Showing 1–49 of 49 results for author: Goel, P

An Empirical Study of Aegis

Authors: Daniel Saragih, Paridhi Goel, Tejas Balaji, Alyssa Li

Abstract: Bit flip** attacks are one class of attacks on neural networks with numerous defense mechanisms invented to mitigate its potency. Due to the importance of ensuring the robustness of these defense mechanisms, we perform an empirical study on the Aegis framework. We evaluate the baseline mechanisms of Aegis on low-entropy data (MNIST), and we evaluate a pre-trained model with the mechanisms fine-t… ▽ More Bit flip** attacks are one class of attacks on neural networks with numerous defense mechanisms invented to mitigate its potency. Due to the importance of ensuring the robustness of these defense mechanisms, we perform an empirical study on the Aegis framework. We evaluate the baseline mechanisms of Aegis on low-entropy data (MNIST), and we evaluate a pre-trained model with the mechanisms fine-tuned on MNIST. We also compare the use of data augmentation to the robustness training of Aegis, and how Aegis performs under other adversarial attacks, such as the generation of adversarial examples. We find that both the dynamic-exit strategy and robustness training of Aegis has some drawbacks. In particular, we see drops in accuracy when testing on perturbed data, and on adversarial examples, as compared to baselines. Moreover, we found that the dynamic exit-strategy loses its uniformity when tested on simpler datasets. The code for this project is available on GitHub. △ Less

Submitted 24 April, 2024; originally announced April 2024.

Comments: 9 pages, 6 figures, 3 tables

Improving the TENOR of Labeling: Re-evaluating Topic Models for Content Analysis

Authors: Zongxia Li, Andrew Mao, Daniel Stephens, Pranav Goel, Emily Walpole, Alden Dima, Juan Fung, Jordan Boyd-Graber

Abstract: Topic models are a popular tool for understanding text collections, but their evaluation has been a point of contention. Automated evaluation metrics such as coherence are often used, however, their validity has been questioned for neural topic models (NTMs) and can overlook a models benefits in real world applications. To this end, we conduct the first evaluation of neural, supervised and classic… ▽ More Topic models are a popular tool for understanding text collections, but their evaluation has been a point of contention. Automated evaluation metrics such as coherence are often used, however, their validity has been questioned for neural topic models (NTMs) and can overlook a models benefits in real world applications. To this end, we conduct the first evaluation of neural, supervised and classical topic models in an interactive task based setting. We combine topic models with a classifier and test their ability to help humans conduct content analysis and document annotation. From simulated, real user and expert pilot studies, the Contextual Neural Topic Model does the best on cluster evaluation metrics and human evaluations; however, LDA is competitive with two other NTMs under our simulated experiment and user study results, contrary to what coherence scores suggest. We show that current automated metrics do not provide a complete picture of topic modeling capabilities, but the right choice of NTMs can be better than classical models on practical task. △ Less

Submitted 19 February, 2024; v1 submitted 29 January, 2024; originally announced January 2024.

Comments: 19 pages, 5 tables, 6 figures, Accepted to EACL Main Conference 2024

Iterative Motion Editing with Natural Language

Authors: Purvi Goel, Kuan-Chieh Wang, C. Karen Liu, Kayvon Fatahalian

Abstract: Text-to-motion diffusion models can generate realistic animations from text prompts, but do not support fine-grained motion editing controls. In this paper, we present a method for using natural language to iteratively specify local edits to existing character animations, a task that is common in most computer animation workflows. Our key idea is to represent a space of motion edits using a set of… ▽ More Text-to-motion diffusion models can generate realistic animations from text prompts, but do not support fine-grained motion editing controls. In this paper, we present a method for using natural language to iteratively specify local edits to existing character animations, a task that is common in most computer animation workflows. Our key idea is to represent a space of motion edits using a set of kinematic motion editing operators (MEOs) whose effects on the source motion is well-aligned with user expectations. We provide an algorithm that leverages pre-existing language models to translate textual descriptions of motion edits into source code for programs that define and execute sequences of MEOs on a source animation. We execute MEOs by first translating them into keyframe constraints, and then use diffusion-based motion models to generate output motions that respect these constraints. Through a user study and quantitative evaluation, we demonstrate that our system can perform motion edits that respect the animator's editing intent, remain faithful to the original animation (it edits the original animation, but does not dramatically change it), and yield realistic character animation results. △ Less

Submitted 3 June, 2024; v1 submitted 15 December, 2023; originally announced December 2023.

Mainstream News Articles Co-Shared with Fake News Buttress Misinformation Narratives

Authors: Pranav Goel, Jon Green, David Lazer, Philip Resnik

Abstract: Most prior and current research examining misinformation spread on social media focuses on reports published by 'fake' news sources. These approaches fail to capture another potential form of misinformation with a much larger audience: factual news from mainstream sources ('real' news) repurposed to promote false or misleading narratives. We operationalize narratives using an existing unsupervised… ▽ More Most prior and current research examining misinformation spread on social media focuses on reports published by 'fake' news sources. These approaches fail to capture another potential form of misinformation with a much larger audience: factual news from mainstream sources ('real' news) repurposed to promote false or misleading narratives. We operationalize narratives using an existing unsupervised NLP technique and examine the narratives present in misinformation content. We find that certain articles from reliable outlets are shared by a disproportionate number of users who also shared fake news on Twitter. We consider these 'real' news articles to be co-shared with fake news. We show that co-shared articles contain existing misinformation narratives at a significantly higher rate than articles from the same reliable outlets that are not co-shared with fake news. This holds true even when articles are chosen following strict criteria of reliability for the outlets and after accounting for the alternative explanation of partisan curation of articles. For example, we observe that a recent article published by The Washington Post titled "Vaccinated people now make up a majority of COVID deaths" was disproportionately shared by Twitter users with a history of sharing anti-vaccine false news reports. Our findings suggest a strategic repurposing of mainstream news by conveyors of misinformation as a way to enhance the reach and persuasiveness of misleading narratives. We also conduct a comprehensive case study to help highlight how such repurposing can happen on Twitter as a consequence of the inclusion of particular narratives in the framing of mainstream news. △ Less

Submitted 11 August, 2023; originally announced August 2023.

M3Act: Learning from Synthetic Human Group Activities

Authors: Che-Jui Chang, Danrui Li, Deep Patel, Parth Goel, Honglu Zhou, Seonghyeon Moon, Samuel S. Sohn, Sejong Yoon, Vladimir Pavlovic, Mubbasir Kapadia

Abstract: The study of complex human interactions and group activities has become a focal point in human-centric computer vision. However, progress in related tasks is often hindered by the challenges of obtaining large-scale labeled datasets from real-world scenarios. To address the limitation, we introduce M3Act, a synthetic data generator for multi-view multi-group multi-person human atomic actions and g… ▽ More The study of complex human interactions and group activities has become a focal point in human-centric computer vision. However, progress in related tasks is often hindered by the challenges of obtaining large-scale labeled datasets from real-world scenarios. To address the limitation, we introduce M3Act, a synthetic data generator for multi-view multi-group multi-person human atomic actions and group activities. Powered by Unity Engine, M3Act features multiple semantic groups, highly diverse and photorealistic images, and a comprehensive set of annotations, which facilitates the learning of human-centered tasks across single-person, multi-person, and multi-group conditions. We demonstrate the advantages of M3Act across three core experiments. The results suggest our synthetic dataset can significantly improve the performance of several downstream methods and replace real-world datasets to reduce cost. Notably, M3Act improves the state-of-the-art MOTRv2 on DanceTrack dataset, leading to a hop on the leaderboard from 10th to 2nd place. Moreover, M3Act opens new research for controllable 3D group activity generation. We define multiple metrics and propose a competitive baseline for the novel task. Our code and data are available at our project page: http://cjerry1243.github.io/M3Act. △ Less

Submitted 2 May, 2024; v1 submitted 29 June, 2023; originally announced June 2023.

Natural Language Decompositions of Implicit Content Enable Better Text Representations

Authors: Alexander Hoyle, Rupak Sarkar, Pranav Goel, Philip Resnik

Abstract: When people interpret text, they rely on inferences that go beyond the observed language itself. Inspired by this observation, we introduce a method for the analysis of text that takes implicitly communicated content explicitly into account. We use a large language model to produce sets of propositions that are inferentially related to the text that has been observed, then validate the plausibilit… ▽ More When people interpret text, they rely on inferences that go beyond the observed language itself. Inspired by this observation, we introduce a method for the analysis of text that takes implicitly communicated content explicitly into account. We use a large language model to produce sets of propositions that are inferentially related to the text that has been observed, then validate the plausibility of the generated content via human judgments. Incorporating these explicit representations of implicit content proves useful in multiple problem settings that involve the human interpretation of utterances: assessing the similarity of arguments, making sense of a body of opinion data, and modeling legislative behavior. Our results suggest that modeling the meanings behind observed language, rather than the literal text alone, is a valuable direction for NLP and particularly its applications to social science. △ Less

Submitted 24 October, 2023; v1 submitted 23 May, 2023; originally announced May 2023.

