Speech keyword spotting system is a critical component of human-computer interfaces. And connectionist temporal classifier (CTC) has been proven to be an effective tool for that task. However, the standard training process speech faces data imbalance issue where positive samples are usually far less than negative samples. Numerous easy-training examples overwhelm training, resulting in degenera...

In the era of Internet Things (IoT), voice control has enhanced human–machine interaction and accuracy keyword spotting (KWS) algorithms reached 97%; however, high power consumption KWS caused by their huge computing storage requirements limited application in Artificial Intelligence (AIoT) devices. this study, features are extracted utilizing fast discrete cosine transform (FDCT) for frequency...

The word-spotting task is analogous to text-based informarion retrieval tasks and message-understanding tasks in that an exhaustive accounting of the input is not required: only a useful subset of the full information need be extracted in the task. Traditional approaches have focussed on the keywords involved. We have shown that accounting for more of the data, by using a large-vocabulary recog...

Keyword Spotting (KWS) has been the subject of research in recent years given increase embedded systems for command recognition such as Alexa, Google Home, and Siri. Performance, model size, processing time, robustness to noise are fundamental these systems. Furthermore, applications demand computationally efficient models that can be implemented current technology. In this work, an approach ke...

We introduce a method for situation understanding from natural, face-to-face conversation. Our method combines a network of commonsense knowledge with keyword spotting and contextual information automatically obtained from a wearable device such as a PDA or cell phone. Using this method we demonstrate the potential for high accuracy, detailed classification of conversation topic.

We submitted two approaches as the required runs: Acoustic Keyword Spotting as the primary one (AKWS) and Dynamic Time Wrapping as the secondary one (DTW) for the Spoken Web Search task. We aimed at building a simple phone based language-dependent system. We experimented with universal context bottle-neck neural network classifier with 3-state phone posterior features or bottle-neck features.

Arabic speech recognition suffers from the scarcity of properly labeled data. In this project, we introduce a pipeline that performs semi-supervised segmentation of audio then— after hand-labeling a small dataset—feeds labeled segments to a supervised learning framework to select, through many rounds of hyperparameter optimization, an ensemble of models to infer labels for a larger dataset; usi...

Three experiments used word-spotting to examine influences of phonetic and musical parameters on intelligibility of closed-set but unpredictable words in polytextual singing. Main comparisons were: 3 musical genres (medieval polyphonic motet, similar but homophonic motet, jingle); harmony (consonant, dissonant); keyword phonetic properties (‘acoustic contrast’, vowel length, vowel quality); con...