$, VC-density of the pair \[ (\mathcal{X},\overline{\mathcal{Z}} = (\mathcal{Z}^{(0)},\ldots,\mathcal{Z}^{(q-1)}). \] The classical notion of VC-density within this topological framework can be recovered by setting $p=0, q=1$. For $p=0, q > 0$, we recover Shelah's notion of higher-order VC-density for $q$-dependent families. Our definition introduces a new notion when $p>0$. Our main result establishes that that in any model of these theories \[ \mathrm{vcd}^{p,q}_{\mathcal{X},\overline{\mathcal{Z}}} \leq (p+q) \dim X. \] This result generalizes known VC-density bounds in these structures, extending them in multiple ways, as well as providing a uniform proof paradigm applicable to all of them. We give examples to show that our bounds are optimal. We present combinatorial applications of our higher-degree VC-density bounds, deriving topological analogs of well-known results such as the existence of $\varepsilon$-nets and the fractional Helly theorem. We show that with certain restrictions, these results extend to our higher-degree topological setting., Comment: 52 pages. Comments welcome

We present here a real-time analysis of diffraction images acquired at high frame-rate (925 Hz) and its application to macromolecular serial crystallography. The software uses a new signal separation algorithm, able to distinguish the amorphous (or powder diffraction) component from the diffraction signal originating from single crystals. It relies on the ability to work efficiently in azimuthal space and derives from the work performed on pyFAI, the fast azimuthal integration library. Two applications are built upon this separation algorithm: a lossy compression algorithm and a peak-picking algorithm; the performances of both is assessed by comparing data quality after reduction with XDS and CrystFEL., Comment: 43 pages, 12 figures

Learning representations of underlying environmental dynamics from partial observations is a critical challenge in machine learning. In the context of Partially Observable Markov Decision Processes (POMDPs), state representations are often inferred from the history of past observations and actions. We demonstrate that incorporating future information is essential to accurately capture causal dynamics and enhance state representations. To address this, we introduce a Dynamical Variational Auto-Encoder (DVAE) designed to learn causal Markovian dynamics from offline trajectories in a POMDP. Our method employs an extended hindsight framework that integrates past, current, and multi-step future information within a factored-POMDP setting. Empirical results reveal that this approach uncovers the causal graph governing hidden state transitions more effectively than history-based and typical hindsight-based models.

Rotation is an important, yet poorly-modelled phenomenon of stellar structure and evolution. Accurate estimates of internal rotation rates are therefore valuable for constraining stellar evolution models. We aim to assess the accuracy of asteroseismic estimates of internal rotation rates and how these depend on the fundamental stellar parameters. We apply the recently-developed method called extended-MOLA inversions to infer localised estimates of internal rotation rates of synthetic observations of red giants. We search for suitable reference stellar models following a grid-based approach, and assess the robustness of the resulting inferences to the choice of reference model. We find that matching the mixed mode pattern between the observation and the reference model is an important criterion to select suitable reference models. We propose to i) select a set of reference models based on the correlation between the observed rotational splittings and the mode-trapping parameter ii) compute rotation rates for all these models iii) use the mean value obtained across the whole set as the estimate of the internal rotation rates. We find that the effect of a near surface perturbation in the synthetic observations on the rotation rates estimated based on the correlation between the observed rotational splittings and the mode-trapping parameter is negligible. We conclude that when using an ensemble of reference models, constructed based on matching the mixed mode pattern, the input rotation rates can be recovered across a range of fundamental stellar parameters like mass, mixing-length parameter and composition. Further, red-giant rotation rates determined in this way are also independent of a near surface perturbation of stellar structure., Comment: 23 pages, 16 figures, Accepted for publication in Astronomy and Astrophysics

Agricultural landscapes are quite complex, especially in the Global South where fields are smaller, and agricultural practices are more varied. In this paper we report on our progress in digitizing the agricultural landscape (natural and man-made) in our study region of India. We use high resolution imagery and a UNet style segmentation model to generate the first of its kind national-scale multi-class panoptic segmentation output. Through this work we have been able to identify individual fields across 151.7M hectares, and delineating key features such as water resources and vegetation. We share how this output was validated by our team and externally by downstream users, including some sample use cases that can lead to targeted data driven decision making. We believe this dataset will contribute towards digitizing agriculture by generating the foundational baselayer., Comment: 34 pages, 7 tables, 15 figs

