MOHAMMADI ZAKI

Followers

Following

Public Views

Interests

Uploads

Papers by MOHAMMADI ZAKI

Isometric Neural Machine Translation using Phoneme Count Ratio Reward-based Reinforcement Learning

arXiv (Cornell University), Mar 20, 2024

Traditional Automatic Video Dubbing (AVD) pipeline consists of three key modules, namely, Automat... more Traditional Automatic Video Dubbing (AVD) pipeline consists of three key modules, namely, Automatic Speech Recognition (ASR), Neural Machine Translation (NMT), and Textto-Speech (TTS). Within AVD pipelines, isometric-NMT algorithms are employed to regulate the length of the synthesized output text. This is done to guarantee synchronization with respect to the alignment of video and audio subsequent to the dubbing process. Previous approaches have focused on aligning the number of characters and words in the source and target language texts of Machine Translation models. However, our approach aims to align the number of phonemes instead, as they are closely associated with speech duration. In this paper, we present the development of an isometric NMT system using Reinforcement Learning (RL), with a focus on optimizing the alignment of phoneme counts in the source and target language sentence pairs. To evaluate our models, we propose the Phoneme Count Compliance (PCC) score, which is a measure of length compliance. Our approach demonstrates a substantial improvement of approximately 36% in the PCC score compared to the state-of-the-art models when applied to English-Hindi language pairs. Moreover, we propose a student-teacher architecture within the framework of our RL approach to maintain a trade-off between the phoneme count and translation quality.

Download

$Research paper thumbnail of Effectiveness of multiscale fractal dimension for improvement of frame classification rate$

Effectiveness of multiscale fractal dimension for improvement of frame classification rate

Zenodo (CERN European Organization for Nuclear Research), Aug 31, 2015

We propose to use multiscale fractal dimension (FD)-based features for phoneme classification tas... more We propose to use multiscale fractal dimension (FD)-based features for phoneme classification task at frame-level. During speech production, turbulence is created and hence vortices (generated due to presence of separated airflow) may travel along the vocal tract and excite vocal tract resonators. This turbulence and in effect, the embedded features of different phoneme classes, can be captured by invariant property of multiscale FD. To capture complementary information, feature-level fusion of proposed feature with state-of-the-art Mel Frequency Cepstral Coefficients (MFCC) is attempted and found to be effective. In particular, single-hidden layer neural nets were trained to compute the frame classification rate. Proposed feature was able to reduce the error rate by over 1.6 % from MFCC features on TIMIT database. This is supported by significant reduction in % EER (i.e., 0.327 % to 4.795 %) 1 .

Download

Quality assessment of voice converted speech using articulatory features

arXiv (Cornell University), Nov 16, 2015

We propose a novel application based on acoustic-toarticulatory inversion (AAI) towards quality a... more We propose a novel application based on acoustic-toarticulatory inversion (AAI) towards quality assessment of voice converted speech. The ability of humans to speak effortlessly requires coordinated movements of various articulators, muscles, etc. This effortless movement contributes towards naturalness, intelligibility and speaker's identity (which is partially present in voice converted speech). Hence, during voice conversion, the information related to speech production is lost. In this paper, this loss is quantified for male voice, by showing increase in RMSE error (upto 12.7 % in tongue tip) for voice converted speech followed by showing decrease in mutual information (I) (by 8.7 %). Similar results are obtained in case of female voice. This observation is extended by showing that articulatory features can be used as an objective measure. The effectiveness of proposed measure over MCD is illustrated by comparing their correlation with Mean Opinion Score (MOS). Moreover, preference score of MCD contradicted ABX test by 100 %, whereas proposed measure supported ABX test by 45.8 % and 16.7 % in case of female-to-male and male-to-female voice conversion systems, respectively.

