Posted on 05/06/2007 4:53:04 PM PDT by blam
Source: American Roentgen Ray Society
Date: May 6, 2007
Voice Recognition Systems Seem To Make More Errors WIth Women's Dictation
Science Daily — There is a significantly higher rate of transcription error in women compared to men when using commercial voice recognition applications, according to a recent study. "Our residency program and department recently made the transition to speech recognition from a digital dictation system," said Syed Ali, MD, lead author of the study. "This prompted us to ask research questions about how to increase our accuracy rates and what factors adversely impacted speech recognition," said Dr. Ali.
Ten radiology residents, five male and five female, were each trained on a commercial speech recognition application. Each resident was asked to dictate a standardized set of ten radiology reports containing a total of 2,123 words. Utilizing a commercial software solution, the generated reports were then compared with the original reports and error rates were calculated. The error rate was defined as the sum of the number of word insertions and deletions divided by the total word count for a given report.
According to the study, error rates in the male population ranged from 0.025 to 0.139 while the error rates in the females ranged from 0.015 to 0.206. The results show a higher rate of recognition error in the females compared to the males.
"The immediate impact of the study for radiologists is an increased level of awareness that women may need to spend more time training on the system than their male counterparts and may have to work somewhat harder to make the system successful," said Dr. Ali. "This could include using macros or actually altering dictation style to increase recognition rates," he said.
"Any efforts to improve recognition rates will have a positive impact on physicians and patients of course by reducing error rates and improving productivity," said Dr. Ali. The full results of this study will be presented on Monday, May 7, 2007 during the American Roentgen Ray Society's annual meeting in Orlando, Florida.
Note: This story has been adapted from a news release issued by American Roentgen Ray Society.
Well that explains alot. I mean if a computer can make mistakes, why can’t I? I am printing this for future self-defense against my girlfriend.
This might be a reason:
“Girls speak faster on average — 250 words per minute versus 125 for typical males.”
http://itre.cis.upenn.edu/~myl/languagelog/archives/003423.html
Discrimination! We need to lower male voice recognition rates to that of female!
i seem to have the same problem...
Oops! I sense a NOW/ACLU moment here...<.P>
chauvinistic computers? lol
This is too funny.
I’m female and a medical transcriptionist and boy can I vouch for this one!! First off... there are a lot of logical, well-spoken women working as physician assistants, nurse practitioners, therapists of all sorts, and physicians. However, being logical and well-spoken is not a requirement for dictating patient notes. There are also a lot of illogical, completely indecipherable male voices.
However, as a rule the women start out by being more difficult by having higher voices. As our hearing abilities age, we lose the higher pitches earlier than the lower pitches.
However (and here come the flames), on average the females dictate more words, a lot more corrections (”go back and change where I said this to that”), and much less ordered notes than the guys.
The opposite has been true with surgical dictations for some reason - The female-dictated notes I type up for that are better than the guys.(maybe the women that go into this field think more directly and orderly than their internal medicine counterparts).
Voice-recognition systems must be intrinsically male, since they apparently also fail to pay attention to women’s speech...
I just thought of a good idea for a SNL script. Have the voice response system talk back to people in their “dialect”. Black guy from da hood gets a response back in jive, islamic fundamentalist gets response in terrorist lingo, etc. Maybe I should send it in.
Woman say four times as many words a day than men...and during an argument, thst rate soars to ten times more.
Maybe that's why I can't stand to watch a talk show with only women.
The speech/voice recognization chip was invented by a male.
Danger Will Robinson! Go into the bunker!
This is old news and is due in large part to the fact that as is the case with all core technologies, the mathematical algorythms used by the DSP's were developed by men and largely lab tested by men.
I’ve the same prollem. It must be a male thing....
They don’t listen either...
With gusts up to 375 words per minute...
ahh the crux of the problem.. we need more women writing software
Maybe that's why I can't stand to watch a talk show with only women.
They may say four to ten times more words, but they certainly don't communicate four to ten times as much information. That's why I can't tolerate women's talk shows.
It's like going into the hen house after all the chickens have laid an egg. A rooster makes a lot of noise, but only once a day.
