arXiv subjects

Ankur Nath

Publications and source records attributed to Ankur Nath.

11 recordsLinked to original sources

arXiv2024

Theoretically Grounded Pruning of Large Ground Sets for Constrained, Discrete Optimization

Modern instances of combinatorial optimization problems often exhibit billion-scale ground sets, which have many uninformative or redundant elements. In this work, we develop light-weight pruning algorithms to quickly discard elements that are unlikely to be part of an optimal solution. Under mild assumptions on the instance, we prove theoretical guarantees on the fraction of the optimal value retained and the size of the resulting pruned ground set. Through extensive experiments on real-world datasets for various applications, we demonstrate that our algorithm, QuickPrune, efficiently prunes over 90% of the ground set and outperforms state-of-the-art classical and machine learning heuristics for pruning.

cs.DS↗

arXiv2024

Mamba in Vision: A Comprehensive Survey of Techniques and Applications

Mamba is emerging as a novel approach to overcome the challenges faced by Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs) in computer vision. While CNNs excel at extracting local features, they often struggle to capture long-range dependencies without complex architectural modifications. In contrast, ViTs effectively model global relationships but suffer from high computational costs due to the quadratic complexity of their self-attention mechanisms. Mamba addresses these limitations by leveraging Selective Structured State Space Models to effectively capture long-range dependencies with linear computational complexity. This survey analyzes the unique contributions, computational benefits, and applications of Mamba models while also identifying challenges and potential future research directions. We provide a foundational resource for advancing the understanding and growth of Mamba models in computer vision. An overview of this work is available at https://github.com/maklachur/Mamba-in-Computer-Vision.

cs.CV↗

arXiv2024

On Precision of the Leptonic Mixing Angle $\theta_{23}$ and its Implications for the Flavor Models

Among three leptonic mixing angles, $\theta_{23}$ angle, which characterizes the fractional contribution of two flavor eigenstates $\nu_{\mu}$ and $\nu_{\tau}$ to the third mass eigenstate $\nu_3$, is known to be the largest but the least precisely measured. The work investigates possible reach of $\theta_{23}$ precision with two upcoming gigantic accelerator-based long-baseline neutrino experiments, namely Hyper-Kamiokande and DUNE experiments as well as a possible joint analyses of future neutrino facilities. Our simulation yields that each experiment will definitely establish the octant of $\theta_{23}$ angle for all values within 1$\sigma$ parameter interval, while considering the current limitation. However, if the actual value is $0.48\leq \sin^2\theta_{23}\leq 0.54$, it becomes challenging for these two experiments to reject the maximal ($\theta_{23}=\pi/4$) hypothesis and conclude its octant. This octant-blind region can be further explored with the proposed facilities ESSnuSB and a neutrino factory. Accurate determination of the mixing angle $\theta_{23}$, as well as the accuracy of $\delta_{CP}$, is crucial for examining a certain category of discrete non-Abelian leptonic flavor models. Specifically if CP is conserved in leptonic sector, the combined analysis of Hyper-K and DUNE will rule out the majority of these models. However, if the CP is maximally violated, higher precision of $\delta_{CP}$ is necessary for testing these flavor models.

hep-ph↗

arXiv2024

X-Ray spectral and temporal properties of LMXB 4U 1608-52- observed with AstroSat and NICER

We report results from a detailed study of the neutron star X-ray binary, 4U 1608-52 using observations with {\it AstroSat} (LAXPC/SXT) and {\it NICER} during its 2016 and 2020 outbursts. The 0.7--20.0 keV spectra could be well described with the disk blackbody and thermal Comptonization model. The best-fitting inner disk temperature is $\sim$ 1 keV and radius { $\sim$ 22.17$^{+2.57}_{-2.38}$--27.19$^{+2.03}_{-1.85}$} km and no significant evolution was observed in the disk radius after performing flux and time-resolved spectroscopy. We used a multi-Lorentzian approach to fit the power density spectra and obtained broad-band noise variability. We estimated the energy-dependent fractional root mean square and time-lag of the broad-band noise, and these variations are quantitatively modelled as being due to the coherent variation of the disk emission and the coronal heating rate. Thus, the rapid temporal modeling is consistent with the longer term spectral evolution where the inner disk radius does not vary, and instead the variations can be attributed to accretion rate variations which changes the inner disk temperature and the coronal heating rate.

astro-ph.HE↗

arXiv2024

A Benchmark for Maximum Cut: Towards Standardization of the Evaluation of Learned Heuristics for Combinatorial Optimization

