-
Smart Speech Segmentation using Acousto-Linguistic Features with look-ahead
Authors:
Piyush Behre,
Naveen Parihar,
Sharman Tan,
Amy Shah,
Eva Sharma,
Geoffrey Liu,
Shuangyu Chang,
Hosam Khalil,
Chris Basoglu,
Sayan Pathak
Abstract:
Segmentation for continuous Automatic Speech Recognition (ASR) has traditionally used silence timeouts or voice activity detectors (VADs), which are both limited to acoustic features. This segmentation is often overly aggressive, given that people naturally pause to think as they speak. Consequently, segmentation happens mid-sentence, hindering both punctuation and downstream tasks like machine tr…
▽ More
Segmentation for continuous Automatic Speech Recognition (ASR) has traditionally used silence timeouts or voice activity detectors (VADs), which are both limited to acoustic features. This segmentation is often overly aggressive, given that people naturally pause to think as they speak. Consequently, segmentation happens mid-sentence, hindering both punctuation and downstream tasks like machine translation for which high-quality segmentation is critical. Model-based segmentation methods that leverage acoustic features are powerful, but without an understanding of the language itself, these approaches are limited. We present a hybrid approach that leverages both acoustic and language information to improve segmentation. Furthermore, we show that including one word as a look-ahead boosts segmentation quality. On average, our models improve segmentation-F0.5 score by 9.8% over baseline. We show that this approach works for multiple languages. For the downstream task of machine translation, it improves the translation BLEU score by an average of 1.05 points.
△ Less
Submitted 27 October, 2022; v1 submitted 25 October, 2022;
originally announced October 2022.
-
FD Cell-Free mMIMO: Analysis and Optimization
Authors:
Soumyadeep Datta,
Ekant Sharma,
Dheeraj Naidu Amudala,
Rohit Budhiraja,
Shivendra S. Panwar
Abstract:
Cell-free (CF) massive multiple-input-multiple-output (mMIMO) deployments are usually investigated with half-duplex nodes and high-capacity fronthaul links. To leverage the possible gains in throughput and energy efficiency (EE) of full-duplex (FD) communications, we consider a FD CF mMIMO system with practical limited-capacity fronthaul links. We derive closed-form spectral efficiency (SE) lower…
▽ More
Cell-free (CF) massive multiple-input-multiple-output (mMIMO) deployments are usually investigated with half-duplex nodes and high-capacity fronthaul links. To leverage the possible gains in throughput and energy efficiency (EE) of full-duplex (FD) communications, we consider a FD CF mMIMO system with practical limited-capacity fronthaul links. We derive closed-form spectral efficiency (SE) lower bounds for this system with maximum-ratio combining/maximum-ratio transmission processing and optimal uniform quantization. We then optimize the weighted sum EE (WSEE) via downlink and uplink power control by using a {two-layered} approach: the first layer formulates the optimization as a generalized convex program, while the second layer solves the optimization decentrally using alternating direction method of multipliers. We analytically show that the proposed two-layered formulation yields a Karush-Kuhn-Tucker point of the original WSEE optimization. We numerically show the influence of weights on the individual EE of the users, which demonstrates the utility of WSEE metric to incorporate heterogeneous EE requirements of users. We show that the low fronthaul capacity reduces the number of users each AP can support, and the cell-free system, consequently, becomes user-centric.
△ Less
Submitted 31 May, 2021; v1 submitted 27 October, 2020;
originally announced October 2020.
-
Full-Duplex Cell-Free mMIMO Systems: Analysis and Decentralized Optimization
Authors:
Soumyadeep Datta,
Dheeraj Naidu Amudala,
Ekant Sharma,
Rohit Budhiraja,
Shivendra S. Panwar
Abstract:
Cell-free (CF) massive multiple-input-multiple-output (mMIMO) deployments are usually investigated with half-duplex nodes and high-capacity fronthaul links. To leverage the possible gains in throughput and energy efficiency (EE) of full-duplex (FD) communications, we consider a FD CF mMIMO system with practical limited-capacity fronthaul links. We derive closed-form spectral efficiency (SE) lower…
▽ More
Cell-free (CF) massive multiple-input-multiple-output (mMIMO) deployments are usually investigated with half-duplex nodes and high-capacity fronthaul links. To leverage the possible gains in throughput and energy efficiency (EE) of full-duplex (FD) communications, we consider a FD CF mMIMO system with practical limited-capacity fronthaul links. We derive closed-form spectral efficiency (SE) lower bounds for this system with maximum-ratio combining/maximum-ratio transmission processing and optimal uniform quantization. We then optimize the weighted sum EE (WSEE) via downlink and uplink power control by using a two-layered approach: the first layer formulates the optimization as a generalized convex program, while the {second layer} solves the optimization decentrally using the alternating direction method of multipliers. We analytically show that the proposed two-layered formulation yields a Karush-Kuhn-Tucker point of the original WSEE optimization. We numerically show the influence of weights on the individual EE of the users, which demonstrates the utility of the WSEE metric to incorporate heterogeneous EE requirements of users. We show that low fronthaul capacity reduces the number of users each AP can support, and the cell-free system, consequently, becomes user-centric.
△ Less
Submitted 10 December, 2021; v1 submitted 27 October, 2020;
originally announced October 2020.
-
Classifying Songs with EEG
Authors:
Prashant Lawhatre,
Bharatesh R Shiraguppi,
Esha Sharma,
Krishna Prasad Miyapuram,
Derek Lomas
Abstract:
This research study aims to use machine learning methods to characterize the EEG response to music. Specifically, we investigate how resonance in the EEG response correlates with individual aesthetic enjoyment. Inspired by the notion of musical processing as resonance, we hypothesize that the intensity of an aesthetic experience is based on the degree to which a participants EEG entrains to the pe…
▽ More
This research study aims to use machine learning methods to characterize the EEG response to music. Specifically, we investigate how resonance in the EEG response correlates with individual aesthetic enjoyment. Inspired by the notion of musical processing as resonance, we hypothesize that the intensity of an aesthetic experience is based on the degree to which a participants EEG entrains to the perceptual input. To test this and other hypotheses, we have built an EEG dataset from 20 subjects listening to 12 two minute-long songs in random order. After preprocessing and feature construction, we used this dataset to train and test multiple machine learning models.
△ Less
Submitted 1 October, 2020;
originally announced October 2020.
-
Smart grid modeling and simulation - Comparing GridLAB-D and RAPSim via two Case studies
Authors:
Midhat Jdeed,
Ekanki Sharma,
Wilfried Elmenreich
Abstract:
One of the most important tools for the development of the smart grid is simulation. Therefore, analyzing, designing, modeling, and simulating the smart grid will allow to explore future scenarios and support decision making for the grid's development. In this paper, we compare two open source simulation tools for the smart grid, GridLAB-Distribution (GridLAB-D) and Renewable Alternative Power sys…
▽ More
One of the most important tools for the development of the smart grid is simulation. Therefore, analyzing, designing, modeling, and simulating the smart grid will allow to explore future scenarios and support decision making for the grid's development. In this paper, we compare two open source simulation tools for the smart grid, GridLAB-Distribution (GridLAB-D) and Renewable Alternative Power systems Simulation (RAPSim). The comparison is based on the implementation of two case studies related to a power flow problem and the integration of renewable energy resources to the grid. Results show that even for very simple case studies, specific properties such as weather simulation or load modeling are influencing the results in a way that they are not reproducible with a different simulator.
△ Less
Submitted 19 September, 2018;
originally announced September 2018.