Comments: Accepted to EMNLP 2023 (Main conference)

Computational Orbital Mechanics of Marble Motion on a 3D Printed Surface -- 1. Formal Basis

Authors: Pooja Bhambhu, Preety, Paridhi Goel, Chinkey, Manisha Siwach, Ananya Kumari, Sudarshana, Sanjana Yadav, Shikha Yadav, Bharti, Poonam, Anshumali, Athira Vijayan, Divakar Pathak

Abstract: Simulating curvature due to gravity through warped surfaces is a common visualization aid in Physics education. We reprise a recent experiment exploring orbital trajectories on a precise 3D-printed surface to mimic Newtonian gravity, and elevate this analogy past the status of a mere visualization tool. We present a general analysis approach through which this straightforward experiment can be use… ▽ More Simulating curvature due to gravity through warped surfaces is a common visualization aid in Physics education. We reprise a recent experiment exploring orbital trajectories on a precise 3D-printed surface to mimic Newtonian gravity, and elevate this analogy past the status of a mere visualization tool. We present a general analysis approach through which this straightforward experiment can be used to create a reasonably advanced computational orbital mechanics lab at the undergraduate level, creating a convenient hands-on, computational pathway to various non-trivial nuances in this discipline, such as the mean, eccentric, and true anomalies and their computation, Laplace-Runge-Lenz vector conservation, characterization of general orbits, and the extraction of orbital parameters. We show that while the motion of a marble on such a surface does not truly represent a orbital trajectory under Newtonian gravity in a strict theoretical sense, but through a proposed projection procedure, the experimentally recorded trajectories closely resemble the Kepler orbits and approximately respect the known conservation laws for orbital motion. The latter fact is demonstrated through multiple experimentally-recorded elliptical trajectories with wide-ranging eccentricities and semi-major axes. In this first part of this two-part sequence, we lay down the formal basis of this exposition, describing the experiment, its calibration, critical assessment of the results, and the computational procedures for the transformation of raw experimental data into a form useful for orbital analysis. △ Less

Submitted 23 February, 2023; originally announced February 2023.

Comments: Article 1 of a two-part sequence. Article 2 (applications) to be released shortly. Comments are most welcome !

Are Neural Topic Models Broken?

Authors: Alexander Hoyle, Pranav Goel, Rupak Sarkar, Philip Resnik

Abstract: Recently, the relationship between automated and human evaluation of topic models has been called into question. Method developers have staked the efficacy of new topic model variants on automated measures, and their failure to approximate human preferences places these models on uncertain ground. Moreover, existing evaluation paradigms are often divorced from real-world use. Motivated by conten… ▽ More Recently, the relationship between automated and human evaluation of topic models has been called into question. Method developers have staked the efficacy of new topic model variants on automated measures, and their failure to approximate human preferences places these models on uncertain ground. Moreover, existing evaluation paradigms are often divorced from real-world use. Motivated by content analysis as a dominant real-world use case for topic modeling, we analyze two related aspects of topic models that affect their effectiveness and trustworthiness in practice for that purpose: the stability of their estimates and the extent to which the model's discovered categories align with human-determined categories in the data. We find that neural topic models fare worse in both respects compared to an established classical method. We take a step toward addressing both issues in tandem by demonstrating that a straightforward ensembling method can reliably outperform the members of the ensemble. △ Less

Submitted 28 October, 2022; originally announced October 2022.

Comments: Accepted to Findings of EMNLP 2022

Soft Anharmonic Coupled Vibrations of Li and SiO4 Enable Li-ion Diffusion in Amorphous Li2Si2O5

Authors: Sajan Kumar, Mayanak K. Gupta, Prabhatasree Goel, Ranjan Mittal, Sanghamitra Mukhopadhyay, Manh Duc Le, Rakesh Shukla, Srungarpu N. Achary, Avesh K. Tyagi, Samrath L. Chaplot

Abstract: We present the investigations on atomic dynamics and Li+ diffusion in crystalline and amorphous Li2Si2O5 using quasielastic (QENS) and inelastic neutron scattering (INS) studies supplemented by ab-initio molecular dynamics simulations (AIMD). The QENS measurements in the amorphous phase of Li2Si2O5 show a narrow temperature window (700 < T < 775 K), exhibiting significant quasielastic broadening c… ▽ More We present the investigations on atomic dynamics and Li+ diffusion in crystalline and amorphous Li2Si2O5 using quasielastic (QENS) and inelastic neutron scattering (INS) studies supplemented by ab-initio molecular dynamics simulations (AIMD). The QENS measurements in the amorphous phase of Li2Si2O5 show a narrow temperature window (700 < T < 775 K), exhibiting significant quasielastic broadening corresponding to the fast Li+ diffusion and relaxation of SiO4 units to the crystalline phase. Our INS measurements clearly show the presence of large phonon density of states (PDOS) at low energy (low-E) in the superionic amorphous phase, which disappear in the non-superionic crystalline phase, corroborating the role of low-E modes in Li+ diffusion. The frustrated energy landscape and host flexibility (due to random orientation and vibrational motion of SiO4 polyhedral units) play an essential role in diffusing the Li+. We used AIMD simulations to identify that these low-E modes involve a large amplitude of Li vibrations coupled with SiO4 vibrations in the amorphous phase. At elevated temperatures, these vibrational dynamics accelerate the Li+ diffusion via a paddle-wheel like coupling mechanism. Above 775 K, these SiO4 vibrational dynamics drive the system into the crystalline phase by locking SiO4 and Li+ into deeper minima of the free energy landscape and disappear in the crystalline phase. Both experiments and simulations provide valuable information about the atomic level stochastic and vibrational dynamics in Li2Si2O5 and their role in Li+ diffusion and vitrification. △ Less

Submitted 17 October, 2022; originally announced October 2022.

Comments: 31 pages, 7 Figures, Supplementary Materials

Radius Constants of Sigmoid Starlike Functions

Authors: Priyanka Goel, S. Sivaprasad Kumar

Abstract: In the present investigation, we study the class of Sigmoid starlike functions, given by $\mathcal{S}^*_{SG}=\{f\in\mathcal{A}: {zf'(z)}/{f(z)}\prec 2/(1+e^{-z})\}$ in context of estimating the sharp radius constants associated with several known subclasses of starlike functions. Further, graphical validation for the sharpness of results is also provided. In the present investigation, we study the class of Sigmoid starlike functions, given by $\mathcal{S}^*_{SG}=\{f\in\mathcal{A}: {zf'(z)}/{f(z)}\prec 2/(1+e^{-z})\}$ in context of estimating the sharp radius constants associated with several known subclasses of starlike functions. Further, graphical validation for the sharpness of results is also provided. △ Less

Submitted 2 August, 2022; originally announced August 2022.

Sufficient conditions and radius problems for the Silverman class

Authors: S. Sivaprasad Kumar, Priyanka Goel

Abstract: For $0<α\leq1$ and $λ>0,$ let \begin{equation}\label{1} G_{λ,α}=\left\{f\in\mathcal{A}:\left|\dfrac{1-α+αzf''(z)/f'(z)}{zf'(z)/f(z)}-(1-α)\right|<λ, z\in\mathbb{D}\right\}, \end{equation} the general form of Silverman class introduced by Tuneski and Irnak. For this class we derive some sufficient conditions in the form of differential inequalities. Further, we consider the class $Ω,$ given by \b… ▽ More For $0<α\leq1$ and $λ>0,$ let \begin{equation}\label{1} G_{λ,α}=\left\{f\in\mathcal{A}:\left|\dfrac{1-α+αzf''(z)/f'(z)}{zf'(z)/f(z)}-(1-α)\right|<λ, z\in\mathbb{D}\right\}, \end{equation} the general form of Silverman class introduced by Tuneski and Irnak. For this class we derive some sufficient conditions in the form of differential inequalities. Further, we consider the class $Ω,$ given by \begin{equation}\label{omega} Ω=\left\{f\in\mathcal{A}:|zf'(z)-f(z)|<\dfrac{1}{2},\;z\in\mathbb{D}\right\}. \end{equation} For the above two classes, we establish inclusion relations involving some other well known subclasses of $\mathcal{S}^*$ and find radius estimates for different pairs involving these classes. △ Less

Submitted 22 March, 2022; originally announced March 2022.

Application of Pythagorean means and Differential Subordination

Authors: S. Sivaprasad Kumar, Priyanka Goel

Abstract: For $0\leqα\leq 1,$ let $H_α(x,y)$ be the convex weighted harmonic mean of $x$ and $y.$ We establish differential subordination implications of the form \begin{equation*} H_α(p(z),p(z)Θ(z)+zp'(z)Φ(z))\prec h(z)\Rightarrow p(z)\prec h(z), \end{equation*} where $Φ,\;Θ$ are analytic functions and $h$ is a univalent function satisfying some special properties. Further, we prove differential subordin… ▽ More For $0\leqα\leq 1,$ let $H_α(x,y)$ be the convex weighted harmonic mean of $x$ and $y.$ We establish differential subordination implications of the form \begin{equation*} H_α(p(z),p(z)Θ(z)+zp'(z)Φ(z))\prec h(z)\Rightarrow p(z)\prec h(z), \end{equation*} where $Φ,\;Θ$ are analytic functions and $h$ is a univalent function satisfying some special properties. Further, we prove differential subordination implications involving a combination of three classical means. As an application, we generalize many existing results and obtain sufficient conditions for starlikeness and univalence. △ Less

Submitted 22 March, 2022; originally announced March 2022.

