Search

While quantum computing holds immense potential for tackling previously intractable problems, its current practicality remains limited. A critical aspect of realizing quantum utility is the ability to efficiently interface with data from the classical world. This research focuses on the crucial phase of quantum encoding, which enables the transformation of classical information into quantum states for processing within quantum systems. We focus on three prominent encoding models: Phase Encoding, Qubit Lattice, and Flexible Representation of Quantum Images (FRQI) for cost and efficiency analysis. The aim of quantifying their different characteristics is to analyze their impact on quantum processing workflows. This comparative analysis offers valuable insights into their limitations and potential to accelerate the development of practical quantum computing solutions., Comment: 11 Pages, 13 Figures, In press: 31st IEEE International Conference on High Performance Computing, Data, and Analytics (HiPC) - 2024

We present observations of the Type IIP supernova (SN) 2024jlf, including spectroscopy beginning just 0.7 days ($\sim$17 hours) after first light. Rapid follow-up was enabled by the new $\texttt{BTSbot-nearby}$ program, which involves autonomously triggering target-of-opportunity requests for new transients in Zwicky Transient Facility data that are coincident with nearby ($D<60$ Mpc) galaxies and identified by the $\texttt{BTSbot}$ machine learning model. Early photometry and non-detections shortly prior to first light show that SN 2024jlf initially brightened by $>$4 mag/day, quicker than $\sim$90% of Type II SNe. Early spectra reveal weak flash ionization features: narrow, short-lived ($1.3 < \tau ~\mathrm{[d]} < 1.8$) emission lines of H$\alpha$, He II, and C IV. Assuming a wind velocity of $v_w=50$ km s$^{-1}$, these properties indicate that the red supergiant progenitor exhibited enhanced mass-loss in the last year before explosion. We constrain the mass-loss rate to $10^{-4} < \dot{M}~\mathrm{[M_\odot~yr^{-1}]} < 10^{-3}$ by matching observations to model grids from two independent radiative hydrodynamics codes. $\texttt{BTSbot-nearby}$ automation minimizes spectroscopic follow-up latency, enabling the observation of ephemeral early-time phenomena exhibited by transients., Comment: 23 pages, 9 figures

Autonomous systems are increasingly expected to operate in the presence of adversaries, though an adversary may infer sensitive information simply by observing a system, without even needing to interact with it. Therefore, in this work we present a deceptive decision-making framework that not only conceals sensitive information, but in fact actively misleads adversaries about it. We model autonomous systems as Markov decision processes, and we consider adversaries that attempt to infer their reward functions using inverse reinforcement learning. To counter such efforts, we present two regularization strategies for policy synthesis problems that actively deceive an adversary about a system's underlying rewards. The first form of deception is ``diversionary'', and it leads an adversary to draw any false conclusion about what the system's reward function is. The second form of deception is ``targeted'', and it leads an adversary to draw a specific false conclusion about what the system's reward function is. We then show how each form of deception can be implemented in policy optimization problems, and we analytically bound the loss in total accumulated reward that is induced by deception. Next, we evaluate these developments in a multi-agent sequential decision-making problem with one real agent and multiple decoys. We show that diversionary deception can cause the adversary to believe that the most important agent is the least important, while attaining a total accumulated reward that is $98.83\%$ of its optimal, non-deceptive value. Similarly, we show that targeted deception can make any decoy appear to be the most important agent, while still attaining a total accumulated reward that is $99.25\%$ of its optimal, non-deceptive value., Comment: 21 pages, 5 figures

In multiplayer games with sequential decision-making, self-interested players form dynamic coalitions to achieve most-preferred temporal goals beyond their individual capabilities. We introduce a novel procedure to synthesize strategies that jointly determine which coalitions should form and the actions coalition members should choose to satisfy their preferences in a subclass of deterministic multiplayer games on graphs. In these games, a leader decides the coalition during each round and the players not in the coalition follow their admissible strategies. Our contributions are threefold. First, we extend the concept of admissibility to games on graphs with preferences and characterize it using maximal sure winning, a concept originally defined for adversarial two-player games with preferences. Second, we define a value function that assigns a vector to each state, identifying which player has a maximal sure winning strategy for certain subset of objectives. Finally, we present a polynomial-time algorithm to synthesize admissible strategies for all players based on this value function and prove their existence in all games within the chosen subclass. We illustrate the benefits of dynamic coalitions over fixed ones in a blocks-world domain. Interestingly, our experiment reveals that aligned preferences do not always encourage cooperation, while conflicting preferences do not always lead to adversarial behavior., Comment: 9 pages, 3 figures

Ethical intervention prompting has emerged as a tool to counter demographic biases of text-to-image generative AI models. Existing solutions either require to retrain the model or struggle to generate images that reflect desired distributions on gender and race. We propose an inference-time process called DebiasPI for Debiasing-by-Prompt-Iteration that provides prompt intervention by enabling the user to control the distributions of individuals' demographic attributes in image generation. DebiasPI keeps track of which attributes have been generated either by probing the internal state of the model or by using external attribute classifiers. Its control loop guides the text-to-image model to select not yet sufficiently represented attributes, With DebiasPI, we were able to create images with equal representations of race and gender that visualize challenging concepts of news headlines. We also experimented with the attributes age, body type, profession, and skin tone, and measured how attributes change when our intervention prompt targets the distribution of an unrelated attribute type. We found, for example, if the text-to-image model is asked to balance racial representation, gender representation improves but the skin tone becomes less diverse. Attempts to cover a wide range of skin colors with various intervention prompts showed that the model struggles to generate the palest skin tones. We conducted various ablation studies, in which we removed DebiasPI's attribute control, that reveal the model's propensity to generate young, male characters. It sometimes visualized career success by generating two-panel images with a pre-success dark-skinned person becoming light-skinned with success, or switching gender from pre-success female to post-success male, thus further motivating ethical intervention prompting with DebiasPI., Comment: This work was presented at The European Conference on Computer Vision (ECCV) 2024 Workshop "Fairness and ethics towards transparent AI: facing the chalLEnge through model Debiasing" (FAILED), Milano, Italy, on September 29, 2024, https://failed-workshop-eccv-2024.github.io

