320 Hits in 1.4 sec

Multipath Graph Convolutional Neural Networks [article]

Rangan Das, Bikram Boote, Saumik Bhattacharya, Ujjwal Maulik
2021 arXiv   pre-print
Graph convolution networks have recently garnered a lot of attention for representation learning on non-Euclidean feature spaces. Recent research has focused on stacking multiple layers like in convolutional neural networks for the increased expressive power of graph convolution networks. However, simply stacking multiple graph convolution layers lead to issues like vanishing gradient, over-fitting and over-smoothing. Such problems are much less when using shallower networks, even though the
more » ... llow networks have lower expressive power. In this work, we propose a novel Multipath Graph convolutional neural network that aggregates the output of multiple different shallow networks. We train and test our model on various benchmarks datasets for the task of node property prediction. Results show that the proposed method not only attains increased test accuracy but also requires fewer training epochs to converge. The full implementation is available at
arXiv:2105.01510v1 fatcat:mey2t3k72vegpg44foywvckhqm

Quantum Process Randomness [article]

Sreetama Das, Asutosh Kumar, Aditi Sen De, Ujjwal Sen
2019 arXiv   pre-print
We study the optimization of any quantum process by minimizing the "randomness" in the measurement result at the output of that quantum process. We conceptualize and propose a measure of such randomness and inquire whether an optimization of the quantum process based on that measure, can reach the point where the process operates with maximum fidelity. We consider approximate quantum cloning and teleportation processes, and find, in particular, that the optimal approximate state-dependent
more » ... m cloning machine obtained by maximizing the fidelity is different from that obtained by minimizing the randomness.
arXiv:1903.03564v2 fatcat:g2nowvenvzhlrd45i4cn53iexy

UnShadowNet: Illumination Critic Guided Contrastive Learning For Shadow Removal [article]

Subhrajyoti Dasgupta, Arindam Das, Sudip Das, Andrei Bursuc, Ujjwal Bhattacharya, Senthil Yogamani
2022 arXiv   pre-print
Arindam Das is a Lead Engineer in the department of Detection Vision Systems (DVS) at Valeo India. He is elected as Valeo Expert for 2022 in AI.  ...  Ujjwal Bhattacharya at Computer Vision and Pattern Recognition Unit at Indian Statistical Institute, Kolkata and later continued there as a researcher.  ... 
arXiv:2203.15441v1 fatcat:lkwpmcrygndqlnjtcaotpa7kiq

The separability versus entanglement problem [article]

Sreetama Das, Titas Chanda, Maciej Lewenstein, Anna Sanpera, Aditi Sen De, Ujjwal Sen
2017 arXiv   pre-print
We present a review of the problem of finding out whether a quantum state of two or more parties is entangled or separable. After a formal definition of entangled states, we present a few criteria for identifying entangled states and introduce some entanglement measures. We also provide a classification of entangled states with respect to their usefulness in quantum dense coding, and present some aspects of multipartite entanglement.
arXiv:1701.02187v1 fatcat:dvjfwnbscndkdckanymmkc4f2e

A Study on Academic Anxiety and Academic Achievement

Ujjwal Kumar Halder, Samit Kumar Das, Bapi Mishra
2018 Figshare  
21ST century can be indicated as the century of competition. Everywhere there is a race, a competition. With the huge development in science and technology, all part of the world is connected with this competition. Especially in the field of education, this competition rapidly increased among the students. To survive in this competition, educational achievement is necessary and compulsory to all and for good educational achievement, anxiety free,a positive environment is also essential.
more » ... anxiety is a deliberating factor which impact on students' academic achievement. In the present study, the researchers made an attempt to investigate the gender difference in relation to academic anxiety and academic achievement of the students of secondary level. The researchers also tried to find out the correlation between academic anxiety and academic achievement. A sample of 237 (128 boys and 109 girls) of secondary level students of class VIII were selected randomly. Analyzing the data, the result shows that girls students has more academic anxiety than boys. It was also found that there is a negative and significant correlation (r= -0.10) between academic anxiety and academic achievement.
doi:10.6084/m9.figshare.7194926 fatcat:q2y4ocryincadnf2p25byavhe4

Scale-Invariant Multi-Oriented Text Detection in Wild Scene Images [article]