Submodular optimization has become increasingly prominent in machine learning and fairness has drawn much attention. In this paper, we propose to study the fair $k$-submodular maximization problem and develop a $\frac{1}{3}$-approximation greedy algorithm with a running time of $\mathcal{O}(knB)$. To the best of our knowledge, our work is the first to incorporate fairness in the context of $k$-submodular maximization, and our theoretical guarantee matches the best-known $k$-submodular maximization results without fairness constraints. In addition, we have developed a faster threshold-based algorithm that achieves a $(\frac{1}{3} - \epsilon)$ approximation with $\mathcal{O}(\frac{kn}{\epsilon} \log \frac{B}{\epsilon})$ evaluations of the function $f$. Furthermore, for both algorithms, we provide approximation guarantees when the $k$-submodular function is not accessible but only can be approximately accessed. We have extensively validated our theoretical findings through empirical research and examined the practical implications of fairness. Specifically, we have addressed the question: ``What is the price of fairness?" through case studies on influence maximization with $k$ topics and sensor placement with $k$ types. The experimental results show that the fairness constraints do not significantly undermine the quality of solutions., Comment: 17 pages. To appear in IEEE BigData 2024

Reconfigurable intelligent surfaces (RIS) and index modulation (IM) represent key technologies for enabling reliable wireless communication with high energy efficiency. However, to fully take advantage of these technologies in practical deployments, comprehending the impact of the non-ideal nature of the underlying transceivers is paramount. In this context, this paper introduces two RIS-assisted IM communication models, in which the RIS is part of the transmitter and space-shift keying (SSK) is employed for IM, and assesses their performance in the presence of hardware impairments. In the first model, the RIS acts as a passive reflector only, reflecting the oncoming SSK modulated signal intelligently towards the desired receive diversity branch/antenna. The second model employs RIS as a transmitter, employing M-ary phase-shift keying for reflection phase modulation (RPM), and as a reflector for the incoming SSK modulated signal. Considering transmissions subjected to Nakagami-m fading, and a greedy detection rule at the receiver, the performance of both the system configurations is evaluated. Specifically, the pairwise probability of erroneous index detection and the probability of erroneous index detection are adopted as performance metrics, and their closed-form expressions are derived for the RIS-assisted SSK and RIS-assisted SSK-RPM system models. Monte-Carlo simulation studies are carried out to verify the analytical framework, and numerical results are presented to study the dependency of the error performance on the system parameters. The findings highlight the effect of hardware impairment on the performance of the communication system under study., Comment: 12 pages, 8 figures

The proton capture cross-section of the least abundant proton-rich stable isotope of cadmium, $^{108}$Cd (abundance 0.89\%), has been measured near the Gamow window corresponding to a temperature range of 3-4 GK. The measurement of the $^{108}$Cd(p,$\gamma$)$^{109}$In reaction was carried out using the activation technique. The cross-section at the lowest energy point of 3T$_9$, E$_p$$^{lab}$= 2.28 MeV, has been reported for the first time. The astrophysical S-factor was measured in the energy range relevant to the astrophysical p-process, between E$_p$$^{cm}$= 2.29 and 6.79 MeV. The experimental results have been compared with theoretical predictions of Hauser-Feshbach statistical model calculations using TALYS-1.96. A calculated proton-optical potential was implemented to achieve better fitting, with different combinations of available nuclear level densities (NLDs) and $\gamma$-ray strength functions in TALYS-1.96. The calculations provided satisfactory agreement with the experimental results. The reaction rate was calculated using the calculated potential in TALYS-1.96 and compared with the values provided in the REACLIB database.

We explore the stationary densities in totally asymmetric exclusion processes (TASEP) with open boundary conditions and spatially inhomogeneous hopping rates. We calculate the steady state density profiles that characterise the associated phases. We show that in the contrast to the low and high density phases, the stationary density profile in the maximal current phase can be discontinuous, even when the space-dependent hopping rate is continuous. The phase diagrams in the plane of the control parameters show universal topology. The associated phase transitions are explored. We further investigate the domain walls, which are delocalised and calculate their envelops, which reveal their dependence on the spatial nonuniformity of the hopping rates., Comment: 17 pages, 16 figures, preliminary version

We report on the fabrication and characterization of superconducting magnesium diboride (MgB$_2$) thin films intended for quantum-limited devices based non-linear kinetic inductance (NLKI) such as parametric amplifiers with either elevated operating temperatures or expanded frequency ranges. In order to characterize the MgB$_2$ material properties, we have fabricated coplanar waveguide (CPW) transmission lines and microwave resonators using $\approx$ 40 nm thick MgB$_2$ films with a measured kinetic inductance of $\sim$ 5.5 pH/$\square$ and internal quality factors, $Q_i \approx 3 \times 10^4$ at 4.2 K. We measure the NLKI in MgB$_2$ by applying a DC bias to a 6 cm long by 4 $\mu$m wide CPW transmission line, and measuring the resulting phase delay caused by the current dependent NLKI. We also measure the current dependent NLKI through CPW resonators that shift down in frequency with increased power applied through the CPW feedline. Using these measurements, we calculate the characteristic non-linear current parameter, $I_*$, for multiple CPW geometries. We find values for corresponding current density, $J_* = 12-22$~MA/cm$^2$ and a ratio of the critical current to the non-linear current parameter, $I_C/I_* = 0.14-0.26$, similar to or higher than values for other superconductors such as NbTiN and TiN.

