-
DravidianMultiModality: A Dataset for Multi-modal Sentiment Analysis in Tamil and Malayalam
Authors:
Bharathi Raja Chakravarthi,
Jishnu Parameswaran P. K,
Premjith B,
K. P Soman,
Rahul Ponnusamy,
Prasanna Kumar Kumaresan,
Kingston Pal Thamburaj,
John P. McCrae
Abstract:
Human communication is inherently multimodal and asynchronous. Analyzing human emotions and sentiment is an emerging field of artificial intelligence. We are witnessing an increasing amount of multimodal content in local languages on social media about products and other topics. However, there are not many multimodal resources available for under-resourced Dravidian languages. Our study aims to cr…
▽ More
Human communication is inherently multimodal and asynchronous. Analyzing human emotions and sentiment is an emerging field of artificial intelligence. We are witnessing an increasing amount of multimodal content in local languages on social media about products and other topics. However, there are not many multimodal resources available for under-resourced Dravidian languages. Our study aims to create a multimodal sentiment analysis dataset for the under-resourced Tamil and Malayalam languages. First, we downloaded product or movies review videos from YouTube for Tamil and Malayalam. Next, we created captions for the videos with the help of annotators. Then we labelled the videos for sentiment, and verified the inter-annotator agreement using Fleiss's Kappa. This is the first multimodal sentiment analysis dataset for Tamil and Malayalam by volunteer annotators.
△ Less
Submitted 9 June, 2021;
originally announced June 2021.
-
Ensemble Average of Three-Dimensional Minkowski Tensors of a Gaussian Random Field in Redshift Space
Authors:
Stephen Appleby,
Joby P. K.,
Pravabati Chingangbam,
Changbom Park
Abstract:
We present the ensemble expectation values for the translation invariant, rank-2 Minkowski tensors in three-dimensions, for a linearly redshift space distorted Gaussian random field. The Minkowski tensors $W^{0,2}_{1}$, $W^{0,2}_{2}$ are sensitive to global anisotropic signals present within a field, and by extracting these statistics from the low redshift matter density one can place constraints…
▽ More
We present the ensemble expectation values for the translation invariant, rank-2 Minkowski tensors in three-dimensions, for a linearly redshift space distorted Gaussian random field. The Minkowski tensors $W^{0,2}_{1}$, $W^{0,2}_{2}$ are sensitive to global anisotropic signals present within a field, and by extracting these statistics from the low redshift matter density one can place constraints on the redshift space distortion parameter $β= f/b$. We begin by reviewing the calculation of the ensemble expectation values $\langle W^{0,2}_{1} \rangle$, $\langle W^{0,2}_{2} \rangle $ for isotropic, Gaussian random fields, then consider how these results are modified by the presence of a linearly anisotropic signal. Under the assumption that all fields remain Gaussian, we calculate the anisotropic correction due to redshift space distortion in a coordinate system aligned with the line of sight, finding inequality between the diagonal elements of $\langle W^{0,2}_{1} \rangle $, $\langle W^{0,2}_{2} \rangle $. The ratio of diagonal elements of these matrices provides a set of statistics that are sensitive only to the redshift space distortion parameter $β$. We estimate the Fisher information that can be extracted from the Minkowski tensors, and find $W^{0,2}_{1}$ is more sensitive to $β$ than $W^{0,2}_{2}$, and a measurement of $W^{0,2}_{1}$ accurate to $\sim 1\%$ can yield a $\sim 4\%$ constraint on $β$. Finally, we discuss the difference between using the matrix elements of the Minkowski tensors directly against measuring the eigenvalues. For the purposes of cosmological parameter estimation we advocate the use of the matrix elements, to avoid spurious anisotropic signals that can be generated by the eigenvalue decomposition.
△ Less
Submitted 7 August, 2019;
originally announced August 2019.
