Recently, Research at Google sat down with Research Scientist +T.V. Raman
to discuss his work on accessibility (http://goo.gl/7kfA7
) and empowering blind and low vision users with eyes-free access through ChromeOS and Android. Before joining Google in 2005, Raman worked at Digital’s Cambridge Research Lab, Adobe Systems, and IBM Research. Obtaining a PhD in Applied Mathematics and an MS in Computer Science from Cornell University, Raman has incorporated features from his PhD thesis Audio System For Technical Readings (AsTeR)
, awarded with the ACM Doctoral Dissertation Award in 1994, into the Chrome browser.
With research that involves auditory user interfaces and speech interaction, Raman strives to ensure that the Google user experience reflects the best aspects of human communication by creating innovative technologies that benefit all users. Read on to learn what Raman had to say about the current state of accessibility research, his motivations, and what he perceives as the future of accessibility and computing.
--Research at Google:
Can you give us a brief overview of your work at Google?T.V. Raman:
I came to Google 8 years ago, working on what I call “innovative user interfaces” and “eyes-free interaction”. My broad research interest, partially motivated by the fact that I cannot see, was that if you have feature x or y, and you want to use service x or y when you’re not looking at the screen, how would you build it? Another engineer and I started started hacking things together, which led to early versions of speech APIs for Android in 2009. This work quickly progressed and seeded more formal work on implementing accessibility APIs for two key platforms: Android and Chrome. R@G:
How would you say Accessibility Research fits within the work undertaken at Google?TVR:
In my experience, all of Google is involved in two things: The first is to bring all the world’s information online. We don’t just provide search results, we bring data online via Google Maps, Google Earth, Street View, Google Books, etc. Secondly, we build platforms and user interfaces to enable our users connect to all of this information as easily as possible.
I believe that innovations in the Accessibility space lay at the intersection of these two spheres of Google’s work. Think about the opportunities this opens up if you are looking at how to change people’s lives, with respect to not being able to see or hear, to enable them to more easily access all of the available information that is online. Those opportunities are what I’m passionate about, the part that I’ve always focused on.
The Knowledge Graph, for example, I believe is a huge deal: If Google can understand everything on the web, and present clear and concise information gathered from multiple web sources, it can save a blind user a significant amount of time. Also, the volume of information in Google Maps; having all the world’s street data on the phone had never been done before! The things we can build with that for a blind person who wants to navigate the world are really quite amazing. What is possible today is remarkably different than what you could have done 5 or 10 years ago. R@G:
Do you see the work in Accessibility as having application to a broader user community?TVR:
Our devices have gotten small enough and powerful enough that you can wear them. Think, for example, about what you can do wearing a smart camera where your eyes are, having a device that can do vast amounts of image processing via your pocket or cloud. Although I can’t see, maybe now I can read what’s on your t-shirt, or read all the signs in my environment and translate them if they are in a foreign language or even a foreign script. That’s game-changing, because it’s not just for blind users, everyone
can find it useful.
To me, these are the kinds of things that make work in the accessibility space very interesting; you have the opportunity to build something that a particular user community needs way before the rest of the population, and if you do it right, you end up building something useful for a MUCH bigger population. Building use cases that have not been built before, from an innovation point of view, is incredibly interesting. R@G:
It sounds like the work you do really lends itself to intense collaboration across many areas of work, i.e. accessibility for Chrome, for Android.TVR:
Accessibility work is very tightly knit into several product areas. We’ve always had no choice but to be tightly knit, because all of the platforms develop very rapidly. Chrome pushes every six weeks and Android updates every 6-9 months, which means that when an update is pushed out, we are already working on the next one. What we are doing is cutting-edge research, doing new things. We can’t do it isolated from those products, because the rest of the world is moving too fast. The research I do is really embedded within a lot of the engineering here at Google, so you could say it’s a good example of the Hybrid Research Model that +Alfred Spector
, +Peter Norvig
, and +Slav Petrov
In your experience, how is Research at Google unique?TVR:
Before coming to Google, I was at IBM Research for 6 years and before that at Adobe’s Advanced Technology. I think the key difference is that at Google, any single product area likely has more PhDs than does all of the Research organization, and it is all very integrated. Traditional academic research, corporate research doesn’t do this and this creates a culture that is very different from other companies. What happens when you have this spread of talent is that when you build something, you build it on the same infrastructure, the same underpinnings as the rest of Google. Because of this you have the opportunity to reach the user directly; if you build something that you want users to get their hands on, you do it. I find that very empowering.R@G:
You mention traditional academic research. With regards to the research culture at Google, how do you view its impact on the general research community in terms of publication and collaboration? TVR:
Almost every piece of work that I’ve done during my time at Google is open source. I believe open sourcing goes to the heart of scientific publication. Why do you publish? It’s not about promotion or citation; you do it so someone else in the field can recreate the result you claim you saw. In software, the best way I know how to do this, to let you recreate the result, is to give you the source code, and let you verify it. Regularly open sourcing code is actually a very strong means of scientific publication. When you release source code, and write a technical article that describes the architecture, and you get recognized people in the community seeing it...well, that’s peer review. And it is peer review that’s much more timely than traditional methods of publication as well. There is still value in traditional publishing for very large projects on longer timescales, but in the computer science community, things progress so rapidly that I think other forms of publication are equally important.R@G:
And what about external collaboration with academia? TVR:
I like to stay involved in external collaborations. I review research proposals on a regular basis. This year I’m hosting someone from academia who is on sabbatical, and the hope is that when he goes back we will continue that collaboration. R@G:
How is the industry progressing in terms of accessibility? What do you think the focus is, and is it on point? Should it be shifted?TVR:
There are two things. One is the baseline with respect to enabling people with disabilities to do simple things that need to happen, making currently available platforms accessible. This is very important, and is the “bread and butter.”