We present a modular pipeline that automates the generation of stealthy jailbreak prompts derived from high-level content policies, enhancing LLM content moderation. First, we address query inefficiency and jailbreak strength by developing Graph of Attacks with Pruning (GAP), a method that utilizes strategies from prior jailbreaks, resulting in 92% attack success rate on GPT-3.5 using only 54% of the queries of the prior algorithm. Second, we address the cold-start issue by automatically generating seed prompts from the high-level policy using LLMs. Finally, we demonstrate the utility of these generated jailbreak prompts of improving content moderation by fine-tuning PromptGuard, a model trained to detect jailbreaks, increasing its accuracy on the Toxic-Chat dataset from 5.1% to 93.89%., Comment: 15 pages, 7 figures

Battery-powered technologies like pagers and walkie-talkies have long been integral to civilian and military operations. However, the potential for such everyday devices to be weaponized has largely been underestimated in the realm of cybersecurity. In September 2024, Lebanon experienced a series of unprecedented, coordinated explosions triggered through compromised pagers and walkie-talkies, creating a new category of attack in the domain of cyber-physical warfare. This attack not only disrupted critical communication networks but also resulted in injuries, loss of life, and exposed significant national security vulnerabilities, prompting governments and organizations worldwide to reevaluate their cybersecurity frameworks. This article provides an in-depth investigation into the infamous Pager and Walkie-Talkie attacks, analyzing both technical and non-technical dimensions. Furthermore, the study extends its scope to explore vulnerabilities in other battery-powered infrastructures, such as battery management systems, highlighting their potential exploitation. Existing prevention and detection techniques are reviewed, with an emphasis on their limitations and the challenges they face in addressing emerging threats. Finally, the article discusses emerging methodologies, particularly focusing on the role of physical inspection, as a critical component of future security measures. This research aims to provide actionable insights to bolster the resilience of cyber-physical systems in an increasingly interconnected world., Comment: 18 pages, 10 figures

Data science has emerged as fourth paradigm of scientific exploration. However many machine learning models operate as black boxes offering limited insight into the reasoning behind their predictions. This lack of transparency is one of the drawbacks to generate new knowledge from data. Recently Kolmogorov-Arnold Network or KAN has been proposed as an alternative model which embeds explainable AI. This study demonstrates the potential of KAN for new scientific exploration. KAN along with five other popular supervised machine learning models are applied to the well-known problem of airfoil lift prediction in aerospace engineering. Standard data generated from an earlier study on 2900 different airfoils is used. KAN performed the best with an R2 score of 96.17 percent on the test data, surpassing both the baseline model and Multi Layer Perceptron. Explainability of KAN is shown by pruning and symbolizing the model resulting in an equation for coefficient of lift in terms of input variables. The explainable information retrieved from KAN model is found to be consistent with the known physics of lift generation by airfoil thus demonstrating its potential to aid in scientific exploration., Comment: 3 pages, 2 tables, 3 figures

Nash equilibrium is a fundamental solution concept for modeling the behavior of self-interested agents. We develop an algorithm to synthesize pure Nash equilibria in two-player deterministic games on graphs where players have partial preferences over objectives expressed with linear temporal logic over finite traces. Previous approaches for Nash equilibrium synthesis assume that players' preferences are common knowledge. Instead, we allow players' preferences to remain private but enable communication between players. The algorithm we design synthesizes Nash equilibria for a complete-information game, but synthesizes these equilibria in an incomplete-information setting where players' preferences are private. The algorithm is privacy-aware, as instead of requiring that players share private preferences, the algorithm reduces the information sharing to a query interface. Through this interface, players exchange information about states in the game from which they can enforce a more desirable outcome. We prove the algorithm's completeness by showing that it either returns an equilibrium or certifies that one does not exist. We then demonstrate, via numerical examples, the existence of games where the queries the players exchange are insufficient to reconstruct players' preferences, highlighting the privacy-aware nature of the algorithm we propose., Comment: 13 pages, 6 figures

Ensuring that AI systems make strategic decisions aligned with the specified preferences in adversarial sequential interactions is a critical challenge for developing trustworthy AI systems, especially when the environment is stochastic and players' incomplete preferences leave some outcomes unranked. We study the problem of synthesizing preference-satisfying strategies in two-player stochastic games on graphs where players have opposite (possibly incomplete) preferences over a set of temporal goals. We represent these goals using linear temporal logic over finite traces (LTLf), which enables modeling the nuances of human preferences where temporal goals need not be mutually exclusive and comparison between some goals may be unspecified. We introduce a solution concept of non-dominated almost-sure winning, which guarantees to achieve a most preferred outcome aligned with specified preferences while maintaining robustness against the adversarial behaviors of the opponent. Our results show that strategy profiles based on this concept are Nash equilibria in the game where players are risk-averse, thus providing a practical framework for evaluating and ensuring stable, preference-aligned outcomes in the game. Using a drone delivery example, we demonstrate that our contributions offer valuable insights not only for synthesizing rational behavior under incomplete preferences but also for designing games that motivate the desired behavior from the players in adversarial conditions., Comment: 9 pages, 3 figures, accepted at AAAI 2025 (AI alignment track)

Population III stars are possible precursors to early massive and supermassive black holes (BHs). The presence of soft UV Lyman Werner (LW) background radiation can suppress Population III star formation in minihalos and allow them to form in pristine atomic cooling halos. In the absence of molecular hydrogen ($\rm H_2$) cooling, atomic-cooling halos enable rapid collapse with suppressed fragmentation. High background LW fluxes from preceding star-formation have been proposed to dissociate $\rm H_2$. This flux can be supplemented by LW radiation from one or more Population III star(s) in the same halo, reducing the necessary background level. Here we consider atomic-cooling halos in which multiple protostellar cores form close to one another nearly simultaneously. We assess whether the first star's LW radiation can dissociate nearby $\rm H_2$, enabling the prompt formation of a second, supermassive star (SMS) from warm, atomically-cooled gas. We use a set of hydrodynamical simulations with the code ENZO, with identical LW backgrounds centered on a halo with two adjacent collapsing gas clumps. When an additional large local LW flux is introduced, we observe immediate reductions in both the accretion rates and the stellar masses that form within these clumps. While the LW flux reduces the $\text{H}_2$ fraction and increases the gas temperature, the halo core's potential well is too shallow to promptly heat the gas to $\gtrsim$ 1000 K and increase the accretion rate onto the second protostar. We conclude that internal LW feedback inside atomic-cooling halos is unlikely to facilitate the formation of SMSs or massive BH seeds., Comment: 19 pages, 21 figures, and 1 table

