Category Archive: Artificial Intelligence

Talking to Myself with SimulScribe

March 4th, 2007

vocaltract2.gif

Illustration from a 1940’s Bell Labs project investigating human speech synthesis and recognition

I recently signed up for SimulScribe, a new service which replaces your existing voicemail system with one that:

  1. Transcribes the voice message into text (using a speech-to-text (STT) engine)…
  2. wraps the voicemail message into a WAV file…
  3. and then emails the raw text and the WAV file (as an attachment) to your email address.

Setting up SimulScribe couldn’t be easier: The free trial doesn’t even require a credit card to start using right away, and they provide you with explicit and shockingly simple instructions for configuring your voicemail for your particular carrier. You can be set up with the SimulScribe service in literally under 3 minutes.

After setting it up (and this may come has a shock to those of you who still think STT is not ready for prime time), the system has performed almost flawlessly.

Below I’ll present some example transcriptions, followed by some ideas on how this technology might be extended in the future. MORE…

Aura of Inevitability (or: When a Technology’s Time has Come)

February 23rd, 2007

waiting_for_technology_410.jpg

New technology products often take us by surprise. In 1992, for example, we couldn’t possibly have dreamed of how the Internet would transform the world by 1997, only 5 years later. The best innovations are things “you never knew you wanted but cannot live without” kind, inventions that come out of nowhere. YouTube, for example. Or TiVo.

But certain other technology products are so obvious that when they finally emerge many people shrug and wonder “what took it so long?” We knew they were coming, but year after year they never actually materialized.

When they do materialize, we are overjoyed. After years of waiting, for example, we are finally getting MP3 players into cel phones.We are using wireless networks and bluetooth more and more, but we knew we wanted this stuff years ago. The technology consumer will often heap glowing praise on these kinds of new technologies as they emerge, calling them innovative and groundbreaking, when in fact the functionality of the products is merely filling a hole that everyone knew was there.

The Apple iPhone is a perfect example: while the UI is indeed remarkable, almost nothing about it is technologically innovative or new. If you asked me (or just about any of my friends) to describe the perfect cel phone feature set, it would look a lot like an iPhone. In fact, as the owner of a Windows PocketPC phone for nearly 5 years, nothing about the iPhone’s tech specs surprised me. The UI, again, is great and very innovative, but the hardware itself and the basic concept of the device is wholly old news. MORE…

The Empathy Test

September 1st, 2006

bladerunner_empathy_01.jpg

“Let me tell you about my mother…”

In the movie Blade Runner, the “Voight-Kampff Empathy Test” detects whether or not a test subject is a real human being or an android “replicant”. A machine reads the body’s physical reactions to various psychologically- provocative scenarios (“Capillary dilation of the so-called blush response? Fluctuation of the pupil. Involuntary dilation of the iris…”) and reveals whether or not the test subject’s sense of empathy is consistent with that of a real human being.

There has always been a lot of talk about “empathy” in the information architecture world. Information architects regularly describe empathy as both a critical prerequisite for the job and as something fundamental to the professional practice. But in all my years as a user experience designer and information architect, I was never taught empathy or specifically tried to train myself to be more empathetic. I’ve never taken any kind of Voight-Kampff test to see how empathetic I am.

Practices like user research and deliverables like user personas certainly embody the concept of empathy, but so does the work of countless other professions — artists/illustrators, doctors, marketers, cops, salespeople, journalists, social workers, politicians, even management. Many of them even specifically include empathy training as part of their academic curricula and professional development programs.

Conversely, I’ve met information architects who are very good at what they do but who don’t strike me as particularly empathetic. Pig headedness, self-aggrandizement, insensitivity, and other non-empathetic personality traits haven’t stood in the way of people becoming excellent at many important aspects of IA. How have we come to claim empathy as a faculty we posess and utilize more than people in other professions do?