Download

Actor-Critic based Improper Reinforcement Learning

We consider an improper reinforcement learning setting where a learner is given M base controller... more We consider an improper reinforcement learning setting where a learner is given M base controllers for an unknown Markov decision process, and wishes to combine them optimally to produce a potentially new controller that can outperform each of the base ones. This can be useful in tuning across controllers, learnt possibly in mismatched or simulated environments, to obtain a good controller for a given target environment with relatively few trials. Towards this, we propose two algorithms: (1) a Policy Gradient-based approach; and (2) an algorithm that can switch between a simple Actor-Critic (AC) based scheme and a Natural Actor-Critic (NAC) scheme depending on the available information. Both algorithms operate over a class of improper mixtures of the given controllers. For the first case, we derive convergence rate guarantees assuming access to a gradient oracle. For the AC-based approach we provide convergence rate guarantees to a stationary point in the basic AC case and to a global optimum in the NAC case. Numerical results on (i) the standard control theoretic benchmark of stabilizing an cartpole; and (ii) a constrained queueing task show that our improper policy optimization algorithm can stabilize the system even when the base policies at its disposal are unstable.

Download

Towards Optimal and Efficient Best Arm Identification in Linear Bandits. (arXiv:1911.01695v2 [cs.LG] UPDATED)

arXiv Computer Science, Nov 11, 2019

We give a new algorithm for best arm identification in linearly parameterised bandits in the fixe... more We give a new algorithm for best arm identification in linearly parameterised bandits in the fixed confidence setting. The algorithm generalises the well-known LUCB algorithm of Kalyanakrishnan et al. (2012) by playing an arm which minimises a suitable notion of geometric overlap of the statistical confidence set for the unknown parameter, and is fully adaptive and computationally efficient as compared to several state-of-the methods. We theoretically analyse the sample complexity of the algorithm for problems with two and three arms, showing optimality in many cases. Numerical results indicate favourable performance over other algorithms with which we compare.

Download

Improved Pure Exploration in Linear Bandits with No-Regret Learning

Improper Reinforcement Learning with Gradient-based Policy Optimization

We consider an improper reinforcement learning setting where a learner is given M base controller... more We consider an improper reinforcement learning setting where a learner is given M base controllers for an unknown Markov decision process, and wishes to combine them optimally to produce a potentially new controller that can outperform each of the base ones. This can be useful in tuning across controllers, learnt possibly in mismatched or simulated environments, to obtain a good controller for a given target environment with relatively few trials. We propose a gradient-based approach that operates over a class of improper mixtures of the controllers. We derive convergence rate guarantees for the approach assuming access to a gradient oracle. The value function of the mixture and its gradient may not be available in closed-form; however, we show that we can employ rollouts and simultaneous perturbation stochastic approximation (SPSA) for explicit gradient descent optimization. Numerical results on (i) the standard control theoretic benchmark of stabilizing an inverted pendulum and (i...

Download

Analysis of hierarchical bottleneck framework for improved phoneme recognition

2016 International Conference on Signal Processing and Communications (SPCOM), 2016

In this paper, an attempt is made to examine and evaluate the effect of bottleneck and the hierar... more In this paper, an attempt is made to examine and evaluate the effect of bottleneck and the hierarchical bottleneck (HBN) framework in MLP-based Automatic Speech Recognition (ASR) systems. In particular, the bottleneck and hierarchical bottleneck framework are analyzed using Volterra series. Experiments on several architectures with incorporation of systematic hierarchical and bottleneck properties are done. We obtain significant increase in % Phone Recognition Accuracy (PRA) as compared to traditional cepstral features based Hidden Markov Model (HMM) acoustic modeling to more complex architectures such as the HBN framework. To this extent, the best results are achieved when tandem and acoustic features are combined in a deep HBN framework with 73.72 % PRA on entire TIMIT database. In addition, we observe a relative drop in dependence of language model (LM) on final % PRA with this proposed architecture.

Improper Learning with Gradient-based Policy Optimization

ArXiv, 2021

We consider an improper reinforcement learning setting where the learner is given M base controll... more We consider an improper reinforcement learning setting where the learner is given M base controllers for an unknown Markov Decision Process, and wishes to combine them optimally to produce a potentially new controller that can outperform each of the base ones. We propose a gradient-based approach that operates over a class of improper mixtures of the controllers. The value function of the mixture and its gradient may not be available in closed-form; however, we show that we can employ rollouts and simultaneous perturbation stochastic approximation (SPSA) for explicit gradient descent optimization. We derive convergence and convergence rate guarantees for the approach assuming access to a gradient oracle. Numerical results on a challenging constrained queueing task show that our improper policy optimization algorithm can stabilize the system even when each constituent policy at its disposal is unstable.

