arXiv subjects

Souvik Das

Publications and source records attributed to Souvik Das.

At least 37 records · Page 2Linked to original sources

arXiv2024

A theoretic analysis of magnetoactive GES-based turbulent solar plasma instability

A recently reported gravito-electrostatic sheath (GES) model is procedurally applied to study the turbumagnetoactive helioseismic oscillation features on the entire bi-fluidic solar plasma system. The bounded solar interior plasma (SIP, internally self-gravitating) and the unbounded solar wind plasma (SWP, externally point-gravitating) are coupled through the interfacial diffused solar surface boundary (SSB) due to an exact gravito-electrostatic interplay. A numerical platform on the developed theoretic formalism reveals the evolution of both dispersive and non-dispersive features of the modified GES mode fluctuations in new parametric windows. Different colourspectral profiles exhibit important features of the GES-based SIP-SWP perturbations elaborately. It is illustratively shown that the thermostatistical GES stability depends mainly on the radial distance, magnetic field, equilibrium plasma density, and plasma temperature. We see that their dispersive features are more pertinently pronounced in the self-gravitational domains (SIP) than the electrostatic ones (SWP). Besides, different characteristic parameters with accelerating (or decelerating) and stabilizing (or destabilizing) effects influencing the entire solar plasma stability are illustratively portrayed. We speculate that, in the SIP, the long-wave (gravitational-like) helioseismic fluctuations become highly dispersive showing more propagatory nature than the shorter ones (acoustic-like). The short waves show more propagatory propensity than the longer ones in the SSB and SWP regime. The reliability of our proposed investigation is bolstered along with the tentative applicability and future scope in light of the current solar observational scenarios, such as SOHO, STEREO, SDO, PSP, and SolO.

astro-ph.SR↗

arXiv2024

Two-dimensional Shiba lattices as possible platform for crystalline topological superconductivity

Localized or propagating Majorana boundary modes are the key feature of topological superconductors. They are rare in naturally-occurring compounds, but the tailored manipulation of quantum matter offers opportunities for their realization. Specifically, lattices of Yu-Shiba-Rusinov bound states $-$ Shiba lattices $-$ that arise when magnetic adatoms are placed on the surface of a conventional superconductor can be used to create topological bands within the superconducting gap of the substrate. Here, using scanning tunnelling microscopy to create and probe adatom lattices with single atom precision we reveal two signatures consistent with the realization of two types of mirror symmetry protected topological superconductors. The first has edge modes as well as higher-order corner states, and the second has symmetry-protected bulk nodal points. In principle, their topological character and boundary modes should be protected by the spatial symmetries of the adatom lattice. Our results highlight the potential of Shiba lattices as a platform to design the topology and sample geometry of 2D superconductors.

cond-mat.mes-hall↗

arXiv2024

Improving Dialog Safety using Socially Aware Contrastive Learning

State-of-the-art conversational AI systems raise concerns due to their potential risks of generating unsafe, toxic, unethical, or dangerous content. Previous works have developed datasets to teach conversational agents the appropriate social paradigms to respond effectively to specifically designed hazardous content. However, models trained on these adversarial datasets still struggle to recognize subtle unsafe situations that appear naturally in conversations or introduce an inappropriate response in a casual context. To understand the extent of this problem, we study prosociality in both adversarial and casual dialog contexts and audit the response quality of general-purpose language models in terms of propensity to produce unsafe content. We propose a dual-step fine-tuning process to address these issues using a socially aware n-pair contrastive loss. Subsequently, we train a base model that integrates prosocial behavior by leveraging datasets like Moral Integrity Corpus (MIC) and ProsocialDialog. Experimental results on several dialog datasets demonstrate the effectiveness of our approach in generating socially appropriate responses.

cs.CL↗

arXiv2024

A novel trajectory optimization algorithm for continuous-time model predictive control

This article introduces a numerical algorithm that serves as a preliminary step toward solving continuous-time model predictive control (MPC) problems directly without explicit time-discretization. The chief ingredients of the underlying optimal control problem (OCP) are a linear time-invariant system, quadratic instantaneous and terminal cost functions, and convex path constraints. The thrust of the method involves finitely parameterizing the admissible space of control trajectories and solving the OCP satisfying the given constraints at every time instant in a tractable manner without explicit time-discretization. The ensuing OCP turns out to be a convex semi-infinite program (SIP), and some recently developed results are employed to obtain an optimal solution to this convex SIP. Numerical illustrations on some benchmark models are included to show the efficacy of the algorithm.

math.OC↗

arXiv2023

Yu-Shiba-Rusinov tips: imaging spins at the atomic scale with full magnetic sensitivity