Solid-like to Liquid-like Behavior of Cu Diffusion in Superionic Cu2X (X=S, Se): An Inelastic Neutron Scattering and Ab-Initio Molecular Dynamics Investigation

Authors: Sajan Kumar, M. K. Gupta, Prabhatasree Goel, R. Mittal, Olivier Delaire, A. Thamizhavel, S. Rols, S. L. Chaplot

Abstract: Cu2Se and Cu2S are excellent model systems of superionic conductors with large diffusion coefficients that have been reported to exhibit different solid-liquid-like Cu-ion diffusion. In this paper, we clarify the atomic dynamics of these compounds with temperature-dependent ab-initio molecular dynamics (AIMD) simulations and inelastic neutron scattering (INS) experiments. Using the dynamical struc… ▽ More Cu2Se and Cu2S are excellent model systems of superionic conductors with large diffusion coefficients that have been reported to exhibit different solid-liquid-like Cu-ion diffusion. In this paper, we clarify the atomic dynamics of these compounds with temperature-dependent ab-initio molecular dynamics (AIMD) simulations and inelastic neutron scattering (INS) experiments. Using the dynamical structure factor and Van-Hove correlation function, we interrogate the jump-time, hop** length distribution and associated diffusion coefficients. In cubic-Cu2Se at 500 K, we find solid-like diffusion with Cu-jump lengths matching well the first-neighbour Cu-Cu distance of ~3 Å in the crystal, and clearly defined optic phonons involving Cu-vibrations. Above 700 K, the jump-length distribution becomes a broad maximum cantered around 4 Å, spanning the first and second neighbour lattice distances, and a concurrent broadening of the Cu-phonon density of states. Further, above 900 K, the Cu-diffusion becomes close to liquid-like, with distributions of Cu-atoms continuously connecting crystal sites, while the vibrational modes involving Cu motions are highly damped, though still not fully over-damped as in a liquid. At low temperatures, the solid-like diffusion is consistent with previous X-ray diffraction and quasielastic neutron scattering experiments, while the higher-temperature observation of the liquid-like diffusion is in agreement with previous AIMD simulations. We also report AIMD simulations in Cu2S in the hexagonal and cubic superionic phases, and observe similar solid and liquid-like diffusion at low- and high-temperatures, respectively. The calculated ionic-conductivity is in fair agreement with reported experimental values. △ Less

Submitted 3 January, 2022; originally announced January 2022.

Comments: 29 Pages, 7 Figures

Point2Cyl: Reverse Engineering 3D Objects from Point Clouds to Extrusion Cylinders

Authors: Mikaela Angelina Uy, Yen-yu Chang, Minhyuk Sung, Purvi Goel, Joseph Lambourne, Tolga Birdal, Leonidas Guibas

Abstract: We propose Point2Cyl, a supervised network transforming a raw 3D point cloud to a set of extrusion cylinders. Reverse engineering from a raw geometry to a CAD model is an essential task to enable manipulation of the 3D data in shape editing software and thus expand their usages in many downstream applications. Particularly, the form of CAD models having a sequence of extrusion cylinders -- a 2D sk… ▽ More We propose Point2Cyl, a supervised network transforming a raw 3D point cloud to a set of extrusion cylinders. Reverse engineering from a raw geometry to a CAD model is an essential task to enable manipulation of the 3D data in shape editing software and thus expand their usages in many downstream applications. Particularly, the form of CAD models having a sequence of extrusion cylinders -- a 2D sketch plus an extrusion axis and range -- and their boolean combinations is not only widely used in the CAD community/software but also has great expressivity of shapes, compared to having limited types of primitives (e.g., planes, spheres, and cylinders). In this work, we introduce a neural network that solves the extrusion cylinder decomposition problem in a geometry-grounded way by first learning underlying geometric proxies. Precisely, our approach first predicts per-point segmentation, base/barrel labels and normals, then estimates for the underlying extrusion parameters in differentiable and closed-form formulations. Our experiments show that our approach demonstrates the best performance on two recent CAD datasets, Fusion Gallery and DeepCAD, and we further showcase our approach on reverse engineering and editing. △ Less

Submitted 29 May, 2022; v1 submitted 17 December, 2021; originally announced December 2021.

Comments: CVPR 2022

Explanation Container in Case-Based Biomedical Question-Answering

Authors: Prateek Goel, Adam J. Johs, Manil Shrestha, Rosina O. Weber

Abstract: The National Center for Advancing Translational Sciences(NCATS) Biomedical Data Translator (Translator) aims to attenuate problems faced by translational scientists. Translator is a multi-agent architecture consisting of six autonomous relay agents (ARAs) and eight knowledge providers (KPs). In this paper, we present the design of the Explanatory Agent (xARA), a case-based ARA that answers biomedi… ▽ More The National Center for Advancing Translational Sciences(NCATS) Biomedical Data Translator (Translator) aims to attenuate problems faced by translational scientists. Translator is a multi-agent architecture consisting of six autonomous relay agents (ARAs) and eight knowledge providers (KPs). In this paper, we present the design of the Explanatory Agent (xARA), a case-based ARA that answers biomedical queries by accessing multiple KPs, ranking results, and explaining the ranking of results. The Explanatory Agent is designed with five knowledge containers that include the four original knowledge containers and one additional container for explanation - the Explanation Container. The Explanation Container is case-based and designed with its own knowledge containers. △ Less

Submitted 22 December, 2021; v1 submitted 13 December, 2021; originally announced December 2021.

Comments: Incomplete acknowledgments. Paper to be withdrawn until further notice

Dynamic Mirror Descent based Model Predictive Control for Accelerating Robot Learning

Authors: Utkarsh A. Mishra, Soumya R. Samineni, Prakhar Goel, Chandravaran Kunjeti, Himanshu Lodha, Aman Singh, Aditya Sagi, Shalabh Bhatnagar, Shishir Kolathaya

Abstract: Recent works in Reinforcement Learning (RL) combine model-free (Mf)-RL algorithms with model-based (Mb)-RL approaches to get the best from both: asymptotic performance of Mf-RL and high sample-efficiency of Mb-RL. Inspired by these works, we propose a hierarchical framework that integrates online learning for the Mb-trajectory optimization with off-policy methods for the Mf-RL. In particular, two… ▽ More Recent works in Reinforcement Learning (RL) combine model-free (Mf)-RL algorithms with model-based (Mb)-RL approaches to get the best from both: asymptotic performance of Mf-RL and high sample-efficiency of Mb-RL. Inspired by these works, we propose a hierarchical framework that integrates online learning for the Mb-trajectory optimization with off-policy methods for the Mf-RL. In particular, two loops are proposed, where the Dynamic Mirror Descent based Model Predictive Control (DMD-MPC) is used as the inner loop Mb-RL to obtain an optimal sequence of actions. These actions are in turn used to significantly accelerate the outer loop Mf-RL. We show that our formulation is generic for a broad class of MPC-based policies and objectives, and includes some of the well-known Mb-Mf approaches. We finally introduce a new algorithm: Mirror-Descent Model Predictive RL (M-DeMoRL), which uses Cross-Entropy Method (CEM) with elite fractions for the inner loop. Our experiments show faster convergence of the proposed hierarchical approach on benchmark MuJoCo tasks. We also demonstrate hardware training for trajectory tracking in a 2R leg and hardware transfer for robust walking in a quadruped. We show that the inner-loop Mb-RL significantly decreases the number of training iterations required in the real system, thereby validating the proposed approach. △ Less

Submitted 4 November, 2021; originally announced December 2021.

Comments: 8 pages, 4 figures. arXiv admin note: substantial text overlap with arXiv:2110.12239

Studying word order through iterative shuffling

Authors: Nikolay Malkin, Sameera Lanka, Pranav Goel, Nebojsa Jojic

Abstract: As neural language models approach human performance on NLP benchmark tasks, their advances are widely seen as evidence of an increasingly complex understanding of syntax. This view rests upon a hypothesis that has not yet been empirically tested: that word order encodes meaning essential to performing these tasks. We refute this hypothesis in many cases: in the GLUE suite and in various genres of… ▽ More As neural language models approach human performance on NLP benchmark tasks, their advances are widely seen as evidence of an increasingly complex understanding of syntax. This view rests upon a hypothesis that has not yet been empirically tested: that word order encodes meaning essential to performing these tasks. We refute this hypothesis in many cases: in the GLUE suite and in various genres of English text, the words in a sentence or phrase can rarely be permuted to form a phrase carrying substantially different information. Our surprising result relies on inference by iterative shuffling (IBIS), a novel, efficient procedure that finds the ordering of a bag of words having the highest likelihood under a fixed language model. IBIS can use any black-box model without additional training and is superior to existing word ordering algorithms. Coalescing our findings, we discuss how shuffling inference procedures such as IBIS can benefit language modeling and constrained generation. △ Less

Submitted 10 September, 2021; originally announced September 2021.

Comments: EMNLP 2021

Longitudinal Distance: Towards Accountable Instance Attribution

Authors: Rosina O. Weber, Prateek Goel, Shideh Amiri, Gideon Simpson

Abstract: Previous research in interpretable machine learning (IML) and explainable artificial intelligence (XAI) can be broadly categorized as either focusing on seeking interpretability in the agent's model (i.e., IML) or focusing on the context of the user in addition to the model (i.e., XAI). The former can be categorized as feature or instance attribution. Example- or sample-based methods such as those… ▽ More Previous research in interpretable machine learning (IML) and explainable artificial intelligence (XAI) can be broadly categorized as either focusing on seeking interpretability in the agent's model (i.e., IML) or focusing on the context of the user in addition to the model (i.e., XAI). The former can be categorized as feature or instance attribution. Example- or sample-based methods such as those using or inspired by case-based reasoning (CBR) rely on various approaches to select instances that are not necessarily attributing instances responsible for an agent's decision. Furthermore, existing approaches have focused on interpretability and explainability but fall short when it comes to accountability. Inspired in case-based reasoning principles, this paper introduces a pseudo-metric we call Longitudinal distance and its use to attribute instances to a neural network agent's decision that can be potentially used to build accountable CBR agents. △ Less

Submitted 23 August, 2021; originally announced August 2021.

Is Automated Topic Model Evaluation Broken?: The Incoherence of Coherence

Authors: Alexander Hoyle, Pranav Goel, Denis Peskov, Andrew Hian-Cheong, Jordan Boyd-Graber, Philip Resnik

Abstract: Topic model evaluation, like evaluation of other unsupervised methods, can be contentious. However, the field has coalesced around automated estimates of topic coherence, which rely on the frequency of word co-occurrences in a reference corpus. Contemporary neural topic models surpass classical ones according to these metrics. At the same time, topic model evaluation suffers from a validation gap:… ▽ More Topic model evaluation, like evaluation of other unsupervised methods, can be contentious. However, the field has coalesced around automated estimates of topic coherence, which rely on the frequency of word co-occurrences in a reference corpus. Contemporary neural topic models surpass classical ones according to these metrics. At the same time, topic model evaluation suffers from a validation gap: automated coherence, developed for classical models, has not been validated using human experimentation for neural models. In addition, a meta-analysis of topic modeling literature reveals a substantial standardization gap in automated topic modeling benchmarks. To address the validation gap, we compare automated coherence with the two most widely accepted human judgment tasks: topic rating and word intrusion. To address the standardization gap, we systematically evaluate a dominant classical model and two state-of-the-art neural models on two commonly used datasets. Automated evaluations declare a winning model when corresponding human evaluations do not, calling into question the validity of fully automatic evaluations independent of human judgments. △ Less

Submitted 27 October, 2021; v1 submitted 5 July, 2021; originally announced July 2021.