Defining the Term

Personally, I’ve tried to avoid using the term since IMHO it conjures up a lot of unprofessional connotations: Like a person in a job interview who claims “I like working with people“, a claim to be empathetic can suggest just the opposite: that one needs to make an overt concerted effort to be empathetic to compensate for a more fundamental shortcoming of the faculty. I’m not even sure we agree on what it means, since for most people it seems to have more of an emotional connotation, and emotion is a subject that IAs rarely, if ever, actually discuss.

bladerunner_empathy_02.jpg

“Did you ever take that test yourself?”

What do you think? Is it fair of user experience designers or information architects to claim “empathy” as a something that makes us special?

For kicks, here’s a real-world empathy test. I took it myself and scored pretty much in the very middle of the empathy range. It hardly seems very scientifically legit, but give it a shot. I doubt too many IAs would score very highly on this particular measure of empathy.

Elegant Experiments

July 20th, 2006

skinnerbox.jpg

The classic Skinner box experiment.

A comment by Steve Baty on my recent series about design research got me thinking about the fine art of experimental research design:

Where research (in all forms) becomes a waste of time and effort is when the research design and methodology applied invalidate the conclusions _before_ they can be drawn.

One of the things that defines a great scientific thinker is their ability to design and construct meaningful experimental methodologies. Defining the terms and tools of an experiment in such a way as to focus on just the right question while excluding bad data, well, this is where science becomes a high art. Whether it’s physics, medicine, psychology/neurology, human behavior, whatever, the structure of a great experiment is often as interesting, surprising, and elegant as the plot of a great novel — and the insights such experiments reveal are often astonishing.

Sometimes the results of powerful experiments open more questions (often deeper, and more significant questions) than they answer.

I think of things like the Turing Test (measuring the quality of an Artificial Intelligence) or the Double-Slit Experiment (illustrating the quantum nature of light). One of my favorites is a kind of test for autism:

Typically, the child watches two characters in a scene. The first person (A) places some chocolate in a drawer. A leaves the room and B (the second person) moves the chocolate to another drawer. A returns and the observer/child is asked to guess where A believes the chocolate is. Young children (< 3 years) and persons with autism mistakenly think that A will look in the new drawer, whereas older children will correctly claim that A will look in the old hiding place.

Another favorite of mine is the controversial experiment by Benjamin Libet in which it is shown that humans begin an action a full half a second after before their conscious mind makes the decision to act. A shocking conclusion, one that throws our understanding of consciousness and free will for quite a loop. But how in the world can this be measured? Libet’s methodology is ingenius:

Libet asked his experimental subjects to move one hand at an arbitrary moment decided by them, and to report when they made the decision (they timed the decision by noticing the position of a dot circling a clock face). At the same time the electrical activity of their brain was monitored. Now it had already been established by much earlier research that consciously-chosen actions are preceded by a pattern of activity known as a Readiness Potential (or RP). The surprising result was that the reported time of each decision was consistently a short period (some tenths of a second)after the RP appeared.

What are your favorite scientific experiments? I’d love to hear about them.

Tagging -2.0?

March 13th, 2006

I was at first impressed with the power of folksonomy-based tagging (in the sense of allowing users to invent their own taxonomies and metadata for information objects). But now I’m not so sure.

I just attended a SXSW panel called “Taxonomy 2.0″, with Tom Vander Wal, Prentiss Riddle, Rashmi Sinha, Adina Levin and moderated by Don Turnbull. They had a lot of really interesting and thought-provoking ideas about folksonomies and free-tagging, and I learned a lot.

But something stuck in my craw about the concept itself: The whole point behind the technological development of search engines, especially their artificial intelligence and natural language capabilities, is that humans shouldn’t be the ones indexing information. Folksonomies seem, to me, like a step backwards.

MORE…

ZOMBIES!

September 1st, 2001

I just re-read Alan Turing’s “Computing Machinery and Intelligence“, for something like the 5th time. It’s worth reading every few months or so just to see how cogent, thorough, and prescient it is. I can’t punch a hole in it. And everything I’ve ever read where someone else tries to punch a hole in it seems to use crazy circular and overly-complex logic… or worse, relies on intangible, even mystical ideas of what consciousness is. MORE…