The applications of generative AI have become extremely impressive, and the interplay between users and AI is even more so. Current human-AI interaction literature has taken a broad look at how humans interact with generative AI, but it lacks specificity regarding the user interface designs and patterns used to create these applications. Therefore, we present a survey that comprehensively presents taxonomies of how a human interacts with AI and the user interaction patterns designed to meet the needs of a variety of relevant use cases. We focus primarily on user-guided interactions, surveying interactions that are initiated by the user and do not include any implicit signals given by the user. With this survey, we aim to create a compendium of different user-interaction patterns that can be used as a reference for designers and developers alike. In doing so, we also strive to lower the entry barrier for those attempting to learn more about the design of generative AI applications.

Covariance matrix estimation is an important problem in multivariate data analysis, both from theoretical as well as applied points of view. Many simple and popular covariance matrix estimators are known to be severely affected by model misspecification and the presence of outliers in the data; on the other hand robust estimators with reasonably high efficiency are often computationally challenging for modern large and complex datasets. In this work, we propose a new, simple, robust and highly efficient method for estimation of the location vector and the scatter matrix for elliptically symmetric distributions. The proposed estimation procedure is designed in the spirit of the minimum density power divergence (DPD) estimation approach with appropriate modifications which makes our proposal (sequential minimum DPD estimation) computationally very economical and scalable to large as well as higher dimensional datasets. Consistency and asymptotic normality of the proposed sequential estimators of the multivariate location and scatter are established along with asymptotic positive definiteness of the estimated scatter matrix. Robustness of our estimators are studied by means of influence functions. All theoretical results are illustrated further under multivariate normality. A large-scale simulation study is presented to assess finite sample performances and scalability of our method in comparison to the usual maximum likelihood estimator (MLE), the ordinary minimum DPD estimator (MDPDE) and other popular non-parametric methods. The applicability of our method is further illustrated with a real dataset on credit card transactions.

This paper addresses the medical imaging problem of joint detection in the upper limbs, viz. elbow, shoulder, wrist and finger joints. Localization of joints from X-Ray and Computerized Tomography (CT) scans is an essential step for the assessment of various bone-related medical conditions like Osteoarthritis, Rheumatoid Arthritis, and can even be used for automated bone fracture detection. Automated joint localization also detects the corresponding bones and can serve as input to deep learning-based models used for the computerized diagnosis of the aforementioned medical disorders. This in-creases the accuracy of prediction and aids the radiologists with analyzing the scans, which is quite a complex and exhausting task. This paper provides a detailed comparative study between diverse Deep Learning (DL) models - YOLOv3, YOLOv7, EfficientDet and CenterNet in multiple bone joint detections in the upper limbs of the human body. The research analyses the performance of different DL models, mathematically, graphically and visually. These models are trained and tested on a portion of the openly available MURA (musculoskeletal radiographs) dataset. The study found that the best Mean Average Precision (mAP at 0.5:0.95) values of YOLOv3, YOLOv7, EfficientDet and CenterNet are 35.3, 48.3, 46.5 and 45.9 respectively. Besides, it has been found YOLOv7 performed the best for accurately predicting the bounding boxes while YOLOv3 performed the worst in the Visual Analysis test. Code available at https://github.com/Sohambasu07/BoneJointsLocalization

Small Language Models (SLMs) have become increasingly important due to their efficiency and performance to perform various language tasks with minimal computational resources, making them ideal for various settings including on-device, mobile, edge devices, among many others. In this article, we present a comprehensive survey on SLMs, focusing on their architectures, training techniques, and model compression techniques. We propose a novel taxonomy for categorizing the methods used to optimize SLMs, including model compression, pruning, and quantization techniques. We summarize the benchmark datasets that are useful for benchmarking SLMs along with the evaluation metrics commonly used. Additionally, we highlight key open challenges that remain to be addressed. Our survey aims to serve as a valuable resource for researchers and practitioners interested in developing and deploying small yet efficient language models.