Comments: Accepted to NeurIPS 2021 (spotlight presentation). CR version

A short-range structural insight into lithium substituted barium vanadate glasses using Raman and EPR spectroscopy as probes

Authors: Parul Goel, Gajanan V Honnavar

Abstract: We present a corroborative study of structural characterization of lithium substituted barium vanadate glasses using Raman and Electron Paramagnetic Resonance (EPR) spectroscopy. Investigation of the thermal and physical properties of these glasses showed a gradual increase in the concentration of non-bridging oxygen. Raman and EPR analysis gave an insight into the changing structure of the glasse… ▽ More We present a corroborative study of structural characterization of lithium substituted barium vanadate glasses using Raman and Electron Paramagnetic Resonance (EPR) spectroscopy. Investigation of the thermal and physical properties of these glasses showed a gradual increase in the concentration of non-bridging oxygen. Raman and EPR analysis gave an insight into the changing structure of the glasses. Both the spectroscopic techniques confirmed that vanadium is present in the glasses as distorted VO6 octahedra. From the analysis of both spectroscopic techniques, it is proposed that the lithium ion prefers to occupy planar positions of the VO6 octahedra thus reducing the tetragonal distortion and making the environment around the network forming unit in the glass matrix more homogenous as we increase lithium content. The concentration of V4+ showed a non-monotonic variation with an increase in Li2O as indicated by Raman studies and confirmed by EPR which indicates a structural change in the distorted VO6 octahedra. △ Less

Submitted 28 June, 2021; v1 submitted 8 April, 2021; originally announced April 2021.

Comments: 22 pages, 7 figures

Analysis on Image Set Visual Question Answering

Authors: Abhinav Khattar, Aviral Joshi, Har Simrat Singh, Pulkit Goel, Rohit Prakash Barnwal

Abstract: We tackle the challenge of Visual Question Answering in multi-image setting for the ISVQA dataset. Traditional VQA tasks have focused on a single-image setting where the target answer is generated from a single image. Image set VQA, however, comprises of a set of images and requires finding connection between images, relate the objects across images based on these connections and generate a unifie… ▽ More We tackle the challenge of Visual Question Answering in multi-image setting for the ISVQA dataset. Traditional VQA tasks have focused on a single-image setting where the target answer is generated from a single image. Image set VQA, however, comprises of a set of images and requires finding connection between images, relate the objects across images based on these connections and generate a unified answer. In this report, we work with 4 approaches in a bid to improve the performance on the task. We analyse and compare our results with three baseline models - LXMERT, HME-VideoQA and VisualBERT - and show that our approaches can provide a slight improvement over the baselines. In specific, we try to improve on the spatial awareness of the model and help the model identify color using enhanced pre-training, reduce language dependence using adversarial regularization, and improve counting using regression loss and graph based deduplication. We further delve into an in-depth analysis on the language bias in the ISVQA dataset and show how models trained on ISVQA implicitly learn to associate language more strongly with the final answer. △ Less

Submitted 31 March, 2021; originally announced April 2021.

On the Robustness of Monte Carlo Dropout Trained with Noisy Labels

Authors: Purvi Goel, Li Chen

Abstract: The memorization effect of deep learning hinders its performance to effectively generalize on test set when learning with noisy labels. Prior study has discovered that epistemic uncertainty techniques are robust when trained with noisy labels compared with neural networks without uncertainty estimation. They obtain prolonged memorization effect and better generalization performance under the adver… ▽ More The memorization effect of deep learning hinders its performance to effectively generalize on test set when learning with noisy labels. Prior study has discovered that epistemic uncertainty techniques are robust when trained with noisy labels compared with neural networks without uncertainty estimation. They obtain prolonged memorization effect and better generalization performance under the adversarial setting of noisy labels. Due to its superior performance amongst other selected epistemic uncertainty methods under noisy labels, we focus on Monte Carlo Dropout (MCDropout) and investigate why it is robust when trained with noisy labels. Through empirical studies on datasets MNIST, CIFAR-10, Animal-10n, we deep dive into three aspects of MCDropout under noisy label setting: 1. efficacy: understanding the learning behavior and test accuracy of MCDropout when training set contains artificially generated or naturally embedded label noise; 2. representation volatility: studying the responsiveness of neurons by examining the mean and standard deviation on each neuron's activation; 3. network sparsity: investigating the network support of MCDropout in comparison with deterministic neural networks. Our findings suggest that MCDropout further sparsifies and regularizes the deterministic neural networks and thus provides higher robustness against noisy labels. △ Less

Submitted 22 March, 2021; originally announced March 2021.

ICodeNet -- A Hierarchical Neural Network Approach for Source Code Author Identification

Authors: Pranali Bora, Tulika Awalgaonkar, Himanshu Palve, Raviraj Joshi, Purvi Goel

Abstract: With the open-source revolution, source codes are now more easily accessible than ever. This has, however, made it easier for malicious users and institutions to copy the code without giving regards to the license, or credit to the original author. Therefore, source code author identification is a critical task with paramount importance. In this paper, we propose ICodeNet - a hierarchical neural n… ▽ More With the open-source revolution, source codes are now more easily accessible than ever. This has, however, made it easier for malicious users and institutions to copy the code without giving regards to the license, or credit to the original author. Therefore, source code author identification is a critical task with paramount importance. In this paper, we propose ICodeNet - a hierarchical neural network that can be used for source code file-level tasks. The ICodeNet processes source code in image format and is employed for the task of per file author identification. The ICodeNet consists of an ImageNet trained VGG encoder followed by a shallow neural network. The shallow network is based either on CNN or LSTM. Different variations of models are evaluated on a source code author classification dataset. We have also compared our image-based hierarchical neural network model with simple image-based CNN architecture and text-based CNN and LSTM models to highlight its novelty and efficiency. △ Less

Submitted 30 January, 2021; originally announced February 2021.

Comments: Accepted at ICMLC 2021

On a Generalized Briot-Bouquet type Differential Subordination

Authors: S. Sivaprasad Kumar, Priyanka Goel

Abstract: We introduce and study the following special type of differential subordination implication: \begin{equation}\label{abs} p(z)Q(z)+\frac{zp'(z)}{βp(z)+α}\prec h(z)\quad\Rightarrow p(z)\prec h(z), \end{equation} which generalizes the Briot-Bouquet differential subordination, where $Q(z)$ is analytic and $0\neqβ,α\in\mathbb{C}.$ Further, some special cases of our result are also discussed. Finally,… ▽ More We introduce and study the following special type of differential subordination implication: \begin{equation}\label{abs} p(z)Q(z)+\frac{zp'(z)}{βp(z)+α}\prec h(z)\quad\Rightarrow p(z)\prec h(z), \end{equation} which generalizes the Briot-Bouquet differential subordination, where $Q(z)$ is analytic and $0\neqβ,α\in\mathbb{C}.$ Further, some special cases of our result are also discussed. Finally, analogues of open door lemma and integral existence theorem with applications to univalent functions are obtained. △ Less

Submitted 11 March, 2022; v1 submitted 1 January, 2021; originally announced January 2021.

Lexically-constrained Text Generation through Commonsense Knowledge Extraction and Injection

Authors: Yikang Li, Pulkit Goel, Varsha Kuppur Rajendra, Har Simrat Singh, Jonathan Francis, Kaixin Ma, Eric Nyberg, Alessandro Oltramari

Abstract: Conditional text generation has been a challenging task that is yet to see human-level performance from state-of-the-art models. In this work, we specifically focus on the Commongen benchmark, wherein the aim is to generate a plausible sentence for a given set of input concepts. Despite advances in other tasks, large pre-trained language models that are fine-tuned on this dataset often produce sen… ▽ More Conditional text generation has been a challenging task that is yet to see human-level performance from state-of-the-art models. In this work, we specifically focus on the Commongen benchmark, wherein the aim is to generate a plausible sentence for a given set of input concepts. Despite advances in other tasks, large pre-trained language models that are fine-tuned on this dataset often produce sentences that are syntactically correct but qualitatively deviate from a human understanding of common sense. Furthermore, generated sequences are unable to fulfill such lexical requirements as matching part-of-speech and full concept coverage. In this paper, we explore how commonsense knowledge graphs can enhance model performance, with respect to commonsense reasoning and lexically-constrained decoding. We propose strategies for enhancing the semantic correctness of the generated text, which we accomplish through: extracting commonsense relations from Conceptnet, injecting these relations into the Unified Language Model (UniLM) through attention mechanisms, and enforcing the aforementioned lexical requirements through output constraints. By performing several ablations, we find that commonsense injection enables the generation of sentences that are more aligned with human understanding, while remaining compliant with lexical requirements. △ Less

Submitted 19 December, 2020; originally announced December 2020.