-
Search for anomalous alignments of structures in Planck data using Minkowski Tensors
Authors:
Joby P. K.,
Pravabati Chingangbam,
Tuhin Ghosh,
Vidhya Ganesan,
Ravikumar C. D
Abstract:
Minkowski Tensors are tensorial generalizations of the scalar Minkowski Functionals. Due to their tensorial nature they contain additional morphological information of structures, in particular about shape and alignment, in comparison to the scalar Minkowski functionals. They have recently been used [29] to study the statistical isotropy of temperature and E mode data from the Planck satellite. Th…
▽ More
Minkowski Tensors are tensorial generalizations of the scalar Minkowski Functionals. Due to their tensorial nature they contain additional morphological information of structures, in particular about shape and alignment, in comparison to the scalar Minkowski functionals. They have recently been used [29] to study the statistical isotropy of temperature and E mode data from the Planck satellite. The calculation in [29] relied on stereographic projection of the fields to extract the shape and alignment information. In this work, we calculate Minkowski Tensors directly on the sphere and compute the net alignment in the data, based on a recent work that extends the definition of Minkowski Tensors to random fields on curved spaces. This method circumvents numerical errors that can be introduced by the stereographic projection. We compare the resulting net alignment parameter values obtained from the frequency coadded CMB temperature data cleaned by the SMICA pipeline, to those obtained from simulations that include instrumental beam effects and residual foreground and noise. We find very good agreement between the two within $\approx$ 1$σ$ . We further compare the alignments obtained from the beam-convolved CMB maps at individual Planck frequencies to those in the corresponding simulations. We find no significant difference between observed data and simulations across all Planck frequencies, except for the 30 GHz channel. For the 30 GHz channel we find $\approx$ 2$σ$ difference between the data and the simulations. This mild disagreement most likely originates from inaccurate estimation of the instrumental beam at 30 GHz.
△ Less
Submitted 7 July, 2018; v1 submitted 3 July, 2018;
originally announced July 2018.
-
Tensor Minkowski Functionals for random fields on the sphere
Authors:
Pravabati Chingangbam,
K P Yogendran,
Joby P. K.,
Vidhya Ganesan,
Stephen Appleby,
Changbom Park
Abstract:
We generalize the translation invariant tensor-valued Minkowski Functionals which are defined on two-dimensional flat space to the unit sphere. We apply them to level sets of random fields. The contours enclosing boundaries of level sets of random fields give a spatial distribution of random smooth closed curves. We obtain analytic expressions for the ensemble expectation values for the matrix ele…
▽ More
We generalize the translation invariant tensor-valued Minkowski Functionals which are defined on two-dimensional flat space to the unit sphere. We apply them to level sets of random fields. The contours enclosing boundaries of level sets of random fields give a spatial distribution of random smooth closed curves. We obtain analytic expressions for the ensemble expectation values for the matrix elements of the tensor-valued Minkowski Functionals for isotropic Gaussian and Rayleigh fields. We elucidate the way in which the elements of the tensor Minkowski Functionals encode information about the nature and statistical isotropy (or departure from isotropy) of the field. We then implement our method to compute the tensor-valued Minkowski Functionals numerically and demonstrate how they encode statistical anisotropy and departure from Gaussianity by applying the method to maps of the Galactic foreground emissions from the PLANCK data.
△ Less
Submitted 29 November, 2017; v1 submitted 14 July, 2017;
originally announced July 2017.
-
Constraint on noncommutative spacetime from PLANCK data
Authors:
Joby P. K.,
Pravabati Chingangbam,
Subinoy Das
Abstract:
We constrain the energy scale of noncommutativity of spacetime using CMB data from PLANCK. We find that PLANCK data puts the lower bound on the noncommutativity energy scale to about 20 TeV, which is about a factor of two larger than a previous constraint that was obtained using data from WMAP, ACBAR and CBI. We further show that inclusion of data of $E$ mode of CMB polarization will not significa…
▽ More
We constrain the energy scale of noncommutativity of spacetime using CMB data from PLANCK. We find that PLANCK data puts the lower bound on the noncommutativity energy scale to about 20 TeV, which is about a factor of two larger than a previous constraint that was obtained using data from WMAP, ACBAR and CBI. We further show that inclusion of data of $E$ mode of CMB polarization will not significantly change the constraint.
△ Less
Submitted 25 March, 2015; v1 submitted 18 December, 2014;
originally announced December 2014.