We consider time dynamics of entanglement entropy between a filled fermionic system and an empty reservoir. We consider scenarios (i) where the system is subjected to a dephasing mechanism and the reservoir is clean, thereby emulating expansion of effectively interacting fermions in vacuum, and (ii) where both the system and the reservoir are subjected to dephasing and thereby enabling us to address how the entanglement between the part of the effectively interacting system and its complement evolves in time. We consider two different kinds of quantum trajectory approaches, namely stochastic unitary unraveling and quantum state diffusion. For both protocols, we observe and characterize the full Page curve-like dynamics for the entanglement entropy. Depending on the protocol and the setup, we observe very distinct characteristics of the Page curve and the associated Page time and Page value. We also compute the number of fermions leaking to the reservoir and the associated current and shed light on their plausible connections with entanglement entropy. Our findings are expected to hold for a wide variety of generic interacting quantum systems., Comment: 18 pages, 8 figures (including supplementary material)

Automatic speech recognition (ASR) systems are well known to perform poorly on dysarthric speech. Previous works have addressed this by speaking rate modification to reduce the mismatch with typical speech. Unfortunately, these approaches rely on transcribed speech data to estimate speaking rates and phoneme durations, which might not be available for unseen speakers. Therefore, we combine unsupervised rhythm and voice conversion methods based on self-supervised speech representations to map dysarthric to typical speech. We evaluate the outputs with a large ASR model pre-trained on healthy speech without further fine-tuning and find that the proposed rhythm conversion especially improves performance for speakers of the Torgo corpus with more severe cases of dysarthria. Code and audio samples are available at https://idiap.github.io/RnV ., Comment: Accepted at ICASSP 2025 Satellite Workshop: Workshop on Speech Pathology Analysis and DEtection (SPADE)

Scientific advancements rely on high-performance computing (HPC) applications that model real-world phenomena through simulations. These applications process vast amounts of data on specialized accelerators (eg., GPUs) using special libraries. Heterogeneous bugs occur in these applications when managing data movement across different platforms, such as CPUs and GPUs, leading to divergent behavior when using heterogeneous platforms compared to using only CPUs. Existing software testing techniques often fail to detect such bugs because either they do not account for platform-specific characteristics or target specific platforms. To address this problem, we present HeteroBugDetect, an automated approach to detect platform-dependent heterogeneous bugs in HPC scientific applications. HeteroBugDetect combines natural-language processing, off-target testing, custom fuzzing, and differential testing to provide an end-to-end solution for detecting platform-specific bugs in scientific applications. We evaluate HeteroBugDetect on LAMMPS, a molecular dynamics simulator, where it detected multiple heterogeneous bugs, enhancing its reliability across diverse HPC environments.

We introduce the notion of round surgery diagrams in $S^3$ for representing 3-manifolds similar to Dehn surgery diagrams. We give a correspondence between a certain class of round surgery diagrams and Dehn surgery diagrams for 3-manifolds. As a consequence, we recover Asimov's result, stating that any closed connected oriented 3-manifold can be obtained by a round surgery on a framed link in $S^3$. There may be more than one round surgery diagram giving rise to the same 3-manifold. Thus, it is natural to ask whether there is a version of Kirby Calculus for round surgery diagrams, similar to the case of Dehn surgery diagrams with integral framings. In this direction, we define four types of moves on round surgery diagrams such that any two round surgery diagrams corresponding to the same 3-manifold can be obtained one from another by a finite sequence of these moves, thereby establishing a version of Kirby Calculus. As an application, we prove the existence of taut foliations, hence the existence of tight contact structures on 3-manifolds obtained by round 1-surgery on fibred links with two components on $S^3$., Comment: 33 pages, 32 figures

The average-reward formulation of reinforcement learning (RL) has drawn increased interest in recent years due to its ability to solve temporally-extended problems without discounting. Independently, RL algorithms have benefited from entropy-regularization: an approach used to make the optimal policy stochastic, thereby more robust to noise. Despite the distinct benefits of the two approaches, the combination of entropy regularization with an average-reward objective is not well-studied in the literature and there has been limited development of algorithms for this setting. To address this gap in the field, we develop algorithms for solving entropy-regularized average-reward RL problems with function approximation. We experimentally validate our method, comparing it with existing algorithms on standard benchmarks for RL., Comment: Accepted at the AAAI-25 Eighth Workshop on Bridging the Gap Between AI Planning and Reinforcement Learning (PRL)

In reinforcement learning, two objective functions have been developed extensively in the literature: discounted and averaged rewards. The generalization to an entropy-regularized setting has led to improved robustness and exploration for both of these objectives. Recently, the entropy-regularized average-reward problem was addressed using tools from large deviation theory in the tabular setting. This method has the advantage of linearity, providing access to both the optimal policy and average reward-rate through properties of a single matrix. In this paper, we extend that framework to more general settings by developing approaches based on function approximation by neural networks. This formulation reveals new theoretical insights into the relationship between different objectives used in RL. Additionally, we combine our algorithm with a posterior policy iteration scheme, showing how our approach can also solve the average-reward RL problem without entropy-regularization. Using classic control benchmarks, we experimentally find that our method compares favorably with other algorithms in terms of stability and rate of convergence., Comment: Accepted at the AAAI-25 8th Workshop on Generalization in Planning. arXiv admin note: text overlap with arXiv:2501.09080

In large-scale software development, understanding the functionality and intent behind complex codebases is critical for effective development and maintenance. While code summarization has been widely studied, existing methods primarily focus on smaller code units, such as functions, and struggle with larger code artifacts like files and packages. Additionally, current summarization models tend to emphasize low-level implementation details, often overlooking the domain and business context that are crucial for real-world applications. This paper proposes a two-step hierarchical approach for repository-level code summarization, tailored to business applications. First, smaller code units such as functions and variables are identified using syntax analysis and summarized with local LLMs. These summaries are then aggregated to generate higher-level file and package summaries. To ensure the summaries are grounded in business context, we design custom prompts that capture the intended purpose of code artifacts based on the domain and problem context of the business application. We evaluate our approach on a business support system (BSS) for the telecommunications domain, showing that syntax analysis-based hierarchical summarization improves coverage, while business-context grounding enhances the relevance of the generated summaries., Comment: To appear at LLM4Code@ICSE 2025