Comments: AAAI-CSKG 2021

Shape From Tracing: Towards Reconstructing 3D Object Geometry and SVBRDF Material from Images via Differentiable Path Tracing

Authors: Purvi Goel, Loudon Cohen, James Guesman, Vikas Thamizharasan, James Tompkin, Daniel Ritchie

Abstract: Reconstructing object geometry and material from multiple views typically requires optimization. Differentiable path tracing is an appealing framework as it can reproduce complex appearance effects. However, it is difficult to use due to high computational cost. In this paper, we explore how to use differentiable ray tracing to refine an initial coarse mesh and per-mesh-facet material representati… ▽ More Reconstructing object geometry and material from multiple views typically requires optimization. Differentiable path tracing is an appealing framework as it can reproduce complex appearance effects. However, it is difficult to use due to high computational cost. In this paper, we explore how to use differentiable ray tracing to refine an initial coarse mesh and per-mesh-facet material representation. In simulation, we find that it is possible to reconstruct fine geometric and material detail from low resolution input views, allowing high-quality reconstructions in a few hours despite the expense of path tracing. The reconstructions successfully disambiguate shading, shadow, and global illumination effects such as diffuse interreflection from material properties. We demonstrate the impact of different geometry initializations, including space carving, multi-view stereo, and 3D neural networks. Finally, with input captured using smartphone video and a consumer 360? camera for lighting estimation, we also show how to refine initial reconstructions of real-world objects in unconstrained environments. △ Less

Submitted 6 December, 2020; originally announced December 2020.

Comments: Will be published at 3DV 2020

Data Representing Ground-Truth Explanations to Evaluate XAI Methods

Authors: Shideh Shams Amiri, Rosina O. Weber, Prateek Goel, Owen Brooks, Archer Gandley, Brian Kitchell, Aaron Zehm

Abstract: Explainable artificial intelligence (XAI) methods are currently evaluated with approaches mostly originated in interpretable machine learning (IML) research that focus on understanding models such as comparison against existing attribution approaches, sensitivity analyses, gold set of features, axioms, or through demonstration of images. There are problems with these methods such as that they do n… ▽ More Explainable artificial intelligence (XAI) methods are currently evaluated with approaches mostly originated in interpretable machine learning (IML) research that focus on understanding models such as comparison against existing attribution approaches, sensitivity analyses, gold set of features, axioms, or through demonstration of images. There are problems with these methods such as that they do not indicate where current XAI approaches fail to guide investigations towards consistent progress of the field. They do not measure accuracy in support of accountable decisions, and it is practically impossible to determine whether one XAI method is better than the other or what the weaknesses of existing models are, leaving researchers without guidance on which research questions will advance the field. Other fields usually utilize ground-truth data and create benchmarks. Data representing ground-truth explanations is not typically used in XAI or IML. One reason is that explanations are subjective, in the sense that an explanation that satisfies one user may not satisfy another. To overcome these problems, we propose to represent explanations with canonical equations that can be used to evaluate the accuracy of XAI methods. The contributions of this paper include a methodology to create synthetic data representing ground-truth explanations, three data sets, an evaluation of LIME using these data sets, and a preliminary analysis of the challenges and potential benefits in using these data to evaluate existing XAI approaches. Evaluation methods based on human-centric studies are outside the scope of this paper. △ Less

Submitted 18 November, 2020; originally announced November 2020.

Comments: Submitted to the AAAI 2021 Explainable Agency in Artificial Intelligence Workshop, 6 pages, 3 figures and 2 tables

Robust Deep Learning with Active Noise Cancellation for Spatial Computing

Authors: Li Chen, David Yang, Purvi Goel, Ilknur Kabul

Abstract: This paper proposes CANC, a Co-teaching Active Noise Cancellation method, applied in spatial computing to address deep learning trained with extreme noisy labels. Deep learning algorithms have been successful in spatial computing for land or building footprint recognition. However a lot of noise exists in ground truth labels due to how labels are collected in spatial computing and satellite imager… ▽ More This paper proposes CANC, a Co-teaching Active Noise Cancellation method, applied in spatial computing to address deep learning trained with extreme noisy labels. Deep learning algorithms have been successful in spatial computing for land or building footprint recognition. However a lot of noise exists in ground truth labels due to how labels are collected in spatial computing and satellite imagery. Existing methods to deal with extreme label noise conduct clean sample selection and do not utilize the remaining samples. Such techniques can be wasteful due to the cost of data retrieval. Our proposed CANC algorithm not only conserves high-cost training samples but also provides active label correction to better improve robust deep learning with extreme noisy labels. We demonstrate the effectiveness of CANC for building footprint recognition for spatial computing. △ Less

Submitted 16 November, 2020; originally announced November 2020.

Improving Neural Topic Models using Knowledge Distillation

Authors: Alexander Hoyle, Pranav Goel, Philip Resnik

Abstract: Topic models are often used to identify human-interpretable topics to help make sense of large document collections. We use knowledge distillation to combine the best attributes of probabilistic topic models and pretrained transformers. Our modular method can be straightforwardly applied with any neural topic model to improve topic quality, which we demonstrate using two models having disparate ar… ▽ More Topic models are often used to identify human-interpretable topics to help make sense of large document collections. We use knowledge distillation to combine the best attributes of probabilistic topic models and pretrained transformers. Our modular method can be straightforwardly applied with any neural topic model to improve topic quality, which we demonstrate using two models having disparate architectures, obtaining state-of-the-art topic coherence. We show that our adaptable framework not only improves performance in the aggregate over all estimated topics, as is commonly reported, but also in head-to-head comparisons of aligned topics. △ Less

Submitted 5 October, 2020; originally announced October 2020.

Comments: Accepted to EMNLP 2020

SURF-SVM Based Identification and Classification of Gastrointestinal Diseases in Wireless Capsule Endoscopy

Authors: Vanshika Vats, Pooja Goel, Amodini Agarwal, Nidhi Goel

Abstract: Endoscopy provides a major contribution to the diagnosis of the Gastrointestinal Tract (GIT) diseases. With Colon Endoscopy having its certain limitations, Wireless Capsule Endoscopy is gradually taking over it in the terms of ease and efficiency. WCE is performed with a miniature optical endoscope which is swallowed by the patient and transmits colour images wirelessly during its journey through… ▽ More Endoscopy provides a major contribution to the diagnosis of the Gastrointestinal Tract (GIT) diseases. With Colon Endoscopy having its certain limitations, Wireless Capsule Endoscopy is gradually taking over it in the terms of ease and efficiency. WCE is performed with a miniature optical endoscope which is swallowed by the patient and transmits colour images wirelessly during its journey through the GIT, inside the body of the patient. These images are used to implement an effective and computationally efficient approach which aims to detect the abnormal and normal tissues in the GIT automatically, and thus helps in reducing the manual work of the reviewers. The algorithm further aims to classify the diseased tissues into various GIT diseases that are commonly known to be affecting the tract. In this manuscript, the descriptor used for the detection of the interest points is Speeded Up Robust Features (SURF), which uses the colour information contained in the images which is converted to CIELAB space colours for better identification. The features extracted at the interest points are then used to train and test a Support Vector Machine (SVM), so that it automatically classifies the images into normal or abnormal and further detects the specific abnormalities. SVM, along with a few parameters, gives a very high accuracy of 94.58% while classifying normal and abnormal images and an accuracy of 82.91% while classifying into multi-class. The present work is an improvement on the previously reported analyses which were only limited to the bi-class classification using this approach. △ Less

Submitted 2 September, 2020; originally announced September 2020.

Comments: 5 pages, 7 figures, 15 subfigures in total

Towards Automatic Generation of Questions from Long Answers

Authors: Shlok Kumar Mishra, Pranav Goel, Abhishek Sharma, Abhyuday Jagannatha, David Jacobs, Hal Daumé III

Abstract: Automatic question generation (AQG) has broad applicability in domains such as tutoring systems, conversational agents, healthcare literacy, and information retrieval. Existing efforts at AQG have been limited to short answer lengths of up to two or three sentences. However, several real-world applications require question generation from answers that span several sentences. Therefore, we propose… ▽ More Automatic question generation (AQG) has broad applicability in domains such as tutoring systems, conversational agents, healthcare literacy, and information retrieval. Existing efforts at AQG have been limited to short answer lengths of up to two or three sentences. However, several real-world applications require question generation from answers that span several sentences. Therefore, we propose a novel evaluation benchmark to assess the performance of existing AQG systems for long-text answers. We leverage the large-scale open-source Google Natural Questions dataset to create the aforementioned long-answer AQG benchmark. We empirically demonstrate that the performance of existing AQG methods significantly degrades as the length of the answer increases. Transformer-based methods outperform other existing AQG methods on long answers in terms of automatic as well as human evaluation. However, we still observe degradation in the performance of our best performing models with increasing sentence length, suggesting that long answer QA is a challenging benchmark task for future research. △ Less

Submitted 15 April, 2020; v1 submitted 10 April, 2020; originally announced April 2020.

Spin-Phonon Coupling and Thermodynamic Behaviour in YCrO3 and LaCrO3: Inelastic Neutron Scattering and Lattice Dynamics

Authors: Mayanak K. Gupta, Ranjan Mittal, Sanjay K. Mishra, Prabhatasree Goel, Baltej Singh, Stephane Rols, Samrath L. Chaplot

Abstract: We report detailed temperature-dependent inelastic neutron scattering and ab-initio lattice dynamics investigation of magnetic perovskites YCrO3 and LaCrO3. The magnetic neutron scattering from the Cr ions exhibits significant changes with temperature and dominates at low momentum transfer regime. Ab-inito calculations performed including magnetic interactions show that the effect of magnetic inte… ▽ More We report detailed temperature-dependent inelastic neutron scattering and ab-initio lattice dynamics investigation of magnetic perovskites YCrO3 and LaCrO3. The magnetic neutron scattering from the Cr ions exhibits significant changes with temperature and dominates at low momentum transfer regime. Ab-inito calculations performed including magnetic interactions show that the effect of magnetic interaction is very signicant on the low- as well as high-energy phonon modes. We have also shown that the inelastic neutron spectrum of YCrO3 mimics the magnon spectrum from a G-type antiferromagnetic system, which is consistent with previously reported magnetic structure in the compound. The ab-initio lattice dynamics calculations in both the compounds exhibit anisotropic thermal expansion behaviour in the orthorhombic structure and predict negative thermal expansion along the crystallographic a-axis at low temperatures. We identify the anharmonic phonon modes responsible for this anamolous behaviour in LaCrO3 involving low-energy La vibrations and distortions of the CrO6 octahedra. △ Less

Submitted 3 April, 2020; originally announced April 2020.