S ep 2 01 6 Low-rank Bandits with Latent Mixtures

We study the task of maximizing rewards from recommending it ems (actions) to users sequentially ... more We study the task of maximizing rewards from recommending it ems (actions) to users sequentially interacting with a recommender system. Users are modeled as latent mixtures of C many representative user classes, where each class specifies a mean reward profile across actions. Both the user features (mixture distribution over classes) and the item f atures (mean reward vector per class) are unknown a priori. The user identity is the only contextual in formation available to the learner while interacting. This induces a low-rank structure on the matri x of expected rewards ra,b from recommending itema to userb. The problem reduces to the well-known linear bandit when ei ther useror item-side features are perfectly known. In the setting wh ere each user, with its stochastically sampled taste profile, interacts only for a small number of se ssions, we develop a bandit algorithm for the two-sided uncertainty. It combines the Robust Tenso r Power Method ofAnandkumar et al. (2014b) with th...

Download

Improper Reinforcement Learning with Gradient-based Policy Optimization

We consider an improper reinforcement learning setting where a learner is given M base controller... more We consider an improper reinforcement learning setting where a learner is given M base controllers for an unknown Markov decision process, and wishes to combine them optimally to produce a potentially new controller that can outperform each of the base ones. This can be useful in tuning across controllers, learnt possibly in mismatched or simulated environments, to obtain a good controller for a given target environment with relatively few trials. We propose a gradient-based approach that operates over a class of improper mixtures of the controllers. We derive convergence rate guarantees for the approach assuming access to a gradient oracle. The value function of the mixture and its gradient may not be available in closed-form; however, we show that we can employ rollouts and simultaneous perturbation stochastic approximation (SPSA) for explicit gradient descent optimization. Numerical results on (i) the standard control theoretic benchmark of stabilizing an inverted pendulum and (i...

Download

Towards Optimal and Efficient Best Arm Identification in Linear Bandits

ArXiv, 2019

Download

Explicit Best Arm Identification in Linear Bandits Using No-Regret Learners

ArXiv, 2020

We study the problem of best arm identification in linearly parameterised multi-armed bandits. Gi... more We study the problem of best arm identification in linearly parameterised multi-armed bandits. Given a set of feature vectors $\mathcal{X}\subset\mathbb{R}^d,$ a confidence parameter $\delta$ and an unknown vector $\theta^*,$ the goal is to identify $\arg\max_{x\in\mathcal{X}}x^T\theta^*$, with probability at least $1-\delta,$ using noisy measurements of the form $x^T\theta^*.$ For this fixed confidence ($\delta$-PAC) setting, we propose an explicitly implementable and provably order-optimal sample-complexity algorithm to solve this problem. Previous approaches rely on access to minimax optimization oracles. The algorithm, which we call the \textit{Phased Elimination Linear Exploration Game} (PELEG), maintains a high-probability confidence ellipsoid containing $\theta^*$ in each round and uses it to eliminate suboptimal arms in phases. PELEG achieves fast shrinkage of this confidence ellipsoid along the most confusing (i.e., close to, but not optimal) directions by interpreting the ...

Download

Employing No Regret Learners for Pure Exploration in Linear Bandits

We study the best arm identification problem in linear multi-armed bandits (LMAB) in the fixed co... more We study the best arm identification problem in linear multi-armed bandits (LMAB) in the fixed confidence (δ-PAC) setting; this is also the problem of optimizing an unknown linear function over a discrete ground set with noisy, zeroth-order access. We propose an explicitly implementable and provably order-optimal sample-complexity algorithm to solve this problem. Most previous approaches rely on access to a minimax optimization oracle which is at the heart of the complexity of the problem. We propose a method to solve this optimization problem (upto suitable accuracy) by interpreting the problem as a two-player zero-sum game, and attempting to sequentially converge to its saddle point using low-regret learners to compute the players’ strategies in each round which yields a concrete querying algorithm. The algorithm, which we call the Phased Elimination Linear Exploration Game (PELEG), maintains a high-probability confidence ellipsoid containing θ∗ in each round and uses it to elimin...