We present a new empirical framework modeling the metallicity and star-formation history (SFH) dependence of X-ray luminous ($L > 10^{36}$ ergs s$^{-1}$) point-source population luminosity functions (XLFs) in normal galaxies. We expect the X-ray point-source populations are dominated by X-ray binaries (XRBs), with contributions from supernova remnants near the low luminosity end of our observations. Our framework is calibrated using the collective statistical power of 3,731 X-ray detected point-sources within 88 Chandra-observed galaxies at $D <$ 40 Mpc that span broad ranges of metallicity ($Z \approx$ 0.03-2 $Z_\odot$), SFH, and morphology (dwarf irregulars, late-types, and early-types). Our best-fitting models indicate that the XLF normalization per unit stellar mass declines by $\approx$2-3 dex from 10 Myr to 10 Gyr, with a slower age decline for low-metallicity populations. The shape of the XLF for luminous X-ray sources ($L < 10^{38}$ ergs s$^{-1}$) significantly steepens with increasing age and metallicity, while the lower-luminosity XLF appears to flatten with increasing age. Integration of our models provide predictions for X-ray scaling relations that agree very well with past results presented in the literature, including, e.g., the $L_{\rm X}$-SFR-$Z$ relation for high-mass XRBs (HMXBs) in young stellar populations as well as the $L_{\rm X}/M_\star$ ratio observed in early-type galaxies that harbor old populations of low-mass XRBs (LMXBs). The model framework and data sets presented in this paper further provide unique benchmarks that can be used for calibrating binary population synthesis models., Comment: Accepted for publication in ApJS; extended figures/materials available at https://lehmer.uark.edu/downloads/ ; python SED fitting code Lightning available at https://github.com/ebmonson/lightningpy

Pure exploration in bandits models multiple real-world problems, such as tuning hyper-parameters or conducting user studies, where different safety, resource, and fairness constraints on the decision space naturally appear. We study these problems as pure exploration in multi-armed bandits with unknown linear constraints, where the aim is to identify an $r$$\textit{-good feasible policy}$. First, we propose a Lagrangian relaxation of the sample complexity lower bound for pure exploration under constraints. We show how this lower bound evolves with the sequential estimation of constraints. Second, we leverage the Lagrangian lower bound and the properties of convex optimisation to propose two computationally efficient extensions of Track-and-Stop and Gamified Explorer, namely LATS and LAGEX. To this end, we propose a constraint-adaptive stopping rule, and while tracking the lower bound, use pessimistic estimate of the feasible set at each step. We show that these algorithms achieve asymptotically optimal sample complexity upper bounds up to constraint-dependent constants. Finally, we conduct numerical experiments with different reward distributions and constraints that validate efficient performance of LAGEX and LATS with respect to baselines.

From the vasculature of animals to the porous media making up batteries, transport by fluid flow within complex networks is crucial to service all cells or media with resources. Yet, living flow networks have a key advantage over porous media: they are adaptive and can self-organize their geometry to achieve a homogeneous perfusion throughout the network. Here, we show that, through erosion, artificial flow networks self-organize to a geometry where perfusion is more homogeneous. Flowing a pulse of cleaving enzyme through a network patterned into an erodible hydrogel, with initial channels disparate in width, we observe a homogenization in channel resistances. Experimental observations are matched with numerical simulations of the diffusion-advection-sorption dynamics of an eroding enzyme within a network. Analyzing transport dynamics theoretically, we show that homogenization only occurs if the pulse of eroding enzyme lasts longer than the time it takes any channel to equilibrate to the pulse concentration. The equilibration time scale derived analytically is in agreement with simulations. Lastly, we show both numerically and experimentally that erosion leads to homogenization of complex networks containing loops. Erosion being an omnipresent reaction, our results pave the way for a very versatile self-organized increase in the performance of porous media.

We study the cogenesis of baryon and dark matter (DM) in an extended reheating period after the end of slow-roll inflation. Within the regime of perturbative reheating, we consider different monomial potential of the inflaton field during reheating era. The inflaton condensate reheats the Universe by decaying into the Standard Model (SM) bath either via fermionic or bosonic decay modes. Assuming the leptogenesis route to baryogenesis in a canonical seesaw framework, we consider the bath to produce such RHNs during the period of reheating when the maximum temperature $T_{\rm max}$ of the SM bath is well above the reheating temperature $T_{\rm rh}$. The DM, assumed to be a SM gauge singlet field, also gets produced from the bath during the reheating period via UV freeze-in. In addition to obtaining different parameter space for such non-thermal leptogenesis and DM for both bosonic and fermionic reheating modes and the type of monomial potential, we discuss the possibility of probing such scenarios via spectral shape of primordial gravitational waves., Comment: 25 pages, 2 tables and 6 figures