I bought some noise cancelling headphones, but was dusappointed to find that I still hear my wife just fine when using them.
I’m also an MT and I cringe when I see that the dictator is female. I’m not sure what it is, but they seem to dictate too slowly, and in a sing-songy sort of voice that lulls me - which makes me type slower. The male docs, for the most part, seem to know what they want to say, and they don’t vary too much from report to report - much easier IMO.
Ummmm...yeah. The results also show a lower rate.
Perhaps the computer needs a how to interprint “nagging” algorithm.
and
could be the fact there is no “reading your mind” algorithm? (I didn’t MEAN to say that you should have known that!)
Peter Pan: Girls talk too much.
Wendy: -laughing- Yes, girls talk too--- Hmmm? Oh.
no nagging interpritation algorithm.
I have very bad hearing and using a phone is a struggle for me. But that is getting to be a big problem professionally as my job requires me to talk on phones daily, often with people in Singapore, who speak great English with a totally amazing accent.
I have been toying with the idea of buying a voice recognition program and gearing it to run simultaneously with my phone calls in the hope that the combination of both listening and filling in with what the voice recognition software picks up would boost overall comprehension and speed.
I understand Rush used a similar system when he was having bad hearing problems.
Any comments, suggestions?
Just the obvious. Ask Rush what he used...
I agree.
However, I think Rush still uses something daily during his show.
“Voice Recognition Systems Seem To Make More Errors WIth Women’s Dictation”
An inference could be made that men are better at giving Dictation.
This has been known for decades.
Females, as a rule, must consciously modify their natural vocal tones, when using radios, telephones and microphones, which is virtually every form of sound amplification or recording device.
My “telephone voice” is measurably different from my natural voice.
Learning to correctly adjust your vocal pitch when using communications devices is very simple, and quickly becomes automatic.
OTOH, females generally have more natural sensitivity to hearing complex sounds at higher ranges, and identifying problems or issues in sound reproduction and audio equipment calibration.
IE: most females can actually hear the sound that white noise generators produce, and they are actually a constant irritant, not an effective office noise suppression, sound cancellation measure.
Voice recognition diction, tried something like it years ago & soon as I saw how my words are getting translated, it came up with some of the funniest word combinations in response to my laughter, which made me laugh even harder.
Dragon Naturally Speaking 9 is awesome. I haven’t tried it with telephone input, but with a good microphone and a Core Duo processor, it has amazing performance. I see it on eBay now for $63, including shipping.
Thanks, gcruse...that’s too much.
“the women should have to tell them...”
Whoops, should be: the women shouldn’t have to tell them...
I'm a retired chip-maker...that chip/system hasn't been invented...yet.
I’m wondering if this could be the same problem that the human ear has in understanding women’s singing voices. It has to do with the pitch. I know a study has been done with opera singers and it is universally known that the higher the notes, the harder it is to understand the words of female singers, regardless of her good diction.
Maybe the word recognition program needs some tweeking to be able to recognize words in higher pitches. Was the technology developed using male voices first? I wonder.
The idea that women speak faster than men, may be somewhat true, but I would expect that with this study, all the readers were trained to speak at a steady pace and with good diction. So, I still think it may be the pitch.
HA! Try being Southern and using one of these things. When I call information now, I just keep saying operator until it transfers me.
“I’m a retired chip-maker...that chip/system hasn’t been invented...yet.”
Hey, if they can expect it of men, they can certainly expect it of computer chips...
Has anyone thought about the fact that most voice recognition systems were designed, built, and tested by a majority male population?
Now, mash 1 for help
That's interesting. But in general, I would suspect that, as you say, most voice recognition systems were developed using male voices, since it was probably expected that most users (business execs, etc.) would be male.
I use one in my translation work, and I have found that it takes one thing: lots of RAM on the PC. The more you use it, stopping to train it every now and then, the better it gets. But it has to have enough RAM or it will be so slow you could type the text 20 times faster.
That said, I do notice that if I speak loudly and at a lower, more "male" pitch, it understands better.
Did I miss any Leftist clichés?
Maybe the software gets bored.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.