Quantizing the weights of a neural network has two steps: (1) Finding a good low bit-complexity representation for weights (which we call the quantization grid) and (2) Rounding the original weights to values in the quantization grid. In this paper, we study the problem of rounding optimally given any quantization grid. The simplest and most commonly used way to round is Round-to-Nearest (RTN). By rounding in a data-dependent way instead, one can improve the quality of the quantized model significantly. We study the rounding problem from the lens of \emph{discrepancy theory}, which studies how well we can round a continuous solution to a discrete solution without affecting solution quality too much. We prove that given $m=\mathrm{poly}(1/\epsilon)$ samples from the data distribution, we can round all but $O(m)$ model weights such that the expected approximation error of the quantized model on the true data distribution is $\le \epsilon$ as long as the space of gradients of the original model is approximately low rank (which we empirically validate). Our proof, which is algorithmic, inspired a simple and practical rounding algorithm called \emph{DiscQuant}. In our experiments, we demonstrate that DiscQuant significantly improves over the prior state-of-the-art rounding method called GPTQ and the baseline RTN over a range of benchmarks on Phi3mini-3.8B and Llama3.1-8B. For example, rounding Phi3mini-3.8B to a fixed quantization grid with 3.25 bits per parameter using DiscQuant gets 64\% accuracy on the GSM8k dataset, whereas GPTQ achieves 54\% and RTN achieves 31\% (the original model achieves 84\%). We make our code available at https://github.com/jerry-chee/DiscQuant.

In this paper, we present an enhanced Convolutional Neural Network (CNN)-based rooftop solar photovoltaic (PV) panel detection approach using satellite images. We propose to use pre-trained CNN-based model to extract the local convolutional features of rooftops. These local features are then combined using the Vectors of Locally Aggregated Descriptors (VLAD) technique to obtain rooftop-level global features, which are then used to train traditional Machine Learning (ML) models to identify rooftop images that do and do not contain PV panels. On the dataset used in this study, the proposed approach achieved rooftop-PV classification scores exceeding the predefined threshold of 0.9 across all three cities for each of the feature extractor networks evaluated. Moreover, we propose a 3-phase approach to enable efficient utilization of the previously trained models on a new city or region with limited labelled data. We illustrate the effectiveness of this 3-phase approach for multi-city rooftop-PV detection task., Comment: Accepted at CODS-COMAD 2024, December, 2024, Jodhpur, India (https://cods-comad.in/accepted-papers.php)

We present multi-wavelength analysis of ZTF23abelseb (AT 2023sva), an optically discovered fast-fading ($\Delta m_r = 2.2$ mag in $\Delta t = 0.74 $ days), luminous ($M_r \sim -30.0$ mag) and red ($g-r = 0.50$ mag) transient at $z = 2.28$ with accompanying luminous radio emission. AT 2023sva does not possess a $\gamma$-ray burst (GRB) counterpart to an isotropic equivalent energy limit of $E_{\rm{\gamma, \, iso}} < 1.6 \times 10^{52}$ erg, determined through searching $\gamma$-ray satellite archives between the last non-detection and first detection, making it the sixth example of an optically-discovered afterglow with a redshift measurement and no detected GRB counterpart. We analyze AT 2023sva's optical, radio, and X-ray observations to characterize the source. From radio analyses, we find the clear presence of strong interstellar scintillation (ISS) 72 days after the initial explosion, allowing us to place constraints on the source's angular size and bulk Lorentz factor. When comparing the source sizes derived from ISS of orphan events to those of the classical GRB population, we find orphan events have statistically smaller source sizes. We also utilize Bayesian techniques to model the multi-wavelength afterglow. Within this framework, we find evidence that AT 2023sva possesses a shallow power-law structured jet viewed slightly off-axis ($\theta_{\rm{v}} = 0.07 \pm 0.02$) just outside of the jet's core opening angle ($\theta_{\rm{c}} = 0.06 \pm 0.02$). We determine this is likely the reason for the lack of a detected GRB counterpart, but also investigate other scenarios. AT 2023sva's evidence for possessing a structured jet stresses the importance of broadening orphan afterglow search strategies to a diverse range of GRB jet angular energy profiles, to maximize the return of future optical surveys., Comment: 22 pages, 14 Figures, Submitted to MNRAS

The Solar Ultraviolet Imaging Telescope (SUIT) is an instrument on the Aditya-L1 mission of the Indian Space Research Organization (ISRO) launched on September 02, 2023. SUIT continuously provides, near-simultaneous full-disk and region-of-interest images of the Sun, slicing through the photosphere and chromosphere and covering a field of view up to 1.5 solar radii. For this purpose, SUIT uses 11 filters tuned at different wavelengths in the 200{--}400~nm range, including the Mg~{\sc ii} h~and~k and Ca~{\sc ii}~H spectral lines. The observations made by SUIT help us understand the magnetic coupling of the lower and middle solar atmosphere. In addition, for the first time, it allows the measurements of spatially resolved solar broad-band radiation in the near and mid ultraviolet, which will help constrain the variability of the solar ultraviolet irradiance in a wavelength range that is central for the chemistry of the Earth's atmosphere. This paper discusses the details of the instrument and data products., Comment: 37 pages, Accepted for Publication in Solar Physics