Recently, there has been much work on the design of general heuristics for graph-based, combinatorial optimization problems via the incorporation of Graph Neural Networks (GNNs) to learn distribution-specific solution structures.However, there is a lack of consistency in the evaluation of these heuristics, in terms of the baselines and instances chosen, which makes it difficult to assess the relative performance of the algorithms. In this paper, we propose an open-source benchmark suite MaxCut-Bench dedicated to the NP-hard Maximum Cut problem in both its weighted and unweighted variants, based on a careful selection of instances curated from diverse graph datasets. The suite offers a unified interface to various heuristics, both traditional and machine learning-based. Next, we use the benchmark in an attempt to systematically corroborate or reproduce the results of several, popular learning-based approaches, including S2V-DQN [31], ECO-DQN [4], among others, in terms of three dimensions: objective value, generalization, and scalability. Our empirical results show that several of the learned heuristics fail to outperform a naive greedy algorithm, and that only one of them consistently outperforms Tabu Search, a simple, general heuristic based upon local search. Furthermore, we find that the performance of ECO-DQN remains the same or is improved if the GNN is replaced by a simple linear regression on a subset of the features that are related to Tabu Search. Code, data, and pretrained models are available at: \url{https://github.com/ankurnath/MaxCut-Bench}.

cs.AI↗

arXiv2024

Discretely Beyond $1/e$: Guided Combinatorial Algorithms for Submodular Maximization

For constrained, not necessarily monotone submodular maximization, all known approximation algorithms with ratio greater than $1/e$ require continuous ideas, such as queries to the multilinear extension of a submodular function and its gradient, which are typically expensive to simulate with the original set function. For combinatorial algorithms, the best known approximation ratios for both size and matroid constraint are obtained by a simple randomized greedy algorithm of Buchbinder et al. [9]: $1/e \approx 0.367$ for size constraint and $0.281$ for the matroid constraint in $\mathcal O (kn)$ queries, where $k$ is the rank of the matroid. In this work, we develop the first combinatorial algorithms to break the $1/e$ barrier: we obtain approximation ratio of $0.385$ in $\mathcal O (kn)$ queries to the submodular set function for size constraint, and $0.305$ for a general matroid constraint. These are achieved by guiding the randomized greedy algorithm with a fast local search algorithm. Further, we develop deterministic versions of these algorithms, maintaining the same ratio and asymptotic time complexity. Finally, we develop a deterministic, nearly linear time algorithm with ratio $0.377$.

cs.DS↗

arXiv2023

Unveiling the Limits of Learned Local Search Heuristics: Are You the Mightiest of the Meek?

In recent years, combining neural networks with local search heuristics has become popular in the field of combinatorial optimization. Despite its considerable computational demands, this approach has exhibited promising outcomes with minimal manual engineering. However, we have identified three critical limitations in the empirical evaluation of these integration attempts. Firstly, instances with moderate complexity and weak baselines pose a challenge in accurately evaluating the effectiveness of learning-based approaches. Secondly, the absence of an ablation study makes it difficult to quantify and attribute improvements accurately to the deep learning architecture. Lastly, the generalization of learned heuristics across diverse distributions remains underexplored. In this study, we conduct a comprehensive investigation into these identified limitations. Surprisingly, we demonstrate that a simple learned heuristic based on Tabu Search surpasses state-of-the-art (SOTA) learned heuristics in terms of performance and generalizability. Our findings challenge prevailing assumptions and open up exciting avenues for future research and innovation in combinatorial optimization.

cs.AI↗

arXiv2023

Non-zero $θ_{13}$, CP-violation and Neutrinoless Double Beta Decay for Neutrino Mixing in the $A_4\times Z_2\times Z_3$ Flavor Symmetry Model

We study the modification of the Altarelli-Feruglio $A_4$ flavor symmetry model by adding three singlet flavons $ξ'$, $ξ''$ and $ρ$ and the model is augmented with extra $Z_2\times Z_2^ \prime$ symmetry to prevent the unwanted terms in our study. The addition of these three flavons lead to two higher order corrections in the form of two perturbation parameters $ε$ and $ε^\prime$. These corrections yield the deviation from exact tri-bimaximal (TBM) neutrino mixing pattern by producing a non-zero $θ_{13}$ and other neutrino oscillation parameters which are consistent with the latest experimental data. In both the corrections, the neutrino masses are generated via Weinberg operator. The analysis of the perturbation parameters $ε$ and $ε^\prime$, shows that normal hierarchy (NH) and inverted hierarchy (IH) for $ε$ does not change much. However, as the values of $ε^\prime$ increases, $θ_{23}$ occupies the lower octant for NH case. We further investigate the neutrinoless double beta decay parmeter $m_{ββ}$ using the parameter space of the model for both normal and inverted hierarchies of neutrino masses.

