Showing 1–2 of 2 results for author: Malkovsky, N

Techniques for Vocabulary Expansion in Hybrid Speech Recognition Systems

Authors: Nikolay Malkovsky, Vladimir Bataev, Dmitrii Sviridkin, Natalia Kizhaeva, Aleksandr Laptev, Ildar Valiev, Oleg Petrov

Abstract: The problem of out of vocabulary words (OOV) is typical for any speech recognition system, hybrid systems are usually constructed to recognize a fixed set of words and rarely can include all the words that will be encountered during exploitation of the system. One of the popular approach to cover OOVs is to use subword units rather then words. Such system can potentially recognize any previously u… ▽ More The problem of out of vocabulary words (OOV) is typical for any speech recognition system, hybrid systems are usually constructed to recognize a fixed set of words and rarely can include all the words that will be encountered during exploitation of the system. One of the popular approach to cover OOVs is to use subword units rather then words. Such system can potentially recognize any previously unseen word if the word can be constructed from present subword units, but also non-existing words can be recognized. The other popular approach is to modify HMM part of the system so that it can be easily and effectively expanded with custom set of words we want to add to the system. In this paper we explore different existing methods of this solution on both graph construction and search method levels. We also present a novel vocabulary expansion techniques which solve some common internal subroutine problems regarding recognition graph processing. △ Less

Submitted 19 March, 2020; originally announced March 2020.

Comments: Submitted to Interspeech 2020

Optimal Distributed Optimization on Slowly Time-Varying Graphs

Authors: Alexander Rogozin, César A. Uribe, Alexander Gasnikov, Nikolay Malkovsky, Angelia Nedić

Abstract: We study optimal distributed first-order optimization algorithms when the network (i.e., communication constraints between the agents) changes with time. This problem is motivated by scenarios where agents experience network malfunctions. We provide a sufficient condition that guarantees a convergence rate with optimal (up lo logarithmic terms) dependencies on the network and function parameters i… ▽ More We study optimal distributed first-order optimization algorithms when the network (i.e., communication constraints between the agents) changes with time. This problem is motivated by scenarios where agents experience network malfunctions. We provide a sufficient condition that guarantees a convergence rate with optimal (up lo logarithmic terms) dependencies on the network and function parameters if the network changes are constrained to a small percentage $α$ of the total number of iterations. We call such networks slowly time-varying networks. Moreover, we show that Nesterov's method has an iteration complexity of $Ω\big( \big(\sqrt{κ_Φ\cdot \barχ} + α\log(κ_Φ\cdot \barχ)\big) \log(1 / \varepsilon)\big)$ for decentralized algorithms, where $κ_Φ$ is condition number of the objective function, and $\barχ$ is a worst case bound on the condition number of the sequence of communication graphs. Additionally, we provide an explicit upper bound on $α$ in terms of the condition number of the objective function and network topologies. △ Less

Submitted 28 November, 2019; v1 submitted 15 May, 2018; originally announced May 2018.