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

White dwarfs (WDs) are the most abundant compact objects, and recent surveys have suggested that over a third of WDs in accreting binaries host a strong (B $\gtrsim$ 1 MG) magnetic field. However, the origin and evolution of WD magnetism remain under debate. Two WD pulsars, AR Sco and J191213.72-441045.1 (J1912), have been found, which are non-accreting binaries hosting rapidly spinning (1.97-min and 5.30-min, respectively) magnetic WDs. The WD in AR Sco is slowing down on a $P/\dot{P}\approx 5.6\times 10^6$ yr timescale. It is believed they will eventually become polars, accreting systems in which a magnetic WD (B $\approx 10-240$ MG) accretes from a Roche lobe-filling donor spinning in sync with the orbit ($\gtrsim 78$ min). Here, we present multiwavelength data and analysis of Gaia22ayj, which outbursted in March 2022. We find that Gaia22ayj is a magnetic accreting WD that is rapidly spinning down ($P/\dot{P} = 6.1^{+0.3}_{-0.2}\times 10^6$ yr) like WD pulsars, but shows clear evidence of accretion, like polars. Strong linear polarization (40%) is detected in Gaia22ayj; such high levels have only been seen in the WD pulsar AR Sco and demonstrate the WD is magnetic. High speed photometry reveals a 9.36-min period accompanying a high amplitude ($\sim 2$ mag) modulation. We associate this with a WD spin or spin-orbit beat period, not an orbital period as was previously suggested. Fast (60-s) optical spectroscopy reveals a broad ``hump'', reminiscent of cyclotron emission in polars, between 4000-8000 Angstrom. We find an X-ray luminosity of $L_X = 2.7_{-0.8}^{+6.2}\times10^{32} \textrm{ erg s}^{-1}$ in the 0.3-8 keV energy range, while two VLA radio campaigns resulted in a non-detection with a $F_r < 15.8\mu\textrm{Jy}$ 3$ \sigma$ upper limit. The shared properties of both WD pulsars and polars suggest that Gaia22ayj is a missing link between the two classes of magnetic WD binaries., Comment: Submitted to PASP; comments welcome

In reinforcement learning, especially in sparse-reward domains, many environment steps are required to observe reward information. In order to increase the frequency of such observations, "potential-based reward shaping" (PBRS) has been proposed as a method of providing a more dense reward signal while leaving the optimal policy invariant. However, the required "potential function" must be carefully designed with task-dependent knowledge to not deter training performance. In this work, we propose a "bootstrapped" method of reward shaping, termed BSRS, in which the agent's current estimate of the state-value function acts as the potential function for PBRS. We provide convergence proofs for the tabular setting, give insights into training dynamics for deep RL, and show that the proposed method improves training speed in the Atari suite., Comment: Accepted at AAAI-2025, Main Track

We show that bi-frequency solitary waves are generically present in fermionic systems with scalar self-interaction, such as the Dirac--Klein--Gordon system and the Soler model. We develop the approach to stability properties of such waves and use the radial reduction to show that indeed the (linear) stability is available. We conjecture that stable bi-frequency modes serve as storages of the Dark Matter., Comment: 5 pages

For the nonlinear Dirac equation with scalar self-interaction (the Soler model) in three spatial dimensions, we consider the linearization at solitary wave solutions and find the invariant spaces which correspond to different spherical harmonics, thus achieving the radial reduction of the spectral stability analysis. We apply the same technique to the bi-frequency solitary waves (which are generically present in the Soler model) and show that they can also possess linear stability properties similar to those of one-frequency solitary waves., Comment: 23 pages. This work represents supplementary material for an analytical part of the article "Stable bi-frequency spinor modes as Dark Matter candidates"

We study how spatiotemporal chaos in dynamical systems can be controlled by stochastically returning them to their initial conditions. Focusing on discrete nonlinear maps, we analyze how key measures of chaos -- the Lyapunov exponent and butterfly velocity, which quantify sensitivity to initial perturbations and the ballistic spread of information, respectively -- are reduced by stochastic resetting. We identify a critical resetting rate that induces a dynamical phase transition, characterized by the simultaneous vanishing of the Lyapunov exponent and butterfly velocity, effectively arresting the spread of information. These theoretical predictions are validated and illustrated with numerical simulations of the celebrated logistic map and its lattice extension. Beyond discrete maps, our findings offer insights applicable to a broad class of extended classical interacting systems., Comment: (4+epsilon) pages + 7 pages of supplementary material

We study the asymptotic symmetries of near-horizon extremal BTZ black holes in higher derivative theories of gravity, such as New Massive Gravity and Topological Massive Gravity. By employing a particular boundary condition and the regularization prescription proposed earlier for the Einstein gravity, we demonstrate the existence of two centrally extended Virasoro algebras. The central charges evaluated within this framework are in agreement with their corresponding expressions evaluated at the spatial infinity. We also discuss the robustness of the regularization procedure by relating asymptotic and near-horizon geometries., Comment: 1+23 pages, references added

We consider the optimal appointment scheduling problem that incorporates patients' unpunctual behavior, where the unpunctuality is assumed to be time dependent, but additive. Our goal is to develop an optimal scheduling method for a large patient system to maximize expected net revenue. Methods for deriving optimal appointment schedules for large-scale systems often run into computational bottlenecks due to mixed-integer programming or robust optimization formulations and computationally complex search methods. In this work, we model the system as a single-server queueing system, where patients arrive unpunctually and follow the FIFO service discipline to see the doctor (i.e., get into service). Using the heavy traffic fluid approximation, we develop a deterministic control problem, referred to as the fluid control problem (FCP), which serves as an asymptotic upper bound for the original queueing control problem (QCP). Using the optimal solution of the FCP, we establish an asymptotically optimal scheduling policy on a fluid scale. We further investigate the convergence rate of the QCP under the proposed policy. The FCP, due to the incorporation of unpunctuality, is difficult to solve analytically. We thus propose a time-discretized numerical scheme to approximately solve the FCP. The discretized FCP takes the form of a quadratic program with linear constraints. We examine the behavior of these schedules under different unpunctuality assumptions and test the performance of the schedules on real data in a simulation study. Interestingly, the optimal schedules can involve block booking of patients, even if the unpunctuality distributions are continuous.

Interaction of electric fields with biological cells is indispensable for many physiological processes. Thermal electrical noise in the cellular environment has long been considered as the minimum threshold for detection of electrical signals by cells. However, there is compelling experimental evidence that the minimum electric field sensed by certain cells and organisms is many orders of magnitude weaker than the thermal electrical noise limit estimated purely under equilibrium considerations. We resolve this discrepancy by proposing a non-equilibrium statistical mechanics model for active electromechanical membranes and hypothesize the role of activity in modulating the minimum electrical field that can be detected by a biological membrane. Active membranes contain proteins that use external energy sources to carry out specific functions and drive the membrane away from equilibrium. The central idea behind our model is that active mechanisms, attributed to different sources, endow the membrane with the ability to sense and respond to electric fields that are deemed undetectable based on equilibrium statistical mechanics. Our model for active membranes is capable of reproducing different experimental data available in the literature by varying the activity. Elucidating how active matter can modulate the sensitivity of cells to electric signals can open avenues for a deeper understanding of physiological and pathological processes.