Comments: 29 Pages, 13 Figures

Phonons and Oxygen Diffusion in Bi2O3 and (Bi0.7Y0.3)2O3

Authors: Prabhatasree Goel, M. K. Gupta, R. Mittal, S. J. Skinner, S. Mukhopadhyay S. Rols, S. L. Chaplot

Abstract: We report investigation of phonons and oxygen diffusion in Bi2O3 and (Bi0.7Y0.3)2O3. The phonon spectra have been measured in Bi2O3 at high temperatures up to 1083 K using inelastic neutron scattering. Ab-initio calculations have been used to compute the individual contributions of the constituent atoms in Bi2O3 and (Bi0.7Y0.3)2O3 to the total phonon density of states. Our computed results indicat… ▽ More We report investigation of phonons and oxygen diffusion in Bi2O3 and (Bi0.7Y0.3)2O3. The phonon spectra have been measured in Bi2O3 at high temperatures up to 1083 K using inelastic neutron scattering. Ab-initio calculations have been used to compute the individual contributions of the constituent atoms in Bi2O3 and (Bi0.7Y0.3)2O3 to the total phonon density of states. Our computed results indicate that as temperature is increased, there is a complete loss of sharp peak structure in the vibrational density of states. Ab-initio molecular dynamics simulations show that even at 1000 K in δ-phase Bi2O3, Bi-Bi correlations remain ordered in the crystalline lattice while the correlations between O-O show liquid like disordered behavior. In the case of (Bi0.7Y0.3)2O3, the O-O correlations broadened at around 500 K indicating that oxygen conductivity is possible at such low temperatures in (Bi0.7Y0.3)2O3 although the conductivity is much less than that observed in the undoped high temperature δ-phase of Bi2O3. This result is consistent with the calculated diffusion coefficients of oxygen and observation by QENS experiments. Our ab-initio molecular dynamics calculations predict that macroscopic diffusion is attainable in (Bi0.7Y0.3)2O3 at much lower temperatures, which is more suited for technological applications. Our studies elucidate the easy directions of diffusion in δ-Bi2O3 and (Bi0.7Y0.3)2O3. △ Less

Submitted 13 February, 2020; originally announced February 2020.

Deep Learning for Hindi Text Classification: A Comparison

Authors: Ramchandra Joshi, Purvi Goel, Raviraj Joshi

Abstract: Natural Language Processing (NLP) and especially natural language text analysis have seen great advances in recent times. Usage of deep learning in text processing has revolutionized the techniques for text processing and achieved remarkable results. Different deep learning architectures like CNN, LSTM, and very recent Transformer have been used to achieve state of the art results variety on NLP t… ▽ More Natural Language Processing (NLP) and especially natural language text analysis have seen great advances in recent times. Usage of deep learning in text processing has revolutionized the techniques for text processing and achieved remarkable results. Different deep learning architectures like CNN, LSTM, and very recent Transformer have been used to achieve state of the art results variety on NLP tasks. In this work, we survey a host of deep learning architectures for text classification tasks. The work is specifically concerned with the classification of Hindi text. The research in the classification of morphologically rich and low resource Hindi language written in Devanagari script has been limited due to the absence of large labeled corpus. In this work, we used translated versions of English data-sets to evaluate models based on CNN, LSTM and Attention. Multilingual pre-trained sentence embeddings based on BERT and LASER are also compared to evaluate their effectiveness for the Hindi language. The paper also serves as a tutorial for popular text classification techniques. △ Less

Submitted 19 January, 2020; originally announced January 2020.

Comments: Accepted at International Conference on Intelligent Human Computer Interaction(IHCI) 2019

Dis-entangling Mixture of Interventions on a Causal Bayesian Network Using Aggregate Observations

Authors: Gaurav Sinha, Ayush Chauhan, Aurghya Maiti, Naman Poddar, Pulkit Goel

Abstract: We study the problem of separating a mixture of distributions, all of which come from interventions on a known causal bayesian network. Given oracle access to marginals of all distributions resulting from interventions on the network, and estimates of marginals from the mixture distribution, we want to recover the mixing proportions of different mixture components. We show that in the worst case… ▽ More We study the problem of separating a mixture of distributions, all of which come from interventions on a known causal bayesian network. Given oracle access to marginals of all distributions resulting from interventions on the network, and estimates of marginals from the mixture distribution, we want to recover the mixing proportions of different mixture components. We show that in the worst case, mixing proportions cannot be identified using marginals only. If exact marginals of the mixture distribution were known, under a simple assumption of excluding a few distributions from the mixture, we show that the mixing proportions become identifiable. Our identifiability proof is constructive and gives an efficient algorithm recovering the mixing proportions exactly. When exact marginals are not available, we design an optimization framework to estimate the mixing proportions. Our problem is motivated from a real-world scenario of an e-commerce business, where multiple interventions occur at a given time, leading to deviations in expected metrics. We conduct experiments on the well known publicly available ALARM network and on a proprietary dataset from a large e-commerce company validating the performance of our method. △ Less

Submitted 15 January, 2020; v1 submitted 30 November, 2019; originally announced December 2019.

Comments: Accepted at the Ninth International Workshop on Statistical Relational AI, AAAI 2020

Dynamics of Na Ion in the Amorphous Na2Si2O5 Using Quasielastic Neutron Scattering and Molecular Dynamics Simulations

Authors: Mayanak K. Gupta, Sanjay K. Mishra, Ranjan Mittal, Baltej Singh, Prabhatasree Goel, Sanghamitra Mukhopadhyay, Rakesh Shukla, Srungarpu N. Achary, Avesh K. Tyagi, Samrath L. Chaplot

Abstract: We have investigated the dynamics of Na ions in amorphous Na2Si2O5, a potential solid electrolyte material for Na-battery. We have employed quasielastic neutron scattering (QENS) technique in the amorphous Na2Si2O5 from 300 to 748 K to understand the diffusion pathways and relaxation timescales of Na atom dynamics. The microscopic analysis of the QENS data has been performed using ab-initio and cl… ▽ More We have investigated the dynamics of Na ions in amorphous Na2Si2O5, a potential solid electrolyte material for Na-battery. We have employed quasielastic neutron scattering (QENS) technique in the amorphous Na2Si2O5 from 300 to 748 K to understand the diffusion pathways and relaxation timescales of Na atom dynamics. The microscopic analysis of the QENS data has been performed using ab-initio and classical molecular dynamics simulations (MD) to understand the Na-ion diffusion in the amorphous phase. Our experimental studies show that the traditional model, such as the Hall and Ross (H-R) model, fairly well describe the diffusion in the amorphous phase giving a mean jump length of ~3 Å and residence time about 9.1 picoseconds. Our MD simulations have indicated that the diffusion of Na+ ions occurs in the amorphous phase of Na2Si2O5 while that is not observed in the crystalline orthorhombic phase even up to 1100 K. The MD simulations have revealed that in the amorphous phase, due to different orientations of silicon polyhedral units, accessible pathways are opened up for Na+ diffusions. These pathways are not available in the crystalline phase of Na2Si2O5 due to rigid spatial arrangement of silicon polyhedral units. △ Less

Submitted 16 September, 2019; originally announced September 2019.

Comments: 24 Pages, 12 figures

Journal ref: Phys. Rev. Materials 4, 045802 (2020)

On sharp bounds of certain Close-to-Convex functions

Authors: Priyanka Goel, S. Sivaprasad Kumar

Abstract: We derive general formula for the fourth coefficient of the functions belonging to the Carathéodory class involving the parameters lying in the open unit disk. Further, we obtain sharp upper bounds of initial inverse coefficients for certain close-to-convex functions satisfying any one of the inequalities: $\RE((1-z)f'(z))>0,$ $\RE((1-z^2)f'(z))>0,$ $\RE((1-z+z^2)f'(z))>0$ and… ▽ More We derive general formula for the fourth coefficient of the functions belonging to the Carathéodory class involving the parameters lying in the open unit disk. Further, we obtain sharp upper bounds of initial inverse coefficients for certain close-to-convex functions satisfying any one of the inequalities: $\RE((1-z)f'(z))>0,$ $\RE((1-z^2)f'(z))>0,$ $\RE((1-z+z^2)f'(z))>0$ and $\RE((1-z)^2f'(z))>0$. △ Less

Submitted 7 October, 2020; v1 submitted 31 July, 2019; originally announced July 2019.

Comments: 19

MSC Class: 30C45; 30C50

Lithium Diffusion in Li2X(X=O, S and Se): Ab-initio Simulations and Neutron Inelastic Scattering Measurements

Authors: M. K. Gupta, Baltej Singh, Prabhatasree Goel, R. Mittal, S. Rols, S. L. Chaplot

Abstract: We have performed ab-initio lattice dynamics and molecular dynamics studies of Li2X (X=O, S and Se) to understand the ionic conduction in these compounds. The inelastic neutron scattering measurements on Li2O have been performed across its superionic transition temperature of about 1200 K. The experimental spectra show significant changes around the superionic transition temperature, which is attr… ▽ More We have performed ab-initio lattice dynamics and molecular dynamics studies of Li2X (X=O, S and Se) to understand the ionic conduction in these compounds. The inelastic neutron scattering measurements on Li2O have been performed across its superionic transition temperature of about 1200 K. The experimental spectra show significant changes around the superionic transition temperature, which is attributed to large diffusion of lithium as well as its large vibrational amplitude. We have identified a correlation between the chemical pressure (ionic radius of X atom) and the superionic transition temperature. The simulations are able to provide the ionic diffusion pathways in Li2X. △ Less

Submitted 21 March, 2019; originally announced March 2019.

