-
Collective Learning Mechanism based Optimal Transport Generative Adversarial Network for Non-parallel Voice Conversion
Authors:
Sandipan Dhar,
Md. Tousin Akhter,
Nanda Dulal Jana,
Swagatam Das
Abstract:
After demonstrating significant success in image synthesis, Generative Adversarial Network (GAN) models have likewise made significant progress in the field of speech synthesis, leveraging their capacity to adapt the precise distribution of target data through adversarial learning processes. Notably, in the realm of State-Of-The-Art (SOTA) GAN-based Voice Conversion (VC) models, there exists a sub…
▽ More
After demonstrating significant success in image synthesis, Generative Adversarial Network (GAN) models have likewise made significant progress in the field of speech synthesis, leveraging their capacity to adapt the precise distribution of target data through adversarial learning processes. Notably, in the realm of State-Of-The-Art (SOTA) GAN-based Voice Conversion (VC) models, there exists a substantial disparity in naturalness between real and GAN-generated speech samples. Furthermore, while many GAN models currently operate on a single generator discriminator learning approach, optimizing target data distribution is more effectively achievable through a single generator multi-discriminator learning scheme. Hence, this study introduces a novel GAN model named Collective Learning Mechanism-based Optimal Transport GAN (CLOT-GAN) model, incorporating multiple discriminators, including the Deep Convolutional Neural Network (DCNN) model, Vision Transformer (ViT), and conformer. The objective of integrating various discriminators lies in their ability to comprehend the formant distribution of mel-spectrograms, facilitated by a collective learning mechanism. Simultaneously, the inclusion of Optimal Transport (OT) loss aims to precisely bridge the gap between the source and target data distribution, employing the principles of OT theory. The experimental validation on VCC 2018, VCTK, and CMU-Arctic datasets confirms that the CLOT-GAN-VC model outperforms existing VC models in objective and subjective assessments.
△ Less
Submitted 18 April, 2025;
originally announced April 2025.
-
Identification of Traditional Medicinal Plant Leaves Using an effective Deep Learning model and Self-Curated Dataset
Authors:
Deepjyoti Chetia,
Sanjib Kr Kalita,
Prof Partha Pratim Baruah,
Debasish Dutta,
Tanaz Akhter
Abstract:
Medicinal plants have been a key component in producing traditional and modern medicines, especially in the field of Ayurveda, an ancient Indian medical system. Producing these medicines and collecting and extracting the right plant is a crucial step due to the visually similar nature of some plants. The extraction of these plants from nonmedicinal plants requires human expert intervention. To sol…
▽ More
Medicinal plants have been a key component in producing traditional and modern medicines, especially in the field of Ayurveda, an ancient Indian medical system. Producing these medicines and collecting and extracting the right plant is a crucial step due to the visually similar nature of some plants. The extraction of these plants from nonmedicinal plants requires human expert intervention. To solve the issue of accurate plant identification and reduce the need for a human expert in the collection process; employing computer vision methods will be efficient and beneficial. In this paper, we have proposed a model that solves such issues. The proposed model is a custom convolutional neural network (CNN) architecture with 6 convolution layers, max-pooling layers, and dense layers. The model was tested on three different datasets named Indian Medicinal Leaves Image Dataset,MED117 Medicinal Plant Leaf Dataset, and the self-curated dataset by the authors. The proposed model achieved respective accuracies of 99.5%, 98.4%, and 99.7% using various optimizers including Adam, RMSprop, and SGD with momentum.
△ Less
Submitted 16 January, 2025;
originally announced January 2025.
-
On the impact of short laser pulses on cold diluted plasmas
Authors:
Gaetano Fiore,
Sergio De Nicola,
Tahmina Akhter,
Renato Fedele,
Dušan Jovanović
Abstract:
We analytically study the impact of a short laser pulse onto an inhomogeneous cold diluted plasma at rest, in particular: the duration of the hydrodynamic regime; the formation and the features of plasma waves (PWs); their wave-breakings (WBs); the motion of test electrons injected in the PWs.
If the pulse is a plane wave travelling in the $z$-direction, and the initial plasma density (IPD) depe…
▽ More
We analytically study the impact of a short laser pulse onto an inhomogeneous cold diluted plasma at rest, in particular: the duration of the hydrodynamic regime; the formation and the features of plasma waves (PWs); their wave-breakings (WBs); the motion of test electrons injected in the PWs.
If the pulse is a plane wave travelling in the $z$-direction, and the initial plasma density (IPD) depends only on $z$, then suitable matched bounds on the maximum and relative variations of the IPD, as well as the intensity and duration of the pulse, ensure a strictly hydrodynamic evolution of the electron fluid during its whole interaction with the pulse, while ions can be regarded as immobile. This evolution is ruled by a family (parametrized by $Z\ge 0$) of decoupled systems of non-autonomous Hamilton equations with 1 degree of freedom, which determine how electrons initially located in the layer $Z\le z<Z+dZ$ move; $ξ=ct-z$ replaces time $t$ as the independent variable. This family of ODEs is obtained by reduction from the Lorentz-Maxwell and continuity PDEs for the electrons' fluid within the spacetime region where the change of the pulse is negligible. After the laser-plasma interaction the Jacobian of the map from Lagrangian to Eulerian coordinates is linear-quasi-periodic in $ξ$. We determine spacetime locations and features of the first wave-breakings of the wakefield PWs, the motion of test electrons (self-)injected in the PWs. The energy of those trapped in a single PW trough grows linearly with the distance gone, where the IPD is constant.
If the pulse has cylindrical symmetry and a not too small radius, the same conclusions hold for the part of the plasma enclosed within the causal cone swept by it.