hep-ph↗

arXiv2022

AstroSat observation of rapid Type-I thermonuclear burst from the low mass X-ray binary GX 3+1

We report the results of an observation of low mass X-ray binary GX 3+1 with {\it AstroSat}'s Large Area X-ray Proportional Counter (LAXPC) and Soft X-ray Telescope (SXT) instruments on-board for the first time. We have detected one Type-1 thermonuclear burst ($\sim$ 15 s) present in the LAXPC 20 light curve, with a double peak feature at higher energies and our study of the hardness-intensity diagram reveals that the source was in a soft banana state. The pre-burst emission could be described well by a thermally Comptonised model component. The burst spectra is modelled adopting a time-resolved spectroscopic method using a single color blackbody model added to the pre-burst model, to monitor the parametric changes as the burst decays. Based on our time-resolved spectroscopy, we claim that the detected burst is a photospheric radius expansion (PRE) burst. During the PRE phase, the blackbody flux is found to be approximately constant at an averaged value $\sim$ 2.56 in $10^{-8}$ ergs s$^{-1}$ cm$^{-2}$ units. On the basis of literature survey, we infer that \textit{AstroSat}/LAXPC 20 has detected a burst from GX 3+1 after more than a decade which is also a PRE one. Utilising the burst parameters obtained, we provide a new estimation to the source distance, which is $\sim$ 9.3 $\pm$ 0.4 kpc, calculated for an isotropic burst emission. Finally, we discuss and compare our findings with the published literature reports.

astro-ph.HE↗

arXiv2021

Dual-CyCon Net: A Cycle Consistent Dual-Domain Convolutional Neural Network Framework for Detection of Partial Discharge

In the last decade, researchers have been investigating the severity of insulation breakdown caused by partial discharge (PD) in overhead transmission lines with covered conductors or electrical equipment such as generators and motors used in various industries. Developing an effective partial discharge detection system can lead to significant savings on maintenance and prevent power disruptions. Traditional methods rely on hand-crafted features and domain expertise to identify partial discharge patterns in the electrical current. Many data-driven deep learning-based methods have been proposed in recent years to remove these ad hoc feature extraction. However, most of these methods either operate in the time-domain or frequency-domain. Many research approaches have been developed to generate phase-resolved partial discharge (PRPD) patterns from raw PD sensor data. These PRPD diagrams suggest a correlation between partial discharge activities occurring in an alternating electrical waveform's positive and negative half-cycles. However, this correlation criterion between half-cycles has been remained unexplored in deep learning-based methods. This work proposes a novel feature-fusion-based Dual-CyCon Net that can utilize all time, frequency, and phase domain features for joint learning in one cohesive framework. Our proposed cycle-consistency loss exploits any relation between an alternating electrical signal's positive and negative half-cycles to calibrate the model's sensitivity. This loss explores cycle-invariant PD-specific features, enabling the model to learn more robust, noise-invariant features for PD detection. A case study of our proposed framework on a public real-world noisy measurement from high-frequency voltage sensors to detect damaged power lines has achieved a state-of-the-art MCC score of 0.8455.

cs.LG↗

arXiv2018

Detection techniques and investigation of different neutrino experiments

Neutrino physics is an experimentally driven field. So, we investigate the different detection techniques available in the literature and study the various neutrino oscillation experiments in a chronological manner. Our primary focus is on the construction and detection mechanisms of each experiment. Today, we know a lot about this mysterious ghostly particle by performing different experiments at different times with different neutrino sources viz. solar, atmospheric, reactor, accelerators and high energy astrophysical; and they have contributed in the determination of neutrino parameters. Yet the problems are far from over. We need to determine more precise values of the already known parameters and unravel the completely unknown parameters. Some of the unknowns are absolute masses of neutrino, types of neutrino, mass hierarchy, octant degeneracy and existence of leptonic CP Phase(s). We analyse the neutrino experiments into the past, present and the future (or proposed). We include SNO, Kamiokande, K2K, MINOS, MINOS+, Chooz, NEMO and ICARUS in the past; while Borexino, Double Chooz, Super-K, T2K, IceCube, KamLAND, NO$ν$A, RENO and Daya Bay in the present; and SNO+, Hyper-K, JUNO, RENO-50, INO, DUNE, SuperNEMO, KM3NeT, P2O, LBNO and PINGU in the proposed experiments. We also discuss the necessities of upgrading the present ones to those of the proposed ones thereby summarizing the potentials of the future experiments. We conclude this paper with the current status of the neutrinos.

hep-ph↗