Comments: 25 Pages, 15 Figures

Journal ref: Phys. Rev. B 99, 224304 (2019)

Phonons and Anisotropic Thermal Expansion Behaviour of NiX (X = S, Se, Te)

Authors: Prabhatasree Goel, M. K. Gupta, S. K. Mishra, Baltej Singh, R. Mittal, P. U. Sastry, A. Thamizhavel, S. L. Chaplot

Abstract: Metal Chalcogenides have been known for important technological applications and have attracted continuous interest in their structure, electronic, thermal and transport properties. Here we present first principles calculations of the vibrational and thermodynamic properties of NiX (X = S, Se, Te) compounds along with inelastic neutron scattering measurements of the phonon spectrum in NiSe. The me… ▽ More Metal Chalcogenides have been known for important technological applications and have attracted continuous interest in their structure, electronic, thermal and transport properties. Here we present first principles calculations of the vibrational and thermodynamic properties of NiX (X = S, Se, Te) compounds along with inelastic neutron scattering measurements of the phonon spectrum in NiSe. The measured phonon spectrum is in very good agreement with the computed result. We also report the measurement of thermal expansion behavior of NiSe using X-ray diffraction from 13 K to 300 K. The change in the hexagonal c lattice parameter in NiSe is considerably greater as compared to a parameter. The ab-initio calculated anisotropic Grüneisen parameters of the different phonon modes in all the chalcogenides along with the elastic constants are used to compute anisotropic thermal expansion behviour, which is found in good agreement with experiments. The displacement pattern of phonons indicate that difference in amplitudes of Ni and X atoms follow the anisotropy of thermal expansion behavior along c- and a-axis. △ Less

Submitted 21 February, 2019; originally announced February 2019.

Comments: 21 Pages, 9 Figures

Observation of Mixed Alkali Like Behaviour by Fluorine Ion in Mixed Alkali Oxyfluro Vanadate Glasses: Analysis from Conductivity Measurements

Authors: Gajanan V Honnavar, Vaibhav Varade, Parul Goel, K P Ramesh

Abstract: In this communication we report the fluorine ion dynamics in mixed alkali oxyfluro vanadate glasses. We have measured the electrical conductivity using impedance spectroscopy technique Room temperature conductivity falls to 5 orders of magnitude from its single alkali values at 33 mol% of rubidium concentration. We have also estimated the distance between similar mobile ions using the density valu… ▽ More In this communication we report the fluorine ion dynamics in mixed alkali oxyfluro vanadate glasses. We have measured the electrical conductivity using impedance spectroscopy technique Room temperature conductivity falls to 5 orders of magnitude from its single alkali values at 33 mol% of rubidium concentration. We have also estimated the distance between similar mobile ions using the density values. Assuming this distance as the hop** distance between the similar ions we have estimated the anionic (Fluorine ion in our case) conductivity. It is observed that the fluorine ion dynamics mimics the mixed alkali effect and scales as the onset frequency f0. △ Less

Submitted 30 July, 2018; originally announced July 2018.

Comments: submitted to DAE-SSDP 2018 India

A new NS3 Implementation of CCNx 1.0 Protocol

Authors: Marc Mosko, Ramesh Ayyagari, Priti Goel, Eric Holmberg, Mark Konezny

Abstract: The ccns3Sim project is an open source implementation of the CCNx 1.0 protocols for the NS3 simulator. We describe the implementation and several important features including modularity and process delay simulation. The ccns3Sim implementation is a fresh NS3-specific implementation. Like NS3 itself, it uses C++98 standard, NS3 code style, NS3 smart pointers, NS3 xUnit, and integrates with the NS3… ▽ More The ccns3Sim project is an open source implementation of the CCNx 1.0 protocols for the NS3 simulator. We describe the implementation and several important features including modularity and process delay simulation. The ccns3Sim implementation is a fresh NS3-specific implementation. Like NS3 itself, it uses C++98 standard, NS3 code style, NS3 smart pointers, NS3 xUnit, and integrates with the NS3 documentation and manual. A user or developer does not need to learn two systems. If one knows NS3, one should be able to get started with the CCNx code right away. A developer can easily use their own implementation of the layer 3 protocol, layer 4 protocol, forwarder, routing protocol, Pending Interest Table (PIT) or Forwarding Information Base (FIB) or Content Store (CS). A user may configure or specify a new implementation for any of these features at runtime in the simulation script. In this paper, we describe the software architecture and give examples of using the simulator. We evaluate the implementation with several example experiments on ICN caching. △ Less

Submitted 15 July, 2017; originally announced July 2017.

Automatic Identification of Sarcasm Target: An Introductory Approach

Authors: Aditya Joshi, Pranav Goel, Pushpak Bhattacharyya, Mark Carman

Abstract: Past work in computational sarcasm deals primarily with sarcasm detection. In this paper, we introduce a novel, related problem: sarcasm target identification i.e., extracting the target of ridicule in a sarcastic sentence). We present an introductory approach for sarcasm target identification. Our approach employs two types of extractors: one based on rules, and another consisting of a statistica… ▽ More Past work in computational sarcasm deals primarily with sarcasm detection. In this paper, we introduce a novel, related problem: sarcasm target identification i.e., extracting the target of ridicule in a sarcastic sentence). We present an introductory approach for sarcasm target identification. Our approach employs two types of extractors: one based on rules, and another consisting of a statistical classifier. To compare our approach, we use two baselines: a naïve baseline and another baseline based on work in sentiment target identification. We perform our experiments on book snippets and tweets, and show that our hybrid approach performs better than the two baselines and also, in comparison with using the two extractors individually. Our introductory approach establishes the viability of sarcasm target identification, and will serve as a baseline for future work. △ Less

Submitted 25 August, 2017; v1 submitted 22 October, 2016; originally announced October 2016.

New Insights into the Compressibility and High-Pressure Stability of Ni(CN)2 from Neutron Diffraction, Raman Spectroscopy and Inelastic Neutron Scattering

Authors: S. K. Mishra, R. Mittal, M. Zbiri, Rekha Rao, Prabhatasree Goel, S. J. Hibble, A. M. Chippindale, T. Hansen, H. Schober, S. L. Chaplot

Abstract: The layered structure of tetragonal Ni(CN)2, consisting of square-planar Ni(CN)4 units linked in the a-b plane, with no true periodicity along the c-axis, is expected to show anisotropic compression on the application of pressure. High-pressure neutron diffraction (elastic) and inelastic neutron scattering experiments have been performed on polycrystalline Ni(CN)2 to investigate its compressibilit… ▽ More The layered structure of tetragonal Ni(CN)2, consisting of square-planar Ni(CN)4 units linked in the a-b plane, with no true periodicity along the c-axis, is expected to show anisotropic compression on the application of pressure. High-pressure neutron diffraction (elastic) and inelastic neutron scattering experiments have been performed on polycrystalline Ni(CN)2 to investigate its compressibility and stability. The intralayer a lattice parameter does not show any appreciable variation with increase of pressure up to 2.7 kbar. Above this pressure value, a decrease in a is observed. The c lattice parameter decreases slowly up to 1 kbar, then decreases sharply up to 20 kbar. It does not show any significant variation with further pressure increase up to 50 kbar. The response of the lattice parameters to the applied pressure is strongly anisotropic as the interlayer spacing (along the c-axis) shows a significantly larger contraction than the a-b plane. The experimental pressure dependence of the volume data is fitted to a bulk modulus, B0, of 1050 (20) kbar over the pressure range 0-1 kbar, and to 154 (2) kbar in the range 1-50 kbar. The change in the slope of the lattice parameters at 1 kbar is also supported by high-pressure Raman measurements, which indicate a phase transition at 1 kbar. Probably arising from a change in the CN ordering within the Ni(CN)2 layers. Raman measurements, performed up to 200 kbar, highlight the possible existence of a second phase transition taking place at about 70 kbar. Our neutron inelastic scattering measurements of the pressure dependence of the phonon spectra performed up to 2.7 kbar, also support the occurrence of a phase transition at low pressure. △ Less

Submitted 14 July, 2015; v1 submitted 8 September, 2014; originally announced September 2014.

Comments: 18 Pages, 9 Figures

Journal ref: J. Phys.: Condens. Matter 28, 045402 (2016)

Inelastic Neutron Scattering Studies of Phonon Spectra and Simulations in Tungstates, AWO4 (A = Ba, Sr, Ca and Pb)

Authors: Prabhatasree Goel, M. K. Gupta, R. Mittal, S. Rols, S. N. Achary, A. K. Tyagi, S. L. Chaplot

Abstract: Lattice dynamics and high pressure phase transitions in AWO4 (A = Ba, Sr, Ca and Pb) have been investigated using inelastic neutron scattering experiments, ab-initio density functional theory calculations and extensive molecular dynamics simulations. The vibrational modes that are internal to WO4 tetrahedra occur at the highest energies consistent with the relative stability of WO4 tetrahedra. The… ▽ More Lattice dynamics and high pressure phase transitions in AWO4 (A = Ba, Sr, Ca and Pb) have been investigated using inelastic neutron scattering experiments, ab-initio density functional theory calculations and extensive molecular dynamics simulations. The vibrational modes that are internal to WO4 tetrahedra occur at the highest energies consistent with the relative stability of WO4 tetrahedra. The neutron data and the ab-initio calculations are found to be in excellent agreement. The neutron and structural data are used to develop and validate an interatomic potential model. The model is used for classical molecular dynamics simulations to study their response to high pressure. We have calculated the enthalpies of the scheelite and fergusonite phases as a function of pressure, which confirms that the scheelite to fergusonite transition is second order in nature. With increase in pressure, there is a gradual change in the AO8 polyhedra, while there is no apparent change in the WO4 tetrahedra. We found that that all the four tungstates amorphize at high pressure. This is in good agreement with available experimental observations which show amorphization at around 45 GPa in BaWO4 and 40 GPa in CaWO4. On amorphization, there is an abrupt increase in the coordination of the W atom while the bisdisphenoids around A atom are considerably distorted. The pair correlation functions of the various atom pairs corroborate these observations. Our observations aid in predicting the pressure of amorphization in SrWO4 and PbWO4, which have not been experimentally reported. △ Less

Submitted 25 July, 2014; originally announced July 2014.