Measurements of magnetic properties at the atomic scale require probes capable of combining high spatial resolution with spin sensitivity. Spin-polarized scanning tunneling microscopy (SP-STM) fulfills these conditions by using atomically sharp magnetic tips. The imaging of spin structures results from the tunneling magneto-conductance that depends on the imbalance in the local density of spin-up and spin-down electrons. Spin-sensitive tips are generally formed from bulk materials or by coating non-magnetic tips with a thin magnetic layer. However, ferromagnetic materials generate stray magnetic fields which can influence the magnetic structure of the probed system, while the magnetization of antiferromagnetic materials is difficult to set tip by externally applied magnetic fields. Here, we use functionalized Yu-Shiba-Rusinov (YSR) tips prepared by attaching magnetic adatoms at the apex of a superconducting cluster to image magnetic interactions at the atomic scale. We demonstrate that YSR tips are capable of sensing different magnetization directions, conferring them full magnetic sensitivity. We additionally show that the finite size of the tip superconducting cluster makes it robust against relatively strong magnetic fields, making YSR tips capable of visualizing magnetic field driven transitions of the spin texture.

cond-mat.supr-con↗

arXiv2023

A discrete-time Pontryagin maximum principle under rate constraints

Limited bandwidth and limited saturation in actuators are practical concerns in control systems. Mathematically, these limitations manifest as constraints being imposed on the control actions, their rates of change, and more generally, the global behavior of their paths. While the problem of actuator saturation has been studied extensively, little attention has been devoted to the problem of actuators having limited bandwidth. While attempts have been made in the direction of incorporating frequency constraints on state-action trajectories before, rate constraints on the control at the design stage have not been studied extensively in the discrete-time regime. This article contributes toward filling this lacuna. In particular, we establish a new discrete-time Pontryagin maximum principle with rate constraints being imposed on the control trajectories, and derive first-order necessary conditions for optimality. A brief discussion on the existence of optimal control is included, and numerical examples are provided to illustrate the results.

math.OC↗

arXiv2023

Track clustering with a quantum annealer for primary vertex reconstruction at hadron colliders

Clustering of charged particle tracks along the beam axis is the first step in reconstructing the positions of hadronic interactions, also known as primary vertices, at hadron collider experiments. We use a 2036 physical qubit D-Wave quantum annealer to perform track clustering in a limited capacity on artificial events where the positions of primary vertices and tracks resemble those measured by the Compact Muon Solenoid experiment at the Large Hadron Collider. The algorithm, which is not a classical-quantum hybrid but relies entirely on quantum annealing, is tested on a variety of event topologies. We demonstrate a deterministic graph-embedding of the problem on the D-Wave Chimera architecture, a method for optimizing the coupling strengths within logical qubits, and a method for optimizing annealing time. Further, we benchmark it against simulated annealing on a commercial CPU constrained to the same processor time per anneal as the physical annealer. We note a quantum advantage against simulated annealing up to a 56 logical qubit problem that involves 665 physical qubits on average. Our embedding and optimization methods, and the benchmarking paradigm, can be applied generally to other clustering problems on quantum annealers. This algorithm may be used as a building-block for more sophisticated algorithms to reach the number of primary vertices at the LHC.

hep-ex↗

arXiv2023

Diving Deep into Modes of Fact Hallucinations in Dialogue Systems

Knowledge Graph(KG) grounded conversations often use large pre-trained models and usually suffer from fact hallucination. Frequently entities with no references in knowledge sources and conversation history are introduced into responses, thus hindering the flow of the conversation -- existing work attempt to overcome this issue by tweaking the training procedure or using a multi-step refining method. However, minimal effort is put into constructing an entity-level hallucination detection system, which would provide fine-grained signals that control fallacious content while generating responses. As a first step to address this issue, we dive deep to identify various modes of hallucination in KG-grounded chatbots through human feedback analysis. Secondly, we propose a series of perturbation strategies to create a synthetic dataset named FADE (FActual Dialogue Hallucination DEtection Dataset). Finally, we conduct comprehensive data analyses and create multiple baseline models for hallucination detection to compare against human-verified data and already established benchmarks.

cs.CL↗

arXiv2022

Acoustic waves in the Jovian dusty magnetosphere: A brief review and meta-analysis