This computationally light approach may help in a preliminary study of extreme acceleration mechanisms of electrons (LWFA, etc.), before 2D or 3D PIC simulations.
△ Less
Submitted 9 August, 2023; v1 submitted 6 March, 2023;
originally announced March 2023.
-
A gravitational-wave limit on the Chandrasekhar mass of dark matter
Authors:
Divya Singh,
Michael Ryan,
Ryan Magee,
Towsifa Akhter,
Sarah Shandera,
Donghui Jeong,
Chad Hanna
Abstract:
We explore a new paradigm to study dissipative dark matter models using gravitational-wave observations. We consider a dark atomic model which predicts the formation of binary black holes such as GW190425 while obeying constraints from large-scale structure, and improving on the missing satellite problem. Using LIGO and Virgo gravitational-wave data from 12th September 2015 to 1st October 2019, we…
▽ More
We explore a new paradigm to study dissipative dark matter models using gravitational-wave observations. We consider a dark atomic model which predicts the formation of binary black holes such as GW190425 while obeying constraints from large-scale structure, and improving on the missing satellite problem. Using LIGO and Virgo gravitational-wave data from 12th September 2015 to 1st October 2019, we show that interpreting GW190425 as a dark matter black-hole binary limits the Chandrasekhar mass for dark matter to be below 1.4 $M_\odot$ at $> 99.9\%$ confidence implying that the dark proton is heavier than 0.95 GeV, while also suggesting that the molecular energy-level spacing of dark molecules lies near $10^{-3}$ eV and constraining the cooling rate of dark matter at low temperatures.
△ Less
Submitted 3 June, 2021; v1 submitted 10 September, 2020;
originally announced September 2020.
-
Self consistent hydrodynamic description of the plasma wake field excitation induced by a relativistic charged-particle beam in an unmagnetized plasma
Authors:
Dušan Jovanović,
Renato Fedele,
Sergio De Nicola,
Tamina Akhter,
Milivoj Belić
Abstract:
A self-consistent nonlinear hydrodynamic theory is presented of the propagation of a long and thin relativistic electron beam, for a typical plasma wake field acceleration configuration in an unmagnetized and overdense plasma. The random component of the trajectories of the beam particles as well as of their velocity spread is modelled by an anisotropic temperature, allowing the beam dynamics to b…
▽ More
A self-consistent nonlinear hydrodynamic theory is presented of the propagation of a long and thin relativistic electron beam, for a typical plasma wake field acceleration configuration in an unmagnetized and overdense plasma. The random component of the trajectories of the beam particles as well as of their velocity spread is modelled by an anisotropic temperature, allowing the beam dynamics to be approximated as a 3-D adiabatic expansion/compression. {It is shown that even in the absence of the {nonlinear plasma wake force}, the localization of the beam in the transverse direction can be achieved owing to the} nonlinearity associated with the adiabatic compression/rarefaction and a coherent stationary state is constructed. Numerical calculations reveal the possibility of the beam focussing and defocussing, but the lifetime of the beam can be significantly extended by the appropriate adjustments, so that transverse oscillations are observed, {similar to those} predicted within the thermal wave and Vlasov kinetic models.
△ Less
Submitted 30 June, 2017;
originally announced June 2017.
-
Formation and stability of a hollow electron beam in the presence of a plasma wake field driven by an ultra-short electron bunch
Authors:
F. Tanjia,
R. Fedele,
S. De Nicola,
T. Akhter,
D. Jovanovic
Abstract:
A numerical investigation on the spatiotemporal evolution of an electron beam, externally injected in a plasma in the presence of a plasma wake field, is carried out. The latter is driven by an ultra-short relativistic axially-symmetric femtosecond electron bunch. We first derive a novel Poisson-like equation for the wake potential where the driving term is the ultra-short bunch density, taking su…
▽ More
A numerical investigation on the spatiotemporal evolution of an electron beam, externally injected in a plasma in the presence of a plasma wake field, is carried out. The latter is driven by an ultra-short relativistic axially-symmetric femtosecond electron bunch. We first derive a novel Poisson-like equation for the wake potential where the driving term is the ultra-short bunch density, taking suitably into account the interplay between the sharpness and high energy of the bunch. Then, we show that a channel is formed longitudinally, through the externally injected beam while experiencing the effects of the bunch-driven plasma wake field, within the context of thermal wave model. The formation of the channel seems to be a final stage of the 3D evolution of the beam. This involves the appearance of small filaments and bubbles around the longitudinal axis. The bubbles coalesce forming a relatively stable axially-symmetric hollow beam structure.
△ Less
Submitted 25 November, 2015;
originally announced November 2015.
-
Self modulated dynamics of a relativistic charged particle beam in plasma wake field excitation
Authors:
T. Akhter,
R. Fedele,
S. De Nicola,
F. Tanjia,
D. Jovanovic,
A. Mannan
Abstract:
Self modulated dynamics of a relativistic charged particle beam is reviewed within the context of the theory of plasma wake field excitation. The self-consistent description of the beam dynamics is provided by coupling the Vlasov equation with a Poisson-type equation relating the plasma wake potential to the beam density. An analysis of the beam envelope self-modulation is then carried out and the…
▽ More
Self modulated dynamics of a relativistic charged particle beam is reviewed within the context of the theory of plasma wake field excitation. The self-consistent description of the beam dynamics is provided by coupling the Vlasov equation with a Poisson-type equation relating the plasma wake potential to the beam density. An analysis of the beam envelope self-modulation is then carried out and the criteria for the occurrence of the instability are discussed thereby.
△ Less
Submitted 24 November, 2015;
originally announced November 2015.
