When Databases Talk Back

I’ve been reading through DITA blogs the last couple of days. Over 40 blogs and 130 posts of great writing so far from classmates on DITA.  Some totally new to blogging and facing their blogging anxieties and WordPress learning curves; some more experienced.  All of them spreading their blogging wings in new and different ways as we learn together.  Way. To. Go. CityLISters!

Lots of blogs had great articles following last week’s lecture and lab on information retrieval thinking about their personal and workplace experiences of information needs and information seeking and the differences, benefits and possible downsides of natural language queries vs more structured queries in search engines and library discovery systems.  One phrase that really struck me was from Emma’s blog in her post on information retrieval:

“We forget that we’re talking to a database rather than a person; we forget that our clumsy queries are being mechanically parsed and restructured and matched against what is essentially just a bunch of giant spreadsheets.”

I then began to wonder how searching changes when we can query databases using natural language by speaking to the information interface using voice search services such as Siri and OK, Google.

okgoogle

I’d never tried these before so I thought I’d give it a go so activated OK, Google on my computer and allowed it to use my microphone.  I then tried using some searches using examples from “What you can ask Google to do for you” and then some of the queries from our lab session.

The interface is really clear.  On the Google search page you either say OK, Google or click on the microphone in the search box.  Google will then ‘Listen’.  If you are on the main page this listening graphic fills the page; if you are on a search results page then it will overlay the listening graphic as a bar at the top of the screen.

googleislistening

Then you speak … and Google is pretty good at listening.  It recognised my query every time.  It writes the words it hears as you speak.  Sometimes it even got an initial word wrong on a first hearing and then corrected it when it understood the next word to make the sentence make sense.  Kind of creepy …

Google then goes away and executes the search and displays the search results screen.

Even more surprising Google also sometimes speaks back! Goggle will tell you the time, tell you the square root of 2209, tell you the film times for Gone Girl at your local cinema, tell you “check out these pictures” if you search for an image.  It is both cool and creepy at the same time.   I can’t decide if I am amazed or freaked out. It’s a female voice so more Mother (MU/TH/UR) than HAL.

The transactional searches were really good.  I may never us a calculator again.  The fact finding, subject and exploratory queries were also recognised and returned reasonably relevant results.  The only thing that was slightly lacking was finding images for reuse.  It could find images well enough but couldn’t recognise that “licensed/labelled for reuse” related to the usage rights facet rather than the subject of the query.  No doubt it will soon.

This slideshow requires JavaScript.

See all query experiments and results.

Oh and Google?  Reading is a verb; when it’s the British town it’s pronounced “Redding”. Not that clever … yet.

Updated: As we know context is everything and Google is actually fine at pronouncing Reading if it has enough context to know it should be talking about the place.  Try: Ok, Google. Where is reading?  Actually quite clever.

where is reading


Featured Image:
Telephone Switchboard Operators – a vintage circa 1914 photo by reynermedia.  Source: Flickr. (CC BY 2.0)

Advertisements

4 Comments

Add yours →

  1. Very nice. I’ve only ever used the voice search on my phone, and I was quite surprised at how accurate it was. But that was because I was searching for something rather specific. It’s nice to see how the system coped with ‘reading’. I’m sure it will improve in time. 🙂

    Like

  2. This is awesome! I’ve never used the voice search function on any of my devices, partly because I fear I’d be wasting my time being constantly misunderstood. It’s really neat to read about your experiences with Google voice search and how well it worked. I definitely think I’ll start making use of it for calculations, at the very least. Great post! 🙂

    Like

  3. This is fascinating. It always seems to me like voice-activated computer functions are a bit of a gimmick–maybe fun to play with and definitely useful if you’re driving, but hard to incorporate into regular life. This seems especially true within the context of library- or research-oriented searches, where the vocabulary is often so specialized that the voice recognition software would struggle with it. Also, what a paradigm shift it would be if everyone started using voice-activated tools in a library! I’m interested to see how this technology is embraced over the next few years.

    Like

  4. A year ago or so my husband was doing some design incorporating voice recognition and had me, and lots of his colleagues have a ‘day of voice’ where we had to use voice commands whenever possible and then report back on our experience. It was interesting, especially for my colleagues in the cubicles around me!
    If you’re interested in this, he presented on designing for voice at UXAustralia, copies of the slides are here: http://www.slideshare.net/jonnyschneider/designing-for-voice-web

    Liked by 1 person

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: