Martin Fujiki

I wanted to bring up the following issue in relation to Brian Scassellanti's talk for everyone to think about–in particular his comment that no one has yet achieved transfer in their work with robots and children with autism.

The issue of transfer has been a major problem in many aspects of speech language pathology. Training has frequently been approached by breaking down a behavior to its component parts and then putting those parts back together again as the client masters them. At each level the behavior is made more complex–more like the behavior you want to end up with. For example, in teaching a child to articulate the r sound, the clinician might begin with the r in isolation (for example, er). Once the child can do this, the sound is practiced in syllables or simple words (e.g., water). Next the child would work on the sound in simple sentences (e.g., I want water). At each step the sound is worked on in more and more naturalistic contexts.

This type of approach works well for some behaviors, but for more complex behaviors results are often less than satisfactory. One way that clinicians have tried to get around the transfer problem is by working on behaviors in more naturalistic contexts from the beginning (e.g., work on the production of the r sound in conversational speech to begin with).

I bring this up only to point out that we are going to have to be very clever as we design our intervention. Breaking the behavior into small parts and then moving closer and closer toward the ultimate goal seems logical, but has a long history of failure.

Martin Fujiki


Goodrich's response

Transfer from laboratory robotics to practice is currently my biggest concern about trying to use robots to help children with autism. As we think about transfer, I'd like us to explore other uses of robots to help children with autism, including those that may extend beyond clinical work. Here are some brainstorm ideas; perhaps some of these ideas will be the seed for future work that may differentiate us from others working in the field.

Compliance: I'm sure that compliance is probably the wrong word, but I don't know what to call the need to get a child's attention, keep his or her attention, disrupt a child's nonproductive/violent behavior, help a child prepare for and cope with a schedule change, etc. Could robots be useful in helping with compliance needs by facilitating productive behavior? If so, this could be very useful for care-givers, parents, teachers, and therapists.

“Assistive Device:” I picture using a robot not just in therapies but as a device similar to reading glasses – something often with the child that helps them be higher functioning. Are there ways that a robot could be useful in practice by focusing attention, facilitating communication, etc.?

“Parent Coping Tool:” There are many anecdotes from parents/care-givers of a child with severe autism where the parents are touched when the child displays even simple social interaction such as eye contact or joint attention. Could a robot be used at home in a way that a child exhibits some of these behaviors often enough that parents are better able to cope with their challenges by seeing (realistic) reasons for hope?

Mike Goodrich


« Discussion Page

ar/on-transfer.txt · Last modified: 2014/08/11 19:59 by tmburdge
Back to top
CC Attribution-Share Alike 4.0 International
chimeric.de = chi`s home Valid CSS Driven by DokuWiki do yourself a favour and use a real browser - get firefox!! Recent changes RSS feed Valid XHTML 1.0