We use a custom-made calibrator to measure individual detectors' polarization angles of BICEP3, a small aperture telescope observing the cosmic microwave background (CMB) at 95GHz from the South Pole. We describe our calibration strategy and the statistical and systematic uncertainties associated with the measurement. We reach an unprecedented precision for such measurement on a CMB experiment, with a repeatability for each detector pair of $0.02\deg$. We show that the relative angles measured using this method are in excellent agreement with those extracted from CMB data. Because the absolute measurement is currently limited by a systematic uncertainty, we do not derive cosmic birefringence constraints from BICEP3 data in this work. Rather, we forecast the sensitivity of BICEP3 sky maps for such analysis. We investigate the relative contributions of instrument noise, lensing, and dust, as well as astrophysical and instrumental systematics. We also explore the constraining power of different angle estimators, depending on analysis choices. We establish that the BICEP3 2-year dataset (2017--2018) has an on-sky sensitivity to the cosmic birefringence angle of $\sigma = 0.078\deg$, which could be improved to $\sigma = 0.055\deg$ by adding all of the existing BICEP3 data (through 2023). Furthermore, we emphasize the possibility of using the BICEP3 sky patch as a polarization calibration source for CMB experiments, which with the present data could reach a precision of $0.035\deg$. Finally, in the context of inflation searches, we investigate the impact of detector-to-detector variations in polarization angles as they may bias the tensor-to-scalar ratio r. We show that while the effect is expected to remain subdominant to other sources of systematic uncertainty, it can be reliably calibrated using polarization angle measurements such as the ones we present in this paper., Comment: 29 Pages, 17 Figures, 6 Tables, as submitted to PRD

As LLMs are increasingly deployed in global applications, the importance of cultural sensitivity becomes paramount, ensuring that users from diverse backgrounds feel respected and understood. Cultural harm can arise when these models fail to align with specific cultural norms, resulting in misrepresentations or violations of cultural values. This work addresses the challenges of ensuring cultural sensitivity in LLMs, especially in small-parameter models that often lack the extensive training data needed to capture global cultural nuances. We present two key contributions: (1) A cultural harm test dataset, created to assess model outputs across different cultural contexts through scenarios that expose potential cultural insensitivities, and (2) A culturally aligned preference dataset, aimed at restoring cultural sensitivity through fine-tuning based on feedback from diverse annotators. These datasets facilitate the evaluation and enhancement of LLMs, ensuring their ethical and safe deployment across different cultural landscapes. Our results show that integrating culturally aligned feedback leads to a marked improvement in model behavior, significantly reducing the likelihood of generating culturally insensitive or harmful content. Ultimately, this work paves the way for more inclusive and respectful AI systems, fostering a future where LLMs can safely and ethically navigate the complexities of diverse cultural landscapes.

The scaled and centered largest eigenvalue of the $n\times n$ principal minor of an infinite GUE matrix, denoted by $\widetilde{\lambda}_n$, converges to the GUE Tracy-Widom distribution. We show that $\liminf\limits_{n\to \infty}(\log n)^{-1/3}\widetilde{\lambda}_n= -4^{1/3}$ almost surely, confirming a conjecture of Paquette and Zeitouni (Ann. Probab., 2017). Our arguments rely on the recently established sharp lower tail estimate for the largest eigenvalue of GUE from (Baslingker et al., 2024) and a correspondence between the GUE minor process and Brownian last passage percolation, due to Baryshnikov (Probab. Theory Related Fields, 2001). The latter result is used in a geometric argument to show the decorrelation of the largest eigenvalues at the appropriate scale. Our arguments also give a new proof of the corresponding law of fractional logarithms for the limsup, established by Paquette and Zeitouni (Ann. Probab., 2017)., Comment: 16 pages, 4 figures

With the pervasive deployment of Machine Learning (ML) models in real-world applications, verifying and auditing properties of ML models have become a central concern. In this work, we focus on three properties: robustness, individual fairness, and group fairness. We discuss two approaches for auditing ML model properties: estimation with and without reconstruction of the target model under audit. Though the first approach is studied in the literature, the second approach remains unexplored. For this purpose, we develop a new framework that quantifies different properties in terms of the Fourier coefficients of the ML model under audit but does not parametrically reconstruct it. We propose the Active Fourier Auditor (AFA), which queries sample points according to the Fourier coefficients of the ML model, and further estimates the properties. We derive high probability error bounds on AFA's estimates, along with the worst-case lower bounds on the sample complexity to audit them. Numerically we demonstrate on multiple datasets and models that AFA is more accurate and sample-efficient to estimate the properties of interest than the baselines.

