Employability Series: Nuance Communications Visits Swansea University, by Alexia Bowler and Jade Hobby
As part of the College of Arts and Humanities’ employability drive and continuing student engagement programme, the Applied Linguistics department welcomed its first industry speaker, from Nuance Communications, to Swansea to talk to students about the work they do as a language technology and communications software company. As well as showcasing the company’s own cutting edge work, the talk provided a fantastic opportunity for our students to find out what real-world applications an applied linguistics degree can provide them with in such an exciting and innovative industry (and the staff learned a lot too!).
But a little about the company first!
About Nuance Communications, Inc.
Nuance Communications is the pioneer and leader in conversational and cognitive AI innovations that bring intelligence to everyday work and life. The company delivers solutions that can understand, analyse and respond to human language to increase productivity and amplify human intelligence. With decades of domain and artificial intelligence expertise, Nuance works with thousands of organisations – in global industries that include healthcare, telecommunications, automotive, financial services, and retail – to create stronger relationships and better experiences for their customers and workforce.
According to Nuance’s website, their stated aim is to make ‘technology fluent in all things human’. So, rather than having to learn specific commands, or a series of unfamiliar menus with unfamiliar terms, Nuance are interested in making technology work for us: its users. It is the nuance (‘scuse the pun) of human communication that drives the company’s research and its applications, not (the often unwieldy) adaptation to technology.
We are pioneers in making technology fluent in all things human: from understanding spoken words and extracting their meaning to adaptively and seamlessly interpreting the swipe of a fingertip. Every interaction can finally be understood to deliver exactly what a person needs. And we continuously evolve the ability to perceive the nuance of words, actions and meaning – to fit seamlessly into your life, your business and your world. (Nuance Communications Website)
Our guest, Brian Redpath, who is the Public Sector Director at Nuance, spoke about ‘Bringing Intelligence to Life’, giving us an overview of the company and its services, including some fascinating examples to illustrate the work Nuance does. One of the company’s accomplishments Brian told us about was the development of a new Intelligent Telephony Automation (ITA) system to be used on customer service helplines, in this example, HMRC (Her Majesty’s Revenue and Customs…that’s the tax office to you and me).
This ITA system replaces the traditional touchtone IVR (Interactive Voice Recognition, e.g. for tax credit enquires press one, for income tax enquiries please press two, you know the drill…) with sophisticated natural language speech recognition. With this new technology, all customers need to do now is state why they are calling to get through to the right department, then intelligent speech recognition identifies key words and asks additional clarification questions before routing the call to a relevant advisor, by-passing what can often feel like layer-upon-layer of menu options that leaves many callers frustrated (we have all been there…).
ITA also completes security checks automatically, notifying the advisor of the customer’s details and reason for calling ahead of time so that the advisor can deal with the call quickly and effectively, making it a better experience for all. Indeed, a 2015 case study found that the introduction of the IPA reduced the average IVR journey by over 37 seconds and also resulted in a reduction in call-handling times.
So, as well as making life better for both customers and advisors, ITA has also saved the HMRC £5 million due to shorter calls. Everybody’s happy!
Relevance to Our Students?
Where all this became relevant for our students, studying for a degree, or perhaps a higher degree, in English language and/or Applied Linguistics, was in the kinds of roles and tasks that seem to be much in demand in companies such as Nuance. Indeed, we were excited to hear about the variety of roles in companies that work on audio, speech and language technology solutions, all of which clearly play to the skills and knowledge gained by the graduate of an Applied Linguistics department. Indeed, it seems that the language technology company’s work is truly interdisciplinary, involving those with skills in computer science, linguistics and psychology.
As we heard, careers with Nuance ranged from working in the marketing and communications departments, IT, to their sales team, all of which require an understanding of the language technology research the company engages in, as well as an ability to see their application to the wider society and business world. Furthermore, specialised roles in the company that we heard about included the role of speech scientist and Virtual User Interface (VUI) designer, and this didn’t even include their research and development department, all of which might well appeal to those doing higher degrees in linguistics.
For those doing higher degrees in linguistics, we were able to hear about roles such as the speech scientist and the VUI designer. The speech scientist, we were told, fulfilled several functions and was involved in collaboration with the research and development teams and designers. Work involves activities such as writing and tuning speech recognition grammars and building acoustic models. For work on natural language understanding, for example, the speech scientist might build and tune statistical language models as well as statistical semantic models. Involvement with the company’s voice biometric engine would include tuning and calibrating the system’s background models and tuning the system’s configuration. The VUI designer, we were told, works closely with the speech scientist in reviewing the latter’s activities but was more was user focused; that is, making sure the interface is optimised for the user of the system and its performance capabilities maximised.
But why are these language technologies useful, or even desirable, we hear you ask?
Well, aside from the solution provided for HMRC, already mentioned, one of the software applications widely used is that of speech recognition, for example turning speech into text; creating documents, transcribing, and even formatting your written document, all using speech recognition. Where this is useful is in something like a medical setting. Medical professionals can use dictation software to create accurate reports quickly. Indeed, Nuance’s view on its role in healthcare is one of assisting the efficiency in clinical situations. Nuance’s site states:
Nuance is redefining how clinicians interact with clinical documentation on every level. Our solutions allow healthcare professionals in the NHS and other health institutions to capture and document the patient story quickly and accurately, improving the quality of the care record and freeing up time to care. (Nuance Communications Website)
So it’s clear that such activities provide fascinating solutions for users who, for example, might have visual impairments, or mobility issues (including repetitive strain injuries), or other needs that can be assisted by voice recognition, as opposed to traditional input methods. It’s not hard to imagine a time where the language technology software might even do away with cumbersome elements such as the keyboard altogether…
We also heard about the company’s research into Biometrics which enables speaker identification (determining the identity of an unknown speaker), speaker verification (where one can authenticate the claim of someone’s identity via their ‘voiceprint’), language identity, speaker segmentation (identifying the transitions in speakers in the audio) and speaker clustering (grouping segments of speech from speaker characteristics).
It seems that our voices are unique; that is, each person’s acoustic patterns differ due to the size and shape of our bodies – the throat and mouth particularly, as well as our vocal chords. Not only this but prosodic features such as stress, pitch, intonation, rhythm or speaking style, also contribute to the voice’s unique identity. What this means is that, like our fingerprints, our ‘voiceprint’ can be used for identification and verification purposes.
So, the research and development in voice biometrics at Nuance, as well as at other language technology companies, moves beyond what is being said, to who is doing the speaking. It’s application? Well, this technology has implications and applications for things such as telephone banking and solving crimes including fraud, among many other things.
Brian’s talk was fascinating and covered a great many aspects of the company’s activities that can’t be covered in a blog post (even one this long), so if you’d like to know more about Nuance then click on the links to the company or read their blog What’s Next.
We’d like to thank Brian Redpath for spending the time with us and providing us with an informative and energising insight into the life of industry applications for language research in the 21st century!
***
If you would like to find out more about our engagement with industry partners and language technology industries in general, we have another upcoming blog post to celebrate this semester’s final event with Appen. So keep an eye out for our next post!
Contributors:
Jade Hobby is the Employability Assistant in Swansea University’s College of Arts and Humanities. For employability enquires and opportunities within COAH, please email Jade at: jade.hobby@swansea.ac.uk
Dr Alexia L. Bowler is Admissions Tutor for Applied Linguistics and is one of the department’s Employability Officers. She also runs the department’s social media (blog and Twitter account).