The omnipresence of dust particulates in space and astrophysical plasmas has been attracting numerous researchers to study the collective excitation and propagation dynamics of different eigen-mode structures in diversifed astrocosmic circumstances for years. It includes planetary rings, interplanetary space, cometary tails, asteroid zones, planetary atmospheres, etc. The ubiquitous charged dust particulates possess collective degrees of dynamic freedom resulting in the excitation of relatively low-frequency modes, such as dust-ion-acoustic waves (DIAWs), dust-acoustic waves (DAWs), dust-Coulomb waves (DCWs), and so forth. An interesting prevalency of dusty plasma stability research lies in the Jovian magnetosphere (i.e., Jovian plasmas), embedded inside the supersonic solar wind. A brief review of the updated research works on dust-acoustic waves and related collective instability dynamics in the presence of trapped plasma particles is presented herein. The key aim of the proposed explorative meta-analysis is rooted in outlining concisely the main up-to-date investigations on such collective instability processes chronologically. An especial attention is given primarily to the thermostatistical distribution laws of the constitutive lighter electrons and ions against the heavier positively charged dust grains (microspheres). The trapping mechanism of both the lighter species (electrons+ions) is another additive feature revisited here properly. Finally, we clearly extrapolate a number of futuristic directions in light of sensible novelties with a wider scope, both horizontally as well as vertically.

physics.plasm-ph↗

arXiv2022

Charged Particle Tracking in Real-Time Using a Full-Mesh Data Delivery Architecture and Associative Memory Techniques

We present a flexible and scalable approach to address the challenges of charged particle track reconstruction in real-time event filters (Level-1 triggers) in collider physics experiments. The method described here is based on a full-mesh architecture for data distribution and relies on the Associative Memory approach to implement a pattern recognition algorithm that quickly identifies and organizes hits associated to trajectories of particles originating from particle collisions. We describe a successful implementation of a demonstration system composed of several innovative hardware and algorithmic elements. The implementation of a full-size system relies on the assumption that an Associative Memory device with the sufficient pattern density becomes available in the future, either through a dedicated ASIC or a modern FPGA. We demonstrate excellent performance in terms of track reconstruction efficiency, purity, momentum resolution, and processing time measured with data from a simulated LHC-like tracking detector.

hep-ex↗

arXiv2022

Interaction effects in a 1D flat band at a topological crystalline step edge

Step edges of topological crystalline insulators can be viewed as predecessors of higher-order topology, as they embody one-dimensional edge channels embedded in an effective three-dimensional electronic vacuum emanating from the topological crystalline insulator. Using scanning tunneling microscopy and spectroscopy we investigate the behaviour of such edge channels in Pb$_{1-x}$Sn$_{x}$Se under doping. Once the energy position of the step edge is brought close to the Fermi level, we observe the opening of a correlation gap. The experimental results are rationalized in terms of interaction effects which are enhanced since the electronic density is collapsed to a one-dimensional channel. This constitutes a unique system to study how topology and many-body electronic effects intertwine, which we model theoretically through a Hartree-Fock analysis.

cond-mat.mes-hall↗

arXiv2022

Using Multi-Encoder Fusion Strategies to Improve Personalized Response Selection

Personalized response selection systems are generally grounded on persona. However, there exists a co-relation between persona and empathy, which is not explored well in these systems. Also, faithfulness to the conversation context plunges when a contradictory or an off-topic response is selected. This paper attempts to address these issues by proposing a suite of fusion strategies that capture the interaction between persona, emotion, and entailment information of the utterances. Ablation studies on the Persona-Chat dataset show that incorporating emotion and entailment improves the accuracy of response selection. We combine our fusion strategies and concept-flow encoding to train a BERT-based model which outperforms the previous methods by margins larger than 2.3 % on original personas and 1.9 % on revised personas in terms of hits@1 (top-1 accuracy), achieving a new state-of-the-art performance on the Persona-Chat dataset.

cs.CL↗

arXiv2022

Non-planar magnetoactive GES-based solar plasma stability

A laboratory plasma-wall interaction-based astrophysical gravito-electrostatic sheath (GES) model is methodologically applied to study the dynamic stability of the magnetoactive bi-fluidic solar plasma system in the presence of turbulence effect. The spherically symmetric GES-model formalism couples the solar interior plasma (SIP, internally self-gravitating, bounded) and the solar wind plasma (SWP, externally point-gravitating, unbounded) through the diffused solar surface boundary (SSB). A normal spherical mode ansatz results in a generalized linear quadratic dispersion relation depicting the modal fluctuations on both the SIP and SWP scales. A constructive numerical platform reveals the evolution of both dispersive and non-dispersive modal features of the modified-GES mode excitations. The reliability of the derived non-planar dispersion laws is concretized with the help of an exact analytic shape matching the previously reported results founded on the plane-wave approximation. It is found that the thermo-statistical GES stability depends mainly on the magnetic field, equilibrium plasma density, and plasma temperature. It is speculated that the dispersive features are more pronounced in the self-gravitational domains against the electrostatic ones. The magneto-thermal interplay introduces decelerating (accelerating) and destabilizing (stabilizing) influences on the SIP (SWP), and so forth. At last, we briefly indicate the applicability of the proposed analysis to understand diverse helioseismic activities from the collective plasma dynamical viewpoint in accordance with the recent astronomical observational scenarios reported in the literature.