The elementary action of symplectic and orthogonal groups on unimodular rows of length $2n$ is transitive for $2n \geq \max(4, d+2)$ in the symplectic case, and $2n \geq \max(6, 2d+4)$ in the orthogonal case, over monoid rings $R[M]$, where $R$ is a commutative noetherian ring of dimension $d$, and $M$ is commutative cancellative torsion free monoid. As a consequence, one gets the surjective stabilization bound for the $K_1$ for classical groups. This is an extension of J. Gubeladze's results for linear groups., Comment: Transformation Groups (2024)

Independent sets in graphs, i.e., subsets of vertices where no two are adjacent, have long been studied, for instance as a model of hard-core gas. The $d$-dimensional hypercube, $\{0,1\}^d$, with the nearest neighbor structure, has been a particularly appealing choice for the base graph, owing in part to its many symmetries. Results go back to the work of Korshunov and Sapozhenko who proved sharp results on the count of such sets as well as structure theorems for random samples drawn uniformly. Of much interest is the behavior of such Gibbs measures in the presence of disorder. In this direction, Kronenberg and Spinka [KS] initiated the study of independent sets in a random subgraph of the hypercube obtained by considering an instance of bond percolation with probability $p$. Relying on tools from statistical mechanics they obtained a detailed understanding of the moments of the partition function, say $\mathcal{Z}$, of the hard-core model on such random graphs and consequently deduced certain fluctuation information, as well as posed a series of interesting questions. In particular, they showed in the uniform case that there is a natural phase transition at $p=2/3$ where $\mathcal{Z}$ transitions from being concentrated for $p>2/3$ to not concentrated at $p=2/3$. In this article, developing a probabilistic framework, as well as relying on certain cluster expansion inputs from [KS], we present a detailed picture of both the fluctuations of $\mathcal{Z}$ as well as the geometry of a randomly sampled independent set. In particular, we establish that $\mathcal{Z}$, properly centered and scaled, converges to a standard Gaussian for $p>2/3$, and to a sum of two i.i.d. log-normals at $p=2/3$. A particular step in the proof which could be of independent interest involves a non-uniform birthday problem for which collisions emerge at $p=2/3$., Comment: 35 pages, 1 figure. Abstract shortened to meet arXiv requirements

In this work, we investigate the entanglement structure in a $\textrm{T}\bar{\textrm{T}}$-deformed holographic CFT$_2$ with a conserved angular momentum. We utilize conformal perturbation theory to compute the leading order correction to the entanglement entropy and the reflected entropy due to the $\textrm{T}\bar{\textrm{T}}$ deformation in the limit of large central charge. In the dual bulk perspective described by a rotating BTZ black hole with a finite radial cut-off, we compute the holographic entanglement entropy and the entanglement wedge cross-section and observe perfect agreement with our field theoretic computation for small values of the deformation parameter., Comment: 33 pages, 13 figures and 1 appendix

Testing whether a sample survey is a credible representation of the population is an important question to ensure the validity of any downstream research. While this problem, in general, does not have an efficient solution, one might take a task-based approach and aim to understand whether a certain data analysis tool, like linear regression, would yield similar answers both on the population and the sample survey. In this paper, we design an algorithm to test the credibility of a sample survey in terms of linear regression. In other words, we design an algorithm that can certify if a sample survey is good enough to guarantee the correctness of data analysis done using linear regression tools. Nowadays, one is naturally concerned about data privacy in surveys. Thus, we further test the credibility of surveys published in a differentially private manner. Specifically, we focus on Local Differential Privacy (LDP), which is a standard technique to ensure privacy in surveys where the survey participants might not trust the aggregator. We extend our algorithm to work even when the data analysis has been done using surveys with LDP. In the process, we also propose an algorithm that learns with high probability the guarantees a linear regression model on a survey published with LDP. Our algorithm also serves as a mechanism to learn linear regression models from data corrupted with noise coming from any subexponential distribution. We prove that it achieves the optimal estimation error bound for $\ell_1$ linear regression, which might be of broader interest. We prove the theoretical correctness of our algorithms while trying to reduce the sample complexity for both public and private surveys. We also numerically demonstrate the performance of our algorithms on real and synthetic datasets.