The discovery of novel mechanical metamaterials, whose properties are dominated by their engineered structures rather than chemical composition, is a knowledge-intensive and resource-demanding process. To accelerate the design of novel metamaterials, we present MetaScientist, a human-in-the-loop system that integrates advanced AI capabilities with expert oversight with two primary phases: (1) hypothesis generation, where the system performs complex reasoning to generate novel and scientifically sound hypotheses, supported with domain-specific foundation models and inductive biases retrieved from existing literature; (2) 3D structure synthesis, where a 3D structure is synthesized with a novel 3D diffusion model based on the textual hypothesis and refined it with a LLM-based refinement model to achieve better structure properties. At each phase, domain experts iteratively validate the system outputs, and provide feedback and supplementary materials to ensure the alignment of the outputs with scientific principles and human preferences. Through extensive evaluation from human scientists, MetaScientist is able to deliver novel and valid mechanical metamaterial designs that have the potential to be highly impactful in the metamaterial field.

With recent advancements in industrial robots, educating students in new technologies and preparing them for the future is imperative. However, access to industrial robots for teaching poses challenges, such as the high cost of acquiring these robots, the safety of the operator and the robot, and complicated training material. This paper proposes two low-cost platforms built using open-source tools like Robot Operating System (ROS) and its latest version ROS 2 to help students learn and test algorithms on remotely connected industrial robots. Universal Robotics (UR5) arm and a custom mobile rover were deployed in different life-size testbeds, a greenhouse, and a warehouse to create an Autonomous Agricultural Harvester System (AAHS) and an Autonomous Warehouse Management System (AWMS). These platforms were deployed for a period of 7 months and were tested for their efficacy with 1,433 and 1,312 students, respectively. The hardware used in AAHS and AWMS was controlled remotely for 160 and 355 hours, respectively, by students over a period of 3 months., Comment: 14 pages. Accepted at Springer's 16th International Conference on Social Robotics + AI 2024

We present the discovery of 14 new (and recovery of 4 known) low accretion rate magnetic white dwarfs in post-common envelope binaries that emit strong cyclotron emission using the Zwicky Transient Facility (ZTF) light curves, doubling the known sample size. In addition, we discovered a candidate magnetic period bouncer and recovered three known ones. We confirmed the presence of cyclotron emission using low-resolution spectra in 19 objects. Using the ZTF light curves, follow-up spectra, and the spectral energy distribution, we measured the orbital period, magnetic field strength, and white dwarf temperature of each system. Although the phase-folded light curves have diverse shapes and show a much larger variability amplitude, we show that their intrinsic properties (e.g. period distribution, magnetic field strength) are similar to those of previously known systems. The diversity in light curve shapes can be explained by differences in the optical depth of the accretion spot and geometric differences, the inclination angle and the magnetic spot latitude. The evolutionary states of the longer period binaries are somewhat uncertain but are vary; we found systems consistent with being pre-polars, detached polars, or low-state polars. In addition, we discovered two new low-state polars that likely have brown dwarf companions and could be magnetic period bouncers.

Spray drop size distribution generated by atomization of fuel influences several facets of a combustion process such as, fuel-air mixing, reaction kinetics and thrust generation. In a typical spray, the drop size distribution evolves spatially, varying significantly between the near and far regions of the spray. Studies so far have focused on either one of these regions and are unclear on the exact axial location of transition. In this work, we address this crucial gap by considering a swirl atomizer and measuring the droplet characteristics for different liquid flow conditions of the ensuing spray at various radial and axial locations. Our results reveal an axial variation in the scaled radial droplet velocity profiles, not followed by the radial drop size profiles, from which we demarcate the near region as the zone which extends to 2.0 to 2.5 times film breakup length. Beyond this distance, the drop size characteristics are influenced by external factors such as airflow and identified as the far region. Further, we locate the point of origin of the droplet high-velocity stream along the spray centreline to the end of film breakup of the spray. We also find that the global probability density functions for droplet size and velocity which show a bimodal behavior in the near-region and unimodal in the far-region being well represented by the double Gaussian and Gamma distributions, respectively. We further quantify our results by number and volume flux distributions, global mean drop sizes, drop size ($D_d$) axial velocity ($U_a$) correlations, axial velocity based on drop size classification and turbulent kinetic energy (TKE) to reveal the effect of drop inertia and air flow in determining the statistics in both the near and far regions. We anticipate the findings of this work will guide future investigations on combustion processes and combustor design based on spray characteristics.

We investigate the efficacy of a systematic planetary nebula (PN) search in the Census of the Local Universe (CLU) narrowband (H$\alpha$) survey that covers a considerably larger sky region of above declination $-20^\circ$ than most previous surveys. Using PNe observed by the Isaac Newton Telescope Photometric H$\alpha$ Survey (IPHAS) as validation, we are able to visually recover 432 out of 441 cataloged PNe (98\%) within the CLU dataset, with 5 sources having unusable CLU images and 4 missed due to limitations of imaging quality. Moreover, the reference PNe are conventionally divided into three PN classes in decreasing order of identification confidence given their spectra and morphologies. We record consistently high recovery rate across all classes: 95\% of True, 71\% of Likely, and 81\% of Possible sources are readily recovered. To further demonstrate the ability of CLU to find new PNe, we undertake a preliminary search of compact PNe within a sub-region of the validation catalog, mainly utilizing the significance of narrow-band colors ($\Sigma$) as a metric for identification. In a $200\,\rm deg^2$ region, we search the CLU source catalog and find 31 PN candidates after automated and visual scrutiny, of which 12 are new sources not appearing in previous studies. As a demonstration of our ongoing follow-up campaign, we present medium-resolution optical spectra of six candidates and notice that four of them show emission signatures characteristic of confirmed PNe. As we refine our selection methods, CLU promises to provide a systematic catalog of PNe spanning $2/3$ of the sky., Comment: 40 pages, 12 figures, 3 tables; submitted to the Publications of the Astronomical Society of the Pacific