Kinjal Dasgupta, Sudip Das, Ujjwal Bhattacharya
2020 arXiv   pre-print
Automatic detection of scene texts in the wild is a challenging problem, particularly due to the difficulties in handling (i) occlusions of varying percentages, (ii) widely different scales and orientations, (iii) severe degradations in the image quality etc. In this article, we propose a fully convolutional neural network architecture consisting of a novel Feature Representation Block (FRB) capable of efficient abstraction of information. The proposed network has been trained using curriculum
more » ... earning with respect to difficulties in image samples and gradual pixel-wise blurring. It is capable of detecting texts of different scales and orientations suffered by blurring from multiple possible sources, non-uniform illumination as well as partial occlusions of varying percentages. Text detection performance of the proposed framework on various benchmark sample databases including ICDAR 2015, ICDAR 2017 MLT, COCO-Text and MSRA-TD500 improves respective state-of-the-art results significantly. Source code of the proposed architecture will be made available at github.
arXiv:2002.06423v1 fatcat:thz63jrs6fechbkptxqkf246ii

Localization of a nonlocal interaction

Ujjwal LAHA, Sanjib Kr DAS, Jhasaketan BHOI
2017 Turkish Journal of Physics  
A simple method for the localization of a separable nonlocal interaction is formulated in terms of Green's functions and the solutions with regular and irregular boundary conditions. The constructed energy-momentumdependent local potential with regular boundary condition is real while that for the irregular boundary condition is complex in nature. The phase function method is exploited to compute elastic scattering phases for the nucleon-nucleon and alpha-nucleon systems. Reasonable agreements
more » ... n scattering phase shifts with experimental data are obtained, particularly, in the low energy range for the systems under consideration. The phase shifts for the imaginary parts of the potentials derived from the irregular solutions for the alpha-nucleon systems, however, give indications of resonances at very low energies.
doi:10.3906/fiz-1704-23 fatcat:oe3ylvh3rjev3larimv5thgfhy

Beating detection loophole in nonlinear entanglement witnesses [article]

Kornikar Sen, Sreetama Das, Ujjwal Sen
2019 arXiv   pre-print
Detectors in the laboratory are often unlike their ideal theoretical cousins. They have non-ideal efficiencies, which may then lead to non-trivial implications. We show how it is possible to predict correct answers about whether a shared quantum state is entangled in spite of finite detector efficiencies, when the tool for entanglement detection is a nonlinear entanglement witness. We first consider the detection loophole for shared quantum states with nonpositive partial transpose. We
more » ... tly find nonlinear witness operators for bound entangled states with positive partial transpose, and show how the detection loophole can be closed also in such instances.
arXiv:1906.09046v1 fatcat:ygvqwvo4bnbhxl3kmnlb2dkhyi

An End-to-End Framework for Unsupervised Pose Estimation of Occluded Pedestrians [article]

Sudip Das, Perla Sai Raj Kishore, Ujjwal Bhattacharya
2020 arXiv   pre-print
Pose estimation in the wild is a challenging problem, particularly in situations of (i) occlusions of varying degrees and (ii) crowded outdoor scenes. Most of the existing studies of pose estimation did not report the performance in similar situations. Moreover, pose annotations for occluded parts of human figures have not been provided in any of the relevant standard datasets which in turn creates further difficulties to the required studies for pose estimation of the entire figure of occluded
more » ... humans. Well known pedestrian detection datasets such as CityPersons contains samples of outdoor scenes but it does not include pose annotations. Here, we propose a novel multi-task framework for end-to-end training towards the entire pose estimation of pedestrians including in situations of any kind of occlusion. To tackle this problem for training the network, we make use of a pose estimation dataset, MS-COCO, and employ unsupervised adversarial instance-level domain adaptation for estimating the entire pose of occluded pedestrians. The experimental studies show that the proposed framework outperforms the SOTA results for pose estimation, instance segmentation and pedestrian detection in cases of heavy occlusions (HO) and reasonable + heavy occlusions (R + HO) on the two benchmark datasets.
arXiv:2002.06429v1 fatcat:toyid2t235gxldzoonkhl2jwgu

How efficient is transport of quantum cargo through multiple highways? [article]

