It is proverbial that necessity is the mother of discovery. With new tech firms and startups booming into business, the key to survival in this cut-throat market is innovation. Innovation is a must for adaptability. Many have faltered, while other prominent tech giants might have managed to survive the up and downs. IBM is one of the household names whose range of trajectory is exceptional. We are blessed with new technological advancements like Artificial Intelligence, Machine Learning, Edge Computing, etc. Computer hardware and software manufacturers have upgraded. Apple Rolled up with iPhone, iPad, iPod, basically all cool gadgets to brag about when rich. Kidding!
Coming back to IBM, whose’ priority has been to enhance life and society through products and services; since its inception in 1891 as the Dayton Scale Company. From developing a translation system that helped speed up Nuremberg trials in creating magnetic strip we see in Debit and credit cards in 1970 to when Watson was pitted against Ken Jennings and Brad Rutter in Jeopardy in 2011, IBM has a rich and cool history. And the list does not stop here. Recently IBM announced plans to integrate Project Debater technologies into Watson.
Watson as a better communicator?
Watson is IBM’s natural-language-processing (NLP) computer system. And this voice to text speech processing platform system is available through the IBM cloud. One can use this feature to process up to 500 minutes of audio for free per month. Although you do need to create an IBM Bluemix account (free registration).
The new add on features will augment Watson with the potential to help organizations with the most challenging aspects of the English language. That too, with greater clarity and greater insights. The main objective is to mine, identify, understand, and analyze idioms, sentiment shifters, colloquialisms (e.g. go bananas, wanna, gonna). Imagine futuristic voice assistants understanding expressions like break a leg, stay frosty, a dime a dozen, etc.
Indeed it is going to be a mammoth task. In idioms, a set of words has a different meaning from the other combinations. While AI sometimes gives an error output even after being fed with actual data provided, recognizing phrases from common speech can be a win for NPL.
Plus, IBM has also unveiled a new classification technology from IBM Research for a better understanding of business documents including business contracts and PDFs, to their AI models. With Project Debater’s deep learning-based classification technology, the new capability can learn from as few as several hundred samples to do new classifications quickly and easily. This is planned to be added to Watson Discovery later this year. Last year Project Debater, debated against Harish Natarajan, who holds the world record for most debate competition victories, on stage at Think 2019.
Along with that, two new features will be integrated into Watson: Summarization and Advanced Topic Clustering. Former will allow the creation of summaries of textual data from a variety of sources to give an idea of what is being said and written about a particular topic. An early version of Summarization was used at The GRAMMYS this year to analyze over 18 million articles, blogs and bios to produce trivial insights on hundreds of GRAMMY artists and celebrities. The data was then infused into the red carpet live stream, on-demand videos and photos across www.grammy.com to give fans deeper context about the leading topics of the night.
While clustering builds on insights gained from Debater to pull together clusters of data to create meaningful ‘topics’ of related information. This will allow concerned experts to customize and fine-tune the topics to reflect the language of specific genres and associated industry jargon.
With competitions from bulk transcription services Google Cloud Speech-to-Text and Amazon Transcribe, IBM Watson has a lot to catch up with and come up with more creative plans. Once it is presented in the world, the life of an average user will never be the same, baffled by its developing doppelganger.