Supernovae (SNe) come in various flavors and are classified into different types based on emission and absorption lines in their spectra. SN candidates are now abundant with the advent of large systematic dynamic sky surveys like the Zwicky Transient Facility (ZTF), however, the identification bottleneck lies in their spectroscopic confirmation and classification. Fully robotic telescopes with dedicated spectrographs optimized for SN follow-up have eased the burden of data acquisition. However, the task of classifying the spectra still largely rests with the astronomers. Automating this classification step reduces human effort, and can make the SN type available sooner to the public. For this purpose, we have developed a deep-learning based program for classifying core-collapse supernovae (CCSNe) with ultra-low resolution spectra obtained with the SED-Machine IFU spectrograph on the Palomar 60-inch telescope. The program consists of hierarchical classification task layers, with each layer composed of multiple binary classifiers which are run in parallel to produce a reliable classification. The binary classifiers utilize RNN and CNN architecture and are designed to take multiple inputs, to supplement spectra with g- and r-band photometry from ZTF. On non-host-contaminated and good quality SEDM spectra ("gold" test set), CCSNscore is ~94% accurate (correct classifications) in distinguishing between hydrogen-rich (Type II) and hydrogen-poor (Type Ibc) CCSNe. With the help of light curve input, CCSNscore classifies ~83% of the gold set with high confidence (score >= 0.8 and score-error < 0.05), with ~98% accuracy., Comment: Submitted to PASP

The 21-cm forest signal is a promising probe of the Epoch of Reionization complementary to other 21-cm line observables and Ly$\alpha$ forest signal. Prospects of detecting it have significantly improved in the last decade thanks to the discovery of more than 30 radio-loud quasars at these redshifts, upgrades to telescope facilities, and the notion that neutral hydrogen islands persist down to $z\lesssim 5.5$. We forward-model the 21-cm forest signal using semi-numerical simulations and incorporate various instrumental features to explore the potential of detecting the 21-cm forest at $z=6$, both directly and statistically, with the currently available (uGMRT) and forthcoming (SKA1-low) observatories. We show that it is possible to detect the 1D power spectrum of the 21-cm forest spectrum, especially at large scales of $k\lesssim8.5\,\rm MHz^{-1}$ with the $500\,\rm hr$ of the uGMRT time and $k\lesssim32.4\,\rm MHz^{-1}$ with the SKA1-low over $50\,\rm hr$ if the intergalactic medium (IGM) is $25\%$ neutral and these neutral hydrogen regions have a spin temperature of $\lesssim30\,\rm K$. On the other hand, we infer that a null-detection of the signal with such observations of 10 radio-loud sources at $z\approx6$ can be translated into constraints on the thermal and ionization state of the IGM which are tighter than the currently available measurements. Moreover, a null-detection of the 1D 21-cm forest power spectrum with only $50\,\rm hr$ of the uGMRT observations of 10 radio-loud sources can already be competitive with the Ly$\alpha$ forest and 21-cm tomographic observations in disfavouring models of significantly neutral and cold IGM at $z=6$., Comment: 15 pages, 16 figures. Accepted for publiction in MNRAS

Antiferromagnets are emerging as promising alternatives to ferromagnets in spintronics applications. A key feature of antiferromagnets is their anisotropic magnetoresistance (AMR), which has the potential to serve as a sensitive marker for the antiferromagnetic order parameter. However, the underlying origins of this behavior remain poorly understood. In this study, we report the observation of AMR in epitaxial thin films of the collinear L1$_{0}$ antiferromagnet MnPt$_{x}$Pd$_{1-x}$. In thicker films, the AMR is dominated by a non-crystalline two-fold component. As the film thickness is reduced, however, a crystalline four-fold component emerges, accompanied by the appearance of uncompensated magnetic moment, which strongly modifies the magnetotransport properties in the thinner films. We demonstrate that interfacial interactions lead to a large density of states (DOS) at the Fermi level. This enhanced DOS, combined with disorder in thinner films, stabilizes the uncompensated moment and induces a four-fold modulation of the DOS as the Neel vector rotates, explaining the observed AMR behavior., Comment: 8 pages, 6 figures

In recent years, the advances in digitalisation have also adversely contributed to the significant rise in cybercrimes. Hence, building the threat intelligence to shield against rising cybercrimes has become a fundamental requisite. Internet Protocol (IP) addresses play a crucial role in the threat intelligence and prevention of cyber crimes. However, we have noticed the lack of one-stop, free, and open-source tools that can analyse IP addresses. Hence, this work introduces a comprehensive web tool for advanced IP address characterisation. Our tool offers a wide range of features, including geolocation, blocklist check, VPN detection, proxy detection, bot detection, Tor detection, port scan, and accurate domain statistics that include the details about the name servers and registrar information. In addition, our tool calculates a confidence score based on a weighted sum of publicly accessible online results from different reliable sources to give users a dependable measure of accuracy. Further, to improve performance, our tool also incorporates a local database for caching the results, to enable fast content retrieval with minimal external Web API calls. Our tool supports domain names and IPv4 addresses, making it a multi-functional and powerful IP analyser tool for threat intelligence. Our tool is available at www.ipanalyzer.in, Comment: Presented at ICIE 2024

Quality degradation is observed in underwater images due to the effects of light refraction and absorption by water, leading to issues like color cast, haziness, and limited visibility. This degradation negatively affects the performance of autonomous underwater vehicles used in marine applications. To address these challenges, we propose a lightweight phase-based transformer network with 1.77M parameters for underwater image restoration (UIR). Our approach focuses on effectively extracting non-contaminated features using a phase-based self-attention mechanism. We also introduce an optimized phase attention block to restore structural information by propagating prominent attentive features from the input. We evaluate our method on both synthetic (UIEB, UFO-120) and real-world (UIEB, U45, UCCS, SQUID) underwater image datasets. Additionally, we demonstrate its effectiveness for low-light image enhancement using the LOL dataset. Through extensive ablation studies and comparative analysis, it is clear that the proposed approach outperforms existing state-of-the-art (SOTA) methods., Comment: 8 pages, 8 figures, conference

In an accompanying publication, the MeerKAT Pulsar Timing Array (MPTA) collaboration reports tentative evidence for the presence of a stochastic gravitational-wave background, following observations of similar signals from the European and Indian Pulsar Timing Arrays, NANOGrav, the Parkes Pulsar Timing Array and the Chinese Pulsar Timing Array. If such a gravitational-wave background signal originates from a population of inspiraling supermassive black-hole binaries, the signal may be anisotropically distributed on the sky. In this Letter we evaluate the anisotropy of the MPTA signal using a spherical harmonic decomposition. We discuss complications arising from the covariance between pulsar pairs and regularisation of the Fisher matrix. Applying our method to the 4.5 yr dataset, we obtain two forms of sky maps for the three most sensitive MPTA frequency bins between 7 -21 nHz. Our "clean maps'' estimate the distribution of gravitational-wave strain power with minimal assumptions. Our radiometer maps answer the question: is there a statistically significant point source? We find a noteworthy hotspot in the 7 nHz clean map with a $p$-factor of $p=0.015$ (not including trial factors). Future observations are required to determine if this hotspot is of astrophysical origin.

