Artificial Intelligence In Speech Recognition

Now-a-days speech recognition has overcome all of its earlier presumed limitations and is widely used as a well capable security screen. The growing popularity in the application of speech recognition in all phases ; be it homes or even the industries. With the advancements in Machine Learning, Big Data Analytics and above all Artificial Intelligence we are down the exploring the endless possibilities IOT relations with Speech Recognition. IOT has taken the world by storm be it in usage or in monetary development by almost 60-70% every year.

It is easily the need of the hour to ponder and perform on the various uninvestigated parts of IOT driven speech recognition systems for the benefit of the society. 

Basically, it was very difficult to answer the question whether it is possible to develop a portable system for the automatic recognition and translation of spontaneous speech in 1992. Previous research work on speech processing had focused on read speech only and international projects aimed at automated text translation had just been terminated without achieving their objectives.

Get quality help now

Doctor Jennifer

Verified writer

Proficient in: Socialization

5 (893)

“ Thank you so much for accepting my assignment the night before it was due. I look forward to working with you moving forward ”

+84 relevant experts are online

Hire writer

The German Federal Ministry of Education and Research(BMBF) made a careful analysis of national and international research projects conducted in the field of speech and language technology before deciding to launch an eight-year basic-research lead project in which research groups were to cooperate in an interdisciplinary and international effort covering the disciplines of computer science, computational linguistics, translation science, signal processing, communication science and artificial intelligence.

At same point, the project comprised up to 135 works packages with up to 33 research groups working on these packages.

The project was controlled by means of a network plan. Every two years the project situation was assessed and the project goals were updated. An international scientific advisory board provided advice for BMBF. A new scientific approach was chosen for this project: coping with the complexity of spontaneous speech with all its pertinent phenomena such as ambiguities, self-corrections, hesitations and disfluencies took precedence over the intended lexicon size.

Another important aspect was that prosodic information was exploited at all processing stages. In addition, symbolic and statistical processing was integrated from the beginning, thereby including the world and domain knowledge necessary for the translation task. This meant that a total 23,000 rules for the translation of 10,000 words for the language pair German-English and of 2,500 words for the language pair German-Japanese had to be fed into the system.

A new operational approach was used in the Verbmobil project: an absolutely decentralized, large-scale project with research groups from a wide range of different disciplines was implemented by means of the most advanced communication and telecooperation methods. Central scientific control was provided by the German Research Center for Artificial Intelligence in Saaebrticken. An important aspect was that the software packages already existing in different programming languages or developed in Verbmobil were integrated by a central system integration group with-out reimplementation.

The Verbmobil project received funds in the amount of DM 116millions between 1993 and 2000. The project goal defined for Verbmobil in 1992, namely the development of a prototypical, portable translation system for the language pair German-English by the year 2000, has now been achieved on the basis of a notebook concept; other objectives added later on, such as the development of a telephone translation system and a remote maintenance system for PCs, have also been reached. During the implementation of Verbmobil, 800 scientific publications.

In accordance with the present invention a system is provided to reduce noise from a signal of speech that is capable of deciding upto the adjustment of a filter subsystem by distinguishing between noise and speech in the spectrum of the incoming signal of speech plus noise by testing the pattern of a power or envelope function of the frequency spectrum of the incoming signal and deciding that fast changing portions of that envelop denote speech whereas the residual is determined to be the frequency distribution of the noise power, while examining either the whole spectrum or frequency bands thereof, regardless of where the maximum of the spectrum lies.

In another embodiment of the invention, a feedback loop is incorporated which provides incremental adjustments to the filter by employing a gradient search producer to attempt to increase certain speech-like features in the systems output. The present system does not require consideration of minima of functions of the incoming signal or pauses in speech. Instead, the present system employs an artificial intelligence system to which is input the envelope pattern of the incoming signal of speech and noise. The present system then filters out of this envelope signal the rapidly changing variations of the envelope over fixed tie windows.

The RWTH LVCSR system is continuous Gaussian mixture density speech recognition system which has been described by detail by Ney et al. (1998). In this paper we report in detail on:

• acceleration methods for within word recognition,
• search and acceleration methods foe across-word recognition,
• acceleration methods for vocal tract normalization,
• incremental processing methods to reduce the response time

We can easily conclude that our research topic is the best and most diverse topic. Its capabilities are immense and linking the IOT to the speech recognition features is the icing on the top now but experts say, its just the tip of the iceberg. It’s possibilities are endless.The new technology will help reduce the thefts and will take the biometric safety to a whole new level. We have presented in our assignment the gist of the detailed analysis as how to explore and implement these upcoming studies for the benefit of humanity as a whole

Works cited

1. Hinton, G. E., Deng, L., Yu, D., Dahl, G. E., Mohamed, A.-R., Jaitly, N., ... & Kingsbury, B. (2012). Deep neural networks for acoustic modeling in speech recognition: The shared views of four research groups. IEEE Signal Processing Magazine, 29(6), 82-97.
2. Niu, Y., Wang, X., & Wu, J. (2019). Speech recognition with deep learning: A review. Neurocomputing, 364, 22-38.
3. Chen, L., Zhang, X., & Jiang, H. (2018). Recent advances in deep learning-based speech recognition. Journal of Electronic Science and Technology, 16(2), 97-110.
4. Li, X., Li, Y., Li, S., Wang, B., & Li, X. (2019). Review of deep learning-based speech recognition technology. Mobile Information Systems, 2019, 1-9.
5. Gou, L., Wu, Z., Zhang, D., & Wu, X. (2018). An overview of deep learning-based speech recognition. Frontiers of Information Technology & Electronic Engineering, 19(1), 10-24.
6. Kuo, H. H., Chen, J. L., Chen, C. H., & Wang, H. M. (2020). Application of artificial intelligence in Internet of Things: A systematic review. Journal of Industrial Information Integration, 17, 100132.
7. Zhang, Y., & Yang, X. (2018). Speech recognition using deep learning algorithms: A review. Journal of Computer and Communications, 6(05), 86-96.
8. Deng, L., & Yu, D. (2014). Deep learning: methods and applications. Foundations and Trends in Signal Processing, 7(3-4), 197-387.
9. Manasrah, A. M., Aljarah, I., Al-Jarrah, M. A., & Hassonah, M. A. (2018). A survey of speech recognition systems based on deep learning techniques. Journal of Intelligent Systems, 27(4), 627-640.
10. Verbmobil project. (n.d.). Retrieved from https://www.dfki.de/en/web/research/projects-and-publications/projects/project/verbmobil/