Comments: 15 pages, 8 figures

Journal ref: Phys. Rev. B 91, 094304 (2015)

Similarity transformed equation of motion coupled cluster theory revisited: a benchmark study of valence excited states

Authors: J. Sous, P. Goel, M. Nooijen

Abstract: The similarity transformed equation of motion coupled cluster (STEOM-CC) method is benchmarked against CC3 and EOM-CCSDT-3 for a large test set of valence excited states of organic molecules studied by Schreiber et al. [M. Schreiber, M.R. Silva-Junior, S.P. Sauer, and W. Thiel, J. Chem. Phys. $\textbf{128}$, 134110 (2008)]. STEOM-CC is found to behave quite satisfactorily and provides significant… ▽ More The similarity transformed equation of motion coupled cluster (STEOM-CC) method is benchmarked against CC3 and EOM-CCSDT-3 for a large test set of valence excited states of organic molecules studied by Schreiber et al. [M. Schreiber, M.R. Silva-Junior, S.P. Sauer, and W. Thiel, J. Chem. Phys. $\textbf{128}$, 134110 (2008)]. STEOM-CC is found to behave quite satisfactorily and provides significant improvement over EOM-CCSD, CASPT2 and NEVPT2 for singlet excited states, lowering standard deviations of these methods by almost a factor of 2. Triplet excited states are found to be described less accurately, however. Besides the parent version of STEOM-CC, additional variations are considered. STEOM-D includes a perturbative correction from doubly excited determinants. The novel STEOM-H ($ω$) approach presents a sophisticated technique to render the STEOM-CC transformed Hamiltonian hermitian. In STEOM-PT, the expensive CCSD step is replaced by many-body second-order perturbation theory (MBPT(2)), while extended STEOM (EXT-STEOM) provides access to doubly excited states. To study orbital invariance in STEOM, we investigate orbital rotation in the STEOM-ORB approach. Comparison of theses variations of STEOM for the large test set provides a comprehensive statistical basis to gauge the usefulness of these approaches. △ Less

Submitted 4 January, 2018; v1 submitted 12 February, 2014; originally announced February 2014.

Comments: Invited article to the Special Issue of Molecular Physics "proceedings of Molecular Quantum Mechanics 2013: electron correlation: the many-body problem at the heart of chemistry" dedicated as a Festschrift for Professor Rodney Bartlett; 49 pages, 10 figures, 11 tables

Journal ref: Mol. Phys. 112, 616 (2014)

Phonons and Thermodynamics of LiMPO4 (M=Mn, Fe)

Authors: Prabhatasree Goel, M. K. Gupta, R. Mittal, S. Rols, S. J. Patwe, S. N. Achary, A. K. Tyagi, S. L. Chaplot

Abstract: Lithium transition metal phospho-olivines are useful electrode materials, owing to their stability, high safety, low cost and cyclability. We report phonon studies using neutron inelastic scattering experiments, ab-initio density functional theory calculations and potential model calculations on LiMPO4 (M=Mn, Fe) at ambient and high temperature to understand the microscopic picture of Li sub-latti… ▽ More Lithium transition metal phospho-olivines are useful electrode materials, owing to their stability, high safety, low cost and cyclability. We report phonon studies using neutron inelastic scattering experiments, ab-initio density functional theory calculations and potential model calculations on LiMPO4 (M=Mn, Fe) at ambient and high temperature to understand the microscopic picture of Li sub-lattice. The experiments are in good agreement with calculations. The lattice dynamics calculations indicate instability of a zone-centre as well as zone-boundary modes along (100) at volume corresponding to high temperature. The unstable phonon modes show mainly large vibration of Li atoms in the x-z plane of the orthorhombic structure (space group Pbnm). Molecular dynamics simulations with increasing temperature indicate large mean square displacement of Li as compared to other constituent atoms. The computed pair-correlations between various atom pairs show that there is local disorder occurring in the lithium sub-lattice with increasing temperature, while other pairs show minimal changes. The results find the two compounds to be thermally stable up to high temperatures, which is a desirable trait for its battery applications. △ Less

Submitted 12 December, 2013; originally announced December 2013.

Comments: 19 Pages, 12 Figures

Journal ref: J. Mater. Chem. A 2, 14729 (2014)

Learning theories reveal loss of pancreatic electrical connectivity in diabetes as an adaptive response

Authors: Pranay Goel, Anita Mehta

Abstract: Cells of almost all solid tissues are connected with gap junctions which permit the direct transfer of ions and small molecules, integral to regulating coordinated function in the tissue. The pancreatic islets of Langerhans are responsible for secreting the hormone insulin in response to glucose stimulation. Gap junctions are the only electrical contacts between the beta-cells in the tissue of the… ▽ More Cells of almost all solid tissues are connected with gap junctions which permit the direct transfer of ions and small molecules, integral to regulating coordinated function in the tissue. The pancreatic islets of Langerhans are responsible for secreting the hormone insulin in response to glucose stimulation. Gap junctions are the only electrical contacts between the beta-cells in the tissue of these excitable islets. It is generally believed that they are responsible for synchrony of the membrane voltage oscillations among beta-cells, and thereby pulsatility of insulin secretion. Most attempts to understand connectivity in islets are often interpreted, bottom-up, in terms of measurements of gap junctional conductance. This does not, however explain systematic changes, such as a diminished junctional conductance in type 2 diabetes. We attempt to address this deficit via the model presented here, which is a learning theory of gap junctional adaptation derived with analogy to neural systems. Here, gap junctions are modelled as bonds in a beta-cell network, that are altered according to homeostatic rules of plasticity. Our analysis reveals that it is nearly impossible to view gap junctions as homogeneous across a tissue. A modified view that accommodates heterogeneity of junction strengths in the islet can explain why, for example, a loss of gap junction conductance in diabetes is necessary for an increase in plasma insulin levels following hyperglycemia. △ Less

Submitted 29 June, 2013; originally announced July 2013.

Comments: 15 pages, 5 figures. To appear in PLoS One (2013)

Journal ref: PLoS ONE 8(8): e70366 (2013)

Behavior of Lithium Oxide at Superionic Transition: First Principles and Molecular Dynamics Studies

Authors: M. K. Gupta, Prabhatasree Goel, R. Mittal, N. Choudhury, S. L. Chaplot

Abstract: We report studies on the vibrational and elastic behavior of lithium oxide, Li2O around its superionic transition temperature. Phonon frequencies calculated using the ab-initio and empirical potential model are in excellent agreement with the reported experimental data. Further, volume dependence of phonon dispersion relation has been calculated, which indicates softening of zone boundary transver… ▽ More We report studies on the vibrational and elastic behavior of lithium oxide, Li2O around its superionic transition temperature. Phonon frequencies calculated using the ab-initio and empirical potential model are in excellent agreement with the reported experimental data. Further, volume dependence of phonon dispersion relation has been calculated, which indicates softening of zone boundary transverse acoustic phonon mode along [110] at volume corresponding to the superionic transition in Li2O. The instability of phonon mode could be a precursor leading to the dynamical disorder of the lithium sub lattice. Empirical potential model calculations have been carried out to deduce the probable direction of lithium diffusion by constructing a super cell consisting of 12000 atoms. The barrier energy for lithium ion diffusion from one lattice site to another at ambient and elevated temperature has been computed. Barrier energy considerations along various symmetry directions indicate that [001] is the most favourable direction for lithium diffusion in the fast ion phase. This result corroborates our observation of dynamical instability in the transverse mode along (110) wave vector. Using molecular dynamics simulations we have studied the temperature variation of elastic constants, which are important to the high-temperature stability of lithium oxide. △ Less

Submitted 28 December, 2011; originally announced December 2011.

Comments: 13 pages, 7 figures

Journal ref: Phys. Rev. B 85, 184304 (2012)

Fast ion diffusion, superionic conductivity and phase transitions of the nuclear materials UO2 and Li2O

Authors: Prabhatasree Goel, N. Choudhury, S. L. Chaplot

Abstract: Lattice dynamics and molecular dynamics studies of the oxides UO2 and Li2O in their normal as well as superionic phase are reported. Lattice dynamics calculations have been carried out using a shell model in the quasiharmonic approximation. The calculated elastic constants, phonon frequencies and specific heat are in good agreement with reported experimental data, which help validate the interat… ▽ More Lattice dynamics and molecular dynamics studies of the oxides UO2 and Li2O in their normal as well as superionic phase are reported. Lattice dynamics calculations have been carried out using a shell model in the quasiharmonic approximation. The calculated elastic constants, phonon frequencies and specific heat are in good agreement with reported experimental data, which help validate the interatomic potentials required for undertaking molecular dynamics simulations. The calculated free energies reveal high pressure fluorite to cottunite phase transitions at 70 GPa for UO2 and anti-fluorite to anti-cotunnite phase transformation at 25 GPa for Li2O, in agreement with reported experiments. Molecular dynamics studies shed important insights into the mechanisms of diffusion and superionic behavior at high temperatures. The calculated superionic transition temperature of Li2O is 1000 K, while that of UO2 is 2300 K. △ Less

Submitted 29 July, 2007; originally announced July 2007.

Comments: 8 pages, 2 Tables, 5 Figures: J. Phys.: Condensed Matter (In Press)

Journal ref: J. Phys.: Condens. Matter 19, 386239 (2007)