In this paper, we study a non-Hermitian (NH) ladder subjected to a variety of driving protocols. The driven system looses chiral symmetry (CS) whose presence is indispensable for its topological characterization. Further, the bulk boundary correspondence (BBC) gets adversely affected due to the presence of non-Hermitian skin effect (NHSE). Here, we present a formalism that not only retrieves the lost CS, but also restores the BBC via the construction of a generalized Brillouin zone (GBZ). Specifically, we employ delta and step drives to compare and contrast between them with regard to their impact on NHSE. Further, a widely studied harmonic drive is invoked in this context, not only for the sake of completeness, but its distinct computational framework offers valuable insights on the properties of out-of-equilibrium systems. While the delta and the harmonic drives exhibit unidirectional skin effect in the system, the step drive may show bi-directional skin effect. Also, there are specific points in the parameter space that are devoid of skin effect. These act as critical points that distinguish the skin modes to be localized at one boundary or the other. Moreover, for the computation of the non-Bloch invariants, we employ GBZ via a pair of symmetric time frames corresponding to the delta and the step drives, while a high-frequency expansion was carried out to deal with the harmonic drive. Finally, we present phase boundary diagrams that demarcate distinct NH phases obtained via tracking the trajectories of the exceptional points. These diagrams demonstrate a co-existence of the zero and $\pi$ energy modes in the strong NH limit and thus may be relevant for studies of Floquet time crystals.

It has been demonstrated that networks' parameters can be significantly reduced in the frequency domain with a very small decrease in accuracy. However, given the cost of frequency transforms, the computational complexity is not significantly decreased. In this work, we propose performing network inference in the frequency domain to speed up networks whose frequency parameters are sparse. In particular, we propose a frequency inference chain that is dual to the network inference in the spatial domain. In order to handle the non-linear layers, we make a compromise to apply non-linear operations on frequency data directly, which works effectively. Enabled by the frequency inference chain and the strategy for non-linear layers, the proposed approach completes the entire inference in the frequency domain. Unlike previous approaches which require extra frequency or inverse transforms for all layers, the proposed approach only needs the frequency transform and its inverse once at the beginning and once at the end of a network. Comparisons with state-of-the-art methods demonstrate that the proposed approach significantly improves accuracy in the case of a high speedup ratio (over 100x). The source code is available at \url{https://github.com/guanfangdong/FreqNet-Infer}., Comment: accepted by ACM Multimedia Asia 2024

This study employs advanced phase-field modeling to investigate Si-based qubit MOSFETs, integrating electrostatics and quantum mechanical effects. We adopt a comprehensive modeling approach, utilizing full-wave treatment of the Schrodinger equation solutions, coupled with the Poisson equation at cryogenic temperatures. Our analysis explores the influence of interface traps on quantum dot (QD) barrier heights, affecting coupling due to tunneling. A wider trap distribution leads to the decoupling of quantum dots. Furthermore, the oscillations in the transmission and reflection coefficients increase as the plunger/barrier gate length increases, reducing the coupling between the QDs. By optimizing plunger and barrier gate dimensions, spacer configurations, and gap oxide lengths, we enhance control over quantum well depths and minimize unwanted wave function leakage. The modeling algorithm is also validated against the experimental data and can accurately capture the oscillations in the Id Vgs caused by the Coulomb blockade at cryogenic temperature, Comment: 10 pages

Inhalation directs air through a defined pathway, initiating from nostrils, moving through the main nasal cavity, past the pharynx and trachea, and culminating in the lungs. Inhaled particles, of a range of sizes, are ferried by this incoming air but are filtered and trapped by upper airway structures to protect the delicate lower respiratory system. From an energetics perspective, the airflow physics along this convoluted tract is characterized by turbulence. The system approaches a critical stationary state over the time scales during which particles enter the airway and deposit. This stasis can be conjectured to correspond with the emergence of criticality in the complex flow domain. For such systemic criticality (i.e., sensitivity to perturbations), inhaled particle deposition impacted by the surrounding flow processes can act as signature avalanche-like events. Based on the principles of organized criticality, we have explored the emergence of power law trends in particle deposition levels at the nasopharynx, a key initial infection site for airborne pathogens. These trends are derived from numerical data from five anatomic airway geometries for 15-85 L/min inhalation rates, modeled using high-fidelity Large Eddy Simulations., Comment: 9 pages, 12 figures

This paper presents a rapid approach to assessing disaster resilience in Florida, particularly regarding Hurricane Helene (2024). This category four storm made landfall on Florida's Gulf Coast in September 2024. Using the Disaster Resilience Index (DRI) developed in this paper, the preparedness and adaptive capacities of communities across counties in Florida are evaluated, identifying the most resilient areas based on three key variables: population size, average per-person income, and the Social Vulnerability Index (SVI). While the Social Vulnerability Index (SVI) accounts for factors like socioeconomic status, household composition, minority status, and housing conditions-key elements in determining a community's resilience to disasters-incorporating a county's population and per person income provides additional insight. A county's total population is directly linked to the number of individuals impacted by a disaster, while personal income reflects a household's capacity to recover. Spatial analysis was performed on the index to compare the vulnerability and resilience levels across thirty-four counties vulnerable to Hurricane Helene's projected path. The results highlight that counties with high income and lower population densities, such as Monroe and Collier, exhibit greater resilience. In contrast, areas with larger populations and higher social vulnerabilities are at greater risk of damage. This study contributes to disaster management planning by providing a rapid yet comprehensive and reassuring socioeconomic impact assessment, offering actionable insights for anticipatory measures and resource allocation.