Second, I think there are many opportunities in terms of wearables and new technology. If you think of various use cases in terms of accessibility, the innovation enables all users to use new technology, not just the blind, or deaf. And if we do this, we will motivate a lot more smart engineers to work in this accessibility space, and that by definition allows us all to move forward. There is a lot of attention around accessibility due to legal requirements, and those are great for raising awareness. However, it is not so great for attracting young innovative engineers to work in this space. I want people to say: “Look at how this work can change someone’s life!”
Engineers are proud of what they build, and if I can come to you and challenge you by saying: “How would you solve this problem?”, you’ll go and create something innovative, something that I could not have dreamed of, and that is really an opportunity that should be leveraged. The opportunities are enormous; by pigeonholing this work as simply “accessibility”, you may miss the moonshots. You may do the baseline, but you may miss the opportunity to do something great.
To me, that is the more empowering aspect of working in this space. Of course, we must absolutely make sure things that are currently built are accessible, and define the baseline. And the upper limit....well, there shouldn’t be one. The sky should be the limit. You are bringing all these things online that have never been done before; they are pure science fiction. That is fundamental research. You don’t know what will develop, what the impact will be from researching use cases in accessibility. Like Alexander Graham Bell’s research with creating a device for the deaf, which led in part to inventing the telephone.R@G:
That’s an interesting perspective, that accessibility defines the use cases....TVR:
The accessibility space is interesting because the need is greater; the technology becomes available for accessibility space a little earlier. For example, people have been using text to speech for the last 20-30 years. Because blind users were early adopters of that, they were also early adopters of ebooks. In contrast, for ebooks to become viable for the mainstream population, visual displays had to get to at least 200 dpi, otherwise they couldn’t compare to printed paper. And even now, for the mainstream audience to listen to an audio book, it still utilizes professional human readers rather than using computer-driven text to speech. But for the blind population this has been acceptable for 20 years, and in that sense, you could maybe say that blind readers invented ebooks. This phenomenon happens in almost every artifact. Accessibility defines use cases. R@G:
In your opinion, what is the next “big thing” in accessibility research?TVR:
I’m really excited about the possibility of having a network of devices on my body, sets of body sensors. On many phones, we already have NFC readers, and NFC stickers are dirt cheap. What if I had an NFC reader in my shoe, and we stuck all the NFC stickers in various places on the floor, and as you walked along your phone knew exactly where you were? Or perhaps other wearable technology that could vibrate to give you feedback as you got close to a wall? Imagine a network of sensors and actuators on your body that interact with the world, and all the data being exchanged with all the wearable devices. I think that area of research will be a lot of fun for the next 3-4 years. R@G:
So, it is safe to assume you’re excited when it comes to the possibility of Ubiquitous Computing?TVR:
I’m a technologist! Of course, anything can be used or abused. With any technology you use, one has to evaluate what the risks are as a function of the potential payback.
With accessibility, I’m trying to add the ability to know what the users’ capabilities are. It’s important to know, when using a service, how
you are using the service. If I know a user is not looking at the screen, it isn’t my business why
a user isn’t looking at the screen; maybe they are blind, maybe they are driving, maybe they are having a conversation with someone nearby. All that’s important is knowing what your abilities at that given instant are, and being able to deliver a service that is situationally suitable. When we do that, when we get the balance right, it’s a service we can all use and feel comfortable with.