Pulsar Timing Arrays search for nanohertz-frequency gravitational waves by regularly observing ensembles of millisecond pulsars over many years to look for correlated timing residuals. Recently the first evidence for a stochastic gravitational wave background has been presented by the major Arrays, with varying levels of significance ($\sim$2-4$\sigma$). In this paper we present the results of background searches with the MeerKAT Pulsar Timing Array. Although of limited duration (4.5 yr), the $\sim$ 250,000 arrival times with a median error of just $3 \mu$s on 83 pulsars make it very sensitive to spatial correlations. Detection of a gravitational wave background requires careful modelling of noise processes to ensure that any correlations represent a fit to the underlying background and not other misspecified processes. Under different assumptions about noise processes we can produce either what appear to be compelling Hellings-Downs correlations of high significance (3-3.4$\sigma$) with a spectrum close to that which is predicted, or surprisingly, under slightly different assumptions, ones that are insignificant. This appears to be related to the fact that many of the highest precision MeerKAT Pulsar Timing Array pulsars are in close proximity and dominate the detection statistics. The sky-averaged characteristic strain amplitude of the correlated signal in our most significant model is $h_{c, {\rm yr}} = 7.5^{+0.8}_{-0.9} \times 10^{-15}$ measured at a spectral index of $\alpha=-0.26$, decreasing to $h_{c, {\rm yr}} = 4.8^{+0.8}_{-0.9} \times 10^{-15}$ when assessed at the predicted $\alpha=-2/3$. These data will be valuable as the International Pulsar Timing Array project explores the significance of gravitational wave detections and their dependence on the assumed noise models.

Pulsar timing arrays are ensembles of regularly observed millisecond pulsars timed to high precision. Each pulsar in an array could be affected by a suite of noise processes, most of which are astrophysically motivated. Analysing them carefully can be used to understand these physical processes. However, the primary purpose of these experiments is to detect signals that are common to all pulsars, in particular signals associated with a stochastic gravitational wave background. To detect this, it is paramount to appropriately characterise other signals that may otherwise impact array sensitivity or cause a spurious detection. Here we describe the second data release and first detailed noise analysis of the pulsars in the MeerKAT Pulsar Timing Array, comprising high-cadence and high-precision observations of $83$ millisecond pulsars over $4.5$ years. We use this analysis to search for a common signal in the data, finding a process with an amplitude of $\log_{10}\mathrm{A_{CURN}} = -14.25^{+0.21}_{-0.36}$ and spectral index $\gamma_\mathrm{CURN} = 3.60^{+1.31}_{-0.89}$. Fixing the spectral index at the value predicted for a background produced by the inspiral of binary supermassive black holes, we measure the amplitude to be $\log_{10}\mathrm{A_{CURN}} = -14.28^{+0.21}_{-0.21}$ at a significance expressed as a Bayes factor of $\ln(\mathcal{B}) = 4.46$. Under both assumptions, the amplitude that we recover is larger than those reported by other PTA experiments. We use the results of this analysis to forecast our sensitivity to a gravitational wave background possessing the spectral properties of the common signal we have measured.

Recent JWST observations have allowed for the first time to obtain comprehensive measurements of the ionizing photon production efficiency $\xi_\text{ion} $ for a wide range of reionization-epoch galaxies. We explore implications for the inferred UV luminosity functions and escape fractions of ionizing sources in our suite of simulations. These are run with the GPU-based radiative transfer code ATON-HE and are calibrated to the XQR-30 Lyman-alpha forest data at $550$% at $z> 10$, disfavouring the oligarchic source model at very high redshift. The inferred effective clumping factors in our simulations are in the range of $3-6$, suggesting consistency between the observed ionizing properties of reionization-epoch galaxies and the ionizing photon budget in our simulations., Comment: 6 pages. 3 figures

The demand for sophisticated natural language processing (NLP) methods, particularly Named Entity Recognition (NER), has increased due to the exponential growth of Marathi-language digital content. In particular, NER is essential for recognizing distant entities and for arranging and understanding unstructured Marathi text data. With an emphasis on managing long-range entities, this paper offers a comprehensive analysis of current NER techniques designed for Marathi documents. It dives into current practices and investigates the BERT transformer model's potential for long-range Marathi NER. Along with analyzing the effectiveness of earlier methods, the report draws comparisons between NER in English literature and suggests adaptation strategies for Marathi literature. The paper discusses the difficulties caused by Marathi's particular linguistic traits and contextual subtleties while acknowledging NER's critical role in NLP. To conclude, this project is a major step forward in improving Marathi NER techniques, with potential wider applications across a range of NLP tasks and domains.

We present the MahaSUM dataset, a large-scale collection of diverse news articles in Marathi, designed to facilitate the training and evaluation of models for abstractive summarization tasks in Indic languages. The dataset, containing 25k samples, was created by scraping articles from a wide range of online news sources and manually verifying the abstract summaries. Additionally, we train an IndicBART model, a variant of the BART model tailored for Indic languages, using the MahaSUM dataset. We evaluate the performance of our trained models on the task of abstractive summarization and demonstrate their effectiveness in producing high-quality summaries in Marathi. Our work contributes to the advancement of natural language processing research in Indic languages and provides a valuable resource for future research in this area using state-of-the-art models. The dataset and models are shared publicly at https://github.com/l3cube-pune/MarathiNLP