Manraj Singh Grover
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Manraj Singh Grover has a total of 20 co-authors. Here are the top co-authors by number of shared publications:
Rajiv Ratn Shah
3 shared publications- Payman Vafaee
Teachers College Columbia University New York United States
2 shared publications - Rajiv Ratn Shah
Indraprastha Institute of Information Technology Delhi
2 shared publications - Yaman Kumar
Adobe Media and Data Science Research (MDSR) IIIT-Delhi University at Buffalo
2 shared publications Ajit Kumar
1 shared publicationBaani Leen Kaur Jolly
Indraprastha Institute of Information Technology Delhi
1 shared publication
Latest Publications
- Open Access
- 15 March 2022
- Published by Elsevier in International Journal of Artificial Intelligence in Education
- Vol. 33 (1) , 119-154
Automated Speech Scoring System Under The Lens
- 15 March 2022
- Published by Elsevier in International Journal of Artificial Intelligence in Education
- Vol. 33 (1) , 119-154
Open AccessAutomated Speech Scoring System Under The Lens
- Open Access
Multi-modal Automated Speech Scoring using Attention Fusion
In this study, we propose a novel multi-modal end-to-end neural approach for automated assessment of non-native English speakers' spontaneous speech using attention fusion. The pipeline employs Bi-directional Recurrent Convolutional Neural Networks and Bi-directional Long Short-Term Memory Neural Networks to encode acoustic and lexical cues from spectrograms and transcriptions, respectively. Attention fusion is performed on these learned predictive features to learn complex interactions between different modalities before final scoring. We compare our model with strong baselines and find combined attention to both lexical and acoustic cues significantly improves the overall performance of the system. Further, we present a qualitative and quantitative analysis of our model.
Open AccessMulti-modal Automated Speech Scoring using Attention Fusion
In this study, we propose a novel multi-modal end-to-end neural approach for automated assessment of non-native English speakers' spontaneous speech using attention fusion. The pipeline employs Bi-directional Recurrent Convolutional Neural Networks and Bi-directional Long Short-Term Memory Neural Networks to encode acoustic and lexical cues from spectrograms and transcriptions, respectively. Attention fusion is performed on these learned predictive features to learn complex interactions between different modalities before final scoring. We compare our model with strong baselines and find combined attention to both lexical and acoustic cues significantly improves the overall performance of the system. Further, we present a qualitative and quantitative analysis of our model.
- 1 March 2020
- Published by Elsevier in Journal of Parallel and Distributed Computing
- Vol. 137, 65-76
A Hitchhiker’s Guide On Distributed Training Of Deep Neural Networks
- 1 March 2020
- Published by Elsevier in Journal of Parallel and Distributed Computing
- Vol. 137, 65-76
A Hitchhiker’s Guide On Distributed Training Of Deep Neural Networks
- 1 September 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 213-218
Universal EEG Encoder for Learning Diverse Intelligent Tasks
Brain Computer Interfaces (BCI) have become very popular with Electroencephalography (EEG) being one of the most commonly used signal acquisition techniques. A major challenge in BCI studies is the individualistic analysis required for each task. Thus, task-specific feature extraction and classification are performed, which fails to generalize to other tasks with similar time-series EEG input data. To this end, we design a GRU-based universal deep encoding architecture to extract meaningful features from publicly available datasets for five diverse EEG-based classification tasks. Our network can generate task and format-independent data representation and outperform the state of the art EEGNet architecture on most experiments. We also compare our results with CNN-based, and Autoencoder networks, in turn performing local, spatial, temporal and unsupervised analysis on the data.
- 1 September 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 213-218
Universal EEG Encoder for Learning Diverse Intelligent Tasks
Brain Computer Interfaces (BCI) have become very popular with Electroencephalography (EEG) being one of the most commonly used signal acquisition techniques. A major challenge in BCI studies is the individualistic analysis required for each task. Thus, task-specific feature extraction and classification are performed, which fails to generalize to other tasks with similar time-series EEG input data. To this end, we design a GRU-based universal deep encoding architecture to extract meaningful features from publicly available datasets for five diverse EEG-based classification tasks. Our network can generate task and format-independent data representation and outperform the state of the art EEGNet architecture on most experiments. We also compare our results with CNN-based, and Autoencoder networks, in turn performing local, spatial, temporal and unsupervised analysis on the data.
- 1 September 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 58-67
Text2FaceGAN: Face Generation from Fine Grained Textual Descriptions
- 1 September 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 58-67
Text2FaceGAN: Face Generation from Fine Grained Textual Descriptions
- 1 August 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-7
Generating word clusters by graph clustering based on hearst patterns
The process of clustering similar words is crucial for a broad range of applications such as text classification and word sense disambiguation. Several approaches for deriving word similarity have been proposed. Some, like latent semantic analysis, are derived from the distributional hypothesis. Others extract relationships between terms by drawing upon predefined linguistic patterns. In this work, we propose an innovative approach which combines the essence of both these approaches. In the first phase, our algorithm generates a graphical model of terms and their interrelations with the help of special lexico-syntactic patterns called Hearst Patterns. We then apply a graph clustering technique to find semantically related words.
- 1 August 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-7
Generating word clusters by graph clustering based on hearst patterns
The process of clustering similar words is crucial for a broad range of applications such as text classification and word sense disambiguation. Several approaches for deriving word similarity have been proposed. Some, like latent semantic analysis, are derived from the distributional hypothesis. Others extract relationships between terms by drawing upon predefined linguistic patterns. In this work, we propose an innovative approach which combines the essence of both these approaches. In the first phase, our algorithm generates a graphical model of terms and their interrelations with the help of special lexico-syntactic patterns called Hearst Patterns. We then apply a graph clustering technique to find semantically related words.