Saptarshi Roy, Tamoghna Das, Debmalya Das, Aditi Sen De, Ujjwal Sen
2018 arXiv   pre-print
Quantum states can be efficiently transferred over a long distance if the entire quantum channel can be divided into several small blocks. We consider a scenario in which each block consists of two copies of a multiparty state -- one is used for distributing an arbitrary quantum state to multiple parties while the other channel is required to concentrate it back to a single party. Both in noiseless and local noisy scenarios, we find one-shot quantum capacities of these channels in terms of
more » ... ity, when the initial shared states in each block are the generalized Greenberger-Horne-Zeilinger and the generalized W states. We also consider a situation where optimal local measurements transform multipartite states to bipartite ones which can then be used as single-path channels for quantum state transmission in each segment. We show that in some parameter ranges, the former protocol provides strictly better fidelities than that of the latter, thereby establishing the importance of distributing and concentrating arbitrary quantum states via multipartite entangled states in long distance quantum communication, over the local measurement based protocol. Moreover, we show that in presence of bit flip or bit-phase flip noise, shared generalized Greenberger-Horne-Zeilinger states possess an inherent noise detection and correction mechanism, leading to the same fidelity as in the noiseless case. We consider further noise models also, which do not enjoy the same mechanism.
arXiv:1802.10398v1 fatcat:kwu6oikvxfe4lgt62hnmmlawpe

Deep Multi-Task Networks For Occluded Pedestrian Pose Estimation [article]

Arindam Das, Sudip Das, Ganesh Sistu, Jonathan Horgan, Ujjwal Bhattacharya, Edward Jones, Martin Glavin, Ciarán Eising
2022 arXiv   pre-print
References [Das et al., 2020] Das, S., Kishore, P. S. R., and Bhattacharya, U. ( 2020 ). An end-to-end framework for pose estimation of occluded pedestrians.  ...  ., 2019] and later personto-person occlusion was investigated in [Das et al., 2020] .  ... 
arXiv:2206.07510v2 fatcat:7rfc42dgobfutmgklfjeys432y

Medical Imaging and Diagnosis Using Genetic Algorithms [chapter]

Ujjwal Maulik, Sanghamitra Bandyopadhyay, Sajal Das
2005 Handbook of Bioinspired Algorithms and Applications  
doi:10.1201/9781420035063.ch16 fatcat:pvoom55zjngm7avoegrzgeshgi

Resource theory of quantum coherence with probabilistically non-distinguishable pointers and corresponding wave-particle duality [article]

Chirag Srivastava, Sreetama Das, Ujjwal Sen
2020 arXiv   pre-print
One of the fundamental features of quantum mechanics is the superposition principle, a manifestation of which is embodied in quantum coherence. Coherence of a quantum state is invariably defined with respect to a preferred set of pointer states, and there exist quantum coherence measures with respect to deterministically as well as probabilistically distinguishable sets of quantum state vectors. Here we study the resource theory of quantum coherence with respect to an arbitrary set of quantum
more » ... ate vectors, that may not even be probabilistically distinguishable. Geometrically, a probabilistically indistinguishable set of quantum state vectors forms a linearly dependent set. We find the free states of the resource theory, and analyze the corresponding free operations, obtaining a necessary condition for an arbitrary quantum operation to be free. We identify a class of measures of the quantum coherence, and in particular establish a monotonicity property of the measures. We find a connection of an arbitrary set of quantum state vectors with positive operator valued measurements with respect to the resource theory being considered, which paves the way for an alternate definition of the free states. We notice that the resource theory of magic can be looked upon as a resource theory of quantum coherence with respect to a set of quantum state vectors that are probabilistically indistinguishable. We subsequently examine the wave-particle duality in a double-slit set-up in which superposition of probabilistically indistinguishable quantum state vectors is possible. Specifically, we report a complementary relation between quantum coherence and path distinguishability in such a set-up.
arXiv:2005.08296v1 fatcat:otmjf7fosnfghdhvomo2uhzvlq

Seek and You Will Find: A New Optimized Framework for Efficient Detection of Pedestrian [article]

Sudip Das, Partha Sarathi Mukherjee, Ujjwal Bhattacharya
2019 arXiv   pre-print
, , Das et al.  ... 
arXiv:1912.10241v1 fatcat:6nep6ljzo5fa3p53vtfnku7rvm


Ujjwal Das, Nader Ebrahimi
2018 Statistics in Transition New Series  
Recently, Das and Ebraimi (2017) extended the concept for accelerated failure time model.  ...  DAS, U., EBRAHIMI, N., (2017). Covariate selection for accelerated failure time data, Communications in Statistics: Theory and Methods. 46, pp. 4051-64. EBRAHIMI, N., SOOFI, E. S., SOYER, R., (2010).  ... 
doi:10.21307/stattrans-2018-017 fatcat:yh7rtiqc5jh2jj75c4cy2mmdse
« Previous Showing results 1 — 15 out of 320 results