Speech-to-text Technology: Tales of just another knackered college student (Innovative Ideas Challenge)

#hackwithdg #ai #firstpost

Introduction

For as long as I can remember, I have always been obsessed with the idea of automation. Be it schedule-sending mails or walking past an automatic door, the idea of things operating on their own without fail after being pre-programmed really just tickles my fancy. As an undergraduate student of Electrical and Electronics Engineering with a focus on Artificial Intelligence, I am constantly stressed out by virtue of the immense workload. From attending long classes to going for laboratory experiments to preparing for exams to writing term papers whilst managing projects, the hustle and bustle is non-stop. Take into account my social impact and volunteering commitments and you might just have the perfect stress recipe. As such, I find myself constantly thinking of ways to automate as many tasks as possible. I guess this constant urge to reduce labor whilst increasing productivity was the genesis behind my fixation for innovative ideas. Although prior to the Deepgram x DEV Hackathon, I had not encountered Deepgram, given my interest in Artificial Intelligennce, I am not new to the concept of speech recognition technology.

My Deepgram Use-Case

The origin of this idea traces back to my last holiday. In a bid to give back to the community whilst staying proactive, I was tutoring a younger student (let’s call him Dipo) in preparation for his Secondary School Leaving Examinations. The student in question is quite acquainted with and fond of technological gadgets and his action that particular day struck a string so hard, it almost felt like a Eureka moment. I asked him the value of Planck’s constant, a fundamental physics constant used in quantum mechanics calculations only to be answered by his phone. Now that I think of it, I don’t know what was more surprising at that point. The accuracy to which the Google assistant returned the constant or the fact that the Google assistant heard what I said, considering how fast I usually talk. Perhaps, I was distracted by the euphoria that hit me. I am pretty convinced that what I felt is similar to how Archimedes felt when he supposedly hopped out of his bath and ran onto the streets to tell the king, ‘I’ve found it’. I felt like a game-changer. I thought to myself, ‘if Dipo could do that to his tutor, why can’t I do that on my college professors?’ Of course, I would need their permission in my case. Right at that moment, I realized that if I harnessed reliable speech-to-text technology, I could save myself a boatload of stress. And, I believe a lot of college students feel the same way. Wouldn’t life be much easier if instead of typing as the lecturers gave their lectures during the online classes, one simply had a speech-to-text technology convert their lectures to a text document?

Dive into Details

Although speech recognition technology is already a part of our everyday lives, for now, it is still limited in its application areas. Speech-to-text technology, and by extension, Artificial Intelligence has the potential to make far-reaching changes in the educational sector. As in my innovative idea, with an app that automatically transcribes live streaming audio from the lecturers in real time using Deepgram’s SDKs, which are supported for use with the Deepgram API will improve endurance and reduce writing fatigue by eliminating the physical act of composing to paper and keyboard. This will in turn shift focus from the physical act of writing to that of expression and organization of thoughts and knowledge. In a bid to confirm this hypothesis, I tried this locally when I endeavored to turn BBC’s live stream into text. Here was my result.

A sample of my code can be found here.

Conclusion

Finally, as an artificial intelligence enthusiast, I am excited by how much speech-to-text technology, a subset of Artificial Intelligence can change the way we look at Education. With the research I did whilst participating in Deepgram Hackathon "Innovative Ideas" challenge, I have gained a lot of insights as regards how speech-to-text technologies work and how it can help education as the key to development while it opens up a world of endless possibilities.