BZPEERReading, trust and collaboration for research
Menu

Discover

GraphFollow connections around a paper, topic or saved view.

Workspaces

HomePick up where your research flow left off.InboxHandle requests, replies and notifications from one place.CollectionsCurate reading lists, evidence sets and shareable dossiers.ResearchSee your private provenance graph, trail and imports.CollaborationMove from discovery into focused discussion rooms.PublishDraft, preview and publish full-text research articles.

Network

ExpertsFind specialists worth following or asking for help.CommunitiesJoin research circles organized around shared work.PartnersSee external organizations active around the same topics.

Opportunities

ListingsBrowse roles, calls and collaborations with clearer fit signals.

Account

Log inReturn to your reading and collaboration workspace.Create accountStart saving work, following people and joining teams.
Log inCreate account

Researcher profile

Pooja Rani

Pooja Rani contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate
cond-mat.mes-hallcond-mat.mtrl-sciArtificial IntelligenceComputation and LanguageMachine Learning

Trust snapshot

Quick read

Trust 19 - UnverifiedVerification L1Unclaimed author
5works
0followers
5topics
4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Identity and collaboration

How to connect with this researcher

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Log in to claim

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Research graph

See the researcher in context

Open full explorer

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Published work

5 published item(s)

preprint2024arXiv

LEXTREME: A Multi-Lingual and Multi-Task Benchmark for the Legal Domain

Lately, propelled by the phenomenal advances around the transformer architecture, the legal NLP field has enjoyed spectacular growth. To measure progress, well curated and challenging benchmarks are crucial. However, most benchmarks are English only and in legal NLP specifically there is no multilingual benchmark available yet. Additionally, many benchmarks are saturated, with the best models clearly outperforming the best humans and achieving near perfect scores. We survey the legal NLP literature and select 11 datasets covering 24 languages, creating LEXTREME. To provide a fair comparison, we propose two aggregate scores, one based on the datasets and one on the languages. The best baseline (XLM-R large) achieves both a dataset aggregate score a language aggregate score of 61.3. This indicates that LEXTREME is still very challenging and leaves ample room for improvement. To make it easy for researchers and practitioners to use, we release LEXTREME on huggingface together with all the code required to evaluate models and a public Weights and Biases project with all the runs.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

On the Variability of Critical Size for Homogeneous Nucleation in a Solid-State Diffusional Transformation

In a solid-state diffusional phase transformation involving nucleation and growth, the size of the critical nucleus for a homogeneous process (r*homo=r* ) has been assumed to be a time invariant constant of the transformation. The strain associated with the process has a positive energy contribution and leads to an increase in the value of r*, with respect to that for nucleation from a liquid. With the progress of such a transformation, the strain energy stored in the matrix increases and nuclei forming at a later stage encounter a strained matrix. Using devitrification of a bulk metallic glass as a model system, we demonstrate that r* is not a cardinal time invariant constant for homogeneous nucleation and can increase or decrease depending on the strain energy penalty. We show that the assumption regarding the constancy of r* is true only in the early stages of the transformation and establish that the progress of the transformation leads to an altered magnitude of r*, which is a function of the microstructural details, geometrical variables and physical parameters. With the aid of high-resolution lattice fringe imaging and computations of r*, we further argue that, 'liquid-like' homogeneous nucleation can occur and that the conclusions are applicable to a broad set of solid-state diffusional transformations. The above effect 'opens up' a lower barrier transformation pathway arising purely from the internal variables of the system.

0 citations0 reviewsNo code linkedOpen access
preprint2016arXiv

Thermodynamic properties of pure and doped (B, N) graphene

Ab-initio density functional perturbation theory (DFPT) has been employed to study thermodynamical properties of pure and doped graphene sheet and the results have been compared with available theoretical and experimental data. The concentration of B and N has been varied upto 50% of the carbon atoms in graphene. Phonon frequencies are essential ingredients entering into such a calculation, which have been computed by using the dynamical matrix provided by VASP software in combination with phonopy code in the harmonic approximation. This easily provides us the Helmholtz free energy and leads us to numerical estimates of various thermodynamical properties. The results for specific heat are in good agreement with various theoretical and experimental studies obtained earlier for pure graphene. Interesting new results have been reported for B and N substituted structure. It has been observed that specific heat decreases with the increase in concentration of doping while the entropy increases. Further, large doping concentrations result in unstable sheets resulting in imaginary frequencies in the transverse directions. The instability needs to be compensated by external strains and that has been assumed while carrying out Brilluoin summations. These results will be useful for calculation of thermal conductivity of doped graphene and thus feasibility of using these for device applications. Our preliminary results for thermal expansion have indicated negative thermal expansion behavior of pure graphene at low temperatures which needs investigation on doped graphene as well.

0 citations0 reviewsNo code linkedOpen access
preprint2014arXiv

DFT study of optical properties of pure and doped Graphene

Ab-initio calculations based on density functional theory (DFT) have been performed to study the optical properties of pure graphene and have been compared to that of individual boron (B), nitrogen (N) and BN co-doped graphene sheet. The effect of doping has been investigated by varying the concentrations of dopants from 3.125 % (one atom of the dopant in 32 host atoms) to 6.25 % (six dopant atoms in 50 host atoms) for individual B and N doping and from 37.5 % (one B/N pair in 32 host atoms) to 18.75 % for BN co-doping. Positions of the dopants have also been varied for the same concentration of substitution doping. The dielectric matrix has been calculated within the random phase approximation (RPA) using VASP (Vienna ab-initio Simulation Package) code. The dielectric function, absorption spectrum and energy loss-function of single layer graphene sheet have been calculated for light polarization parallel and perpendicular to the plane of graphene sheet and compared with doping graphene. The calculated dielectric functions and energy-loss spectra are in reasonable agreement with the available theoretical and experimental results for pure graphene. It has been found that individual B and N doping does not significantly affect the imaginary dielectric function and hence the absorption spectra. However, significant red shift in absorption towards visible range of the radiation at high doping is found to occur for the B/N co-doping. The results can be used to tailor the optical properties of graphene in visible region.

0 citations0 reviewsNo code linkedOpen access
preprint2012arXiv

Designing band gap of graphene by B and N dopant atoms

Ab-initio calculations have been performed to study the geometry and electronic structure of boron (B) and nitrogen (N) doped graphene sheet. The effect of doping has been investigated by varying the concentrations of dopants from 2 % (one atom of the dopant in 50 host atoms) to 12 % (six dopant atoms in 50 atoms host atoms) and also by considering different doping sites for the same concentration of substitutional doping. All the calculations have been performed by using VASP (Vienna Ab-initio Simulation Package) based on density functional theory. By B and N doping p-type and n-type doping is induced respectively in the graphene sheet. While the planar structure of the graphene sheet remains unaffected on doping, the electronic properties change from semimetal to semiconductor with increasing number of dopants. It has been observed that isomers formed differ significantly in the stability, bond length and band gap introduced. The band gap is maximum when dopants are placed at same sublattice points of graphene due to combined effect of symmetry breaking of sub lattices and the band gap is closed when dopants are placed at adjacent positions (alternate sublattice positions). These interesting results provide the possibility of tuning the band gap of graphene as required and its application in electronic devices such as replacements to Pt based catalysts in Polymer Electrolytic Fuel Cell (PEFC).

0 citations0 reviewsNo code linkedOpen access