Download

Low-rank Bandits with Latent Mixtures

We study the task of maximizing rewards from recommending items (actions) to users sequentially i... more We study the task of maximizing rewards from recommending items (actions) to users sequentially interacting with a recommender system. Users are modeled as latent mixtures of C many representative user classes, where each class specifies a mean reward profile across actions. Both the user features (mixture distribution over classes) and the item features (mean reward vector per class) are unknown a priori. The user identity is the only contextual information available to the learner while interacting. This induces a low-rank structure on the matrix of expected rewards r a,b from recommending item a to user b. The problem reduces to the well-known linear bandit when either user-or item-side features are perfectly known. In the setting where each user, with its stochastically sampled taste profile, interacts only for a small number of sessions, we develop a bandit algorithm for the two-sided uncertainty. It combines the Robust Tensor Power Method of Anandkumar et al. (2014b) with the ...

Download

Quality assessment of voice converted speech using articulatory features

2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2017

Download

$Research paper thumbnail of Effectiveness of multiscale fractal dimension-based phonetic segmentation in speech synthesis for low resource language$

Effectiveness of multiscale fractal dimension-based phonetic segmentation in speech synthesis for low resource language

2014 International Conference on Asian Language Processing (IALP), 2014

ABSTRACT Phonetic segmentation plays a key role in developing various speech applications. In thi... more ABSTRACT Phonetic segmentation plays a key role in developing various speech applications. In this work, we propose to use various features for automatic phonetic segmentation task for forced Viterbi alignment and compare their effectiveness. We propose to use novel multiscale fractal dimension-based features concatenated with Mel- Frequency Cepstral Coefficients (MFCC). The novel features are expected to capture additional nonlinearities in speech production which should improve the performance of segmentation task. However, to evaluate effectiveness of these segmentation algorithms, we require manual accurate phoneme-level labeled data which is not available for low resource languages such as Gujarati (a low resource language and one of the official languages of India). In order to measure effectiveness of various segmentation algorithms, HMM-based speech synthesis system (HTS) for Gujarati have been built. From the subjective and objective evaluations, it is observed that FD-based features for segmentation work moderately better than other state-of- the-art features such as MFCC, Perceptual Linear Prediction Cepstral Coefficients (PLP-CC), Cochlear Filter Cepstral Coefficients (CFCC), and RelAtive SpecTrAl (RASTA)-based PLP-CC. The Mean Opinion Score (MOS) and the Degraded-MOS, which are the measures of naturalness indicate an improvement of 9.69% with the proposed features from the MFCC (which is found to be the best among the other features) based features.

Better than the Best: Gradient-based Improper Reinforcement Learning for Network Scheduling

We consider the problem of scheduling in constrained queueing networks with a view to minimizing ... more We consider the problem of scheduling in constrained queueing networks with a view to minimizing packet delay. We formulate a novel top down approach to scheduling where, given an unknown network and a set of scheduling policies, we use a policy gradient based reinforcement learning algorithm that produces a scheduler that performs better than the available atomic policies. We derive convergence results and analyze finite time performance of the algorithm. Simulation results show that the algorithm performs well even when the arrival rates are nonstationary and can stabilize the system even when the constituent policies are unstable.

Download

$Research paper thumbnail of Effectiveness of fractal dimension for ASR in low resource language$

Effectiveness of fractal dimension for ASR in low resource language

The 9th International Symposium on Chinese Spoken Language Processing, Sep 1, 2014

Influence of various asymmetrical contextual factors for TTS in a low resource language

2014 International Conference on Asian Language Processing, Oct 20, 2014

ABSTRACT The generalized statistical framework of Hidden Markov Model (HMM) has been successfully... more ABSTRACT The generalized statistical framework of Hidden Markov Model (HMM) has been successfully applied from the field of speech recognition to speech synthesis. In this paper, we have applied HMM-based Speech Synthesis (HTS) method to Gujarati (one of the official language of India). Adaption and evaluation of HTS for Gujarati language has been done here. In addition, to understand the influence of asymmetrical contextual factors on quality of synthesized speech, we have conducted series of experiments. Evaluation of different HTS built for Gujarati speech using various asymmetrical contextual factors is done in terms of naturalness and speech intelligibility. From the experimental results, it is evident that when more weightage is given to left phoneme in asymmetrical contextual factor, HTS performance improves compare to conventional symmetrical contextual factors for both triphone and pentaphone case. Furthermore, we achieved best performance for Gujarati HTS with left-left-left-centre-right (i.e., LLLCR) contextual factors.

MOHAMMADI ZAKI

Uploads

Papers by MOHAMMADI ZAKI

Log In