astro-ph.SR↗

arXiv2021

Proto: A Neural Cocktail for Generating Appealing Conversations

In this paper, we present our Alexa Prize Grand Challenge 4 socialbot: Proto. Leveraging diverse sources of world knowledge, and powered by a suite of neural and rule-based natural language understanding modules, state-of-the-art neural generators, novel state-based deterministic generators, an ensemble of neural re-rankers, a robust post-processing algorithm, and an efficient overall conversation strategy, Proto strives to be able to converse coherently about a diverse range of topics of interest to humans, and provide a memorable experience to the user. In this paper we dissect and analyze the different components and conversation strategies implemented by our socialbot, which enables us to generate colloquial, empathetic, engaging, self-rectifying, factually correct, and on-topic response, which has helped us achieve consistent scores throughout the competition.

cs.CL↗

arXiv2021

Similarity Based Label Smoothing For Dialogue Generation

Generative neural conversational systems are generally trained with the objective of minimizing the entropy loss between the training "hard" targets and the predicted logits. Often, performance gains and improved generalization can be achieved by using regularization techniques like label smoothing, which converts the training "hard" targets to "soft" targets. However, label smoothing enforces a data independent uniform distribution on the incorrect training targets, which leads to an incorrect assumption of equi-probable incorrect targets for each correct target. In this paper we propose and experiment with incorporating data dependent word similarity based weighing methods to transforms the uniform distribution of the incorrect target probabilities in label smoothing, to a more natural distribution based on semantics. We introduce hyperparameters to control the incorrect target distribution, and report significant performance gains over networks trained using standard label smoothing based loss, on two standard open domain dialogue corpora.

cs.CL↗

arXiv2021

Medical Literature Mining and Retrieval in a Conversational Setting

The Covid-19 pandemic has caused a spur in the medical research literature. With new research advances in understanding the virus, there is a need for robust text mining tools which can process, extract and present answers from the literature in a concise and consumable way. With a DialoGPT based multi-turn conversation generation module, and BM-25 \& neural embeddings based ensemble information retrieval module, in this paper we present a conversational system, which can retrieve and answer coronavirus-related queries from the rich medical literature, and present it in a conversational setting with the user. We further perform experiments to compare neural embedding-based document retrieval and the traditional BM25 retrieval algorithm and report the results.

cs.IR↗

arXiv2021

Origin of the quasi-quantized Hall effect in ZrTe5

The quantum Hall effect (QHE) is traditionally considered a purely two-dimensional (2D) phenomenon. Recently, a three-dimensional (3D) version of the QHE has been reported in the Dirac semimetal ZrTe5. It was proposed to arise from a magnetic-field-driven Fermi surface instability, transforming the original 3D electron system into a stack of 2D sheets. Here, we report thermodynamic, thermoelectric and charge transport measurements on ZrTe5 in the quantum Hall regime. The measured thermodynamic properties: magnetization and ultrasound propagation, show no signatures of a Fermi surface instability, consistent with in-field single crystal X-ray diffraction. Instead, a direct comparison of the experimental data with linear response calculations based on an effective 3D Dirac Hamiltonian suggests that the quasi-quantization of the observed Hall response is an intrinsic property of the 3D electronic structure. Our findings render the Hall effect in ZrTe5 a truly 3D counterpart of the QHE in 2D systems.

cond-mat.str-el↗

arXiv2020

A multi-agent evolutionary robotics framework to train spiking neural networks

A novel multi-agent evolutionary robotics (ER) based framework, inspired by competitive evolutionary environments in nature, is demonstrated for training Spiking Neural Networks (SNN). The weights of a population of SNNs along with morphological parameters of bots they control in the ER environment are treated as phenotypes. Rules of the framework select certain bots and their SNNs for reproduction and others for elimination based on their efficacy in capturing food in a competitive environment. While the bots and their SNNs are given no explicit reward to survive or reproduce via any loss function, these drives emerge implicitly as they evolve to hunt food and survive within these rules. Their efficiency in capturing food as a function of generations exhibit the evolutionary signature of punctuated equilibria. Two evolutionary inheritance algorithms on the phenotypes, Mutation and Crossover with Mutation, are demonstrated. Performances of these algorithms are compared using ensembles of 100 experiments for each algorithm. We find that Crossover with Mutation promotes 40% faster learning in the SNN than mere Mutation with a statistically significant margin.

cs.NE↗