We have used solar oscillation frequencies and frequency splittings obtained over solar cycles 23, 24 and the rising phase of solar cycle 25 to investigate whether the tachocline properties (jump i.e., the change in the rotation rate across the tachocline, width and position) show any time variation. We confirm that the change in rotation rate across the tachocline changes substantially, however, the change does not show a simple correlation with solar cycle unlike, for instance, changes in mode frequencies. The change during the ascending phase of solar cycle 25 is almost a mirror image of the change during the descending part of solar cycle 24, tempting us to speculate that the tachocline has a much longer period than either the sunspot or the magnetic cycle. We also find that the position of the tachocline, defined as the mid-point of the change in rotation rate, showed significant changes during solar cycle 24. The width of the tachocline, on the other hand, has showed significant changes during solar cycle 23, but not later. The change in the tachocline becomes more visible if we look at the upper and lower extents of the tachocline, defined as (position +/- width). We find that for epochs around solar maxima and minima, the extent decreases before increasing again - a few more years of data should clarify this trend. Our results reinforce the need to continue helioseismic monitoring of the Sun to understand solar activity and its evolution., Comment: Accepted for publication in ApJ

Diseases in plants cause significant danger to productive and secure agriculture. Plant diseases can be detected early and accurately, reducing crop losses and pesticide use. Traditional methods of plant disease identification, on the other hand, are generally time-consuming and require professional expertise. It would be beneficial to the farmers if they could detect the disease quickly by taking images of the leaf directly. This will be a time-saving process and they can take remedial actions immediately. To achieve this a novel feature extraction approach for detecting tomato plant illnesses from leaf photos using low-cost computing systems such as mobile phones is proposed in this study. The proposed approach integrates various types of Deep Learning techniques to extract robust and discriminative features from leaf images. After the proposed feature extraction comparisons have been made on five cutting-edge deep learning models: AlexNet, ResNet50, VGG16, VGG19, and MobileNet. The dataset contains 10,000 leaf photos from ten classes of tomato illnesses and one class of healthy leaves. Experimental findings demonstrate that AlexNet has an accuracy score of 87%, with the benefit of being quick and lightweight, making it appropriate for use on embedded systems and other low-processing devices like smartphones., Comment: 10 Pages, 8 figures, 1 table

Retrieval Augmented Generation (RAG) is a widely used approach for leveraging external context in several natural language applications such as question answering and information retrieval. Yet, the exact nature in which a Language Model (LM) leverages this non-parametric memory or retrieved context isn't clearly understood. This paper mechanistically examines the RAG pipeline to highlight that LMs demonstrate a "shortcut'' effect and have a strong bias towards utilizing the retrieved context to answer questions, while relying minimally on model priors. We propose (a) Causal Mediation Analysis; for proving that parametric memory is minimally utilized when answering a question and (b) Attention Contributions and Knockouts for showing the last token residual stream do not get enriched from the subject token in the question, but gets enriched from tokens of RAG-context. We find this pronounced "shortcut'' behaviour to be true across both LLMs (e.g.,LlaMa) and SLMs (e.g., Phi), Comment: Accepted to Blackbox NLP @ EMNLP 2024

The financial domain presents a complex environment for stock market prediction, characterized by volatile patterns and the influence of multifaceted data sources. Traditional models have leveraged either Convolutional Neural Networks (CNN) for spatial feature extraction or Long Short-Term Memory (LSTM) networks for capturing temporal dependencies, with limited integration of external textual data. This paper proposes a novel Two-Level Conv-LSTM Neural Network integrated with a Large Language Model (LLM) for comprehensive stock advising. The model harnesses the strengths of Conv-LSTM for analyzing time-series data and LLM for processing and understanding textual information from financial news, social media, and reports. In the first level, convolutional layers are employed to identify local patterns in historical stock prices and technical indicators, followed by LSTM layers to capture the temporal dynamics. The second level integrates the output with an LLM that analyzes sentiment and contextual information from textual data, providing a holistic view of market conditions. The combined approach aims to improve prediction accuracy and provide contextually rich stock advising., Comment: 8 pages, 2 figures, 2 tables