artificial intelligence picks the right formula

August 5, 2006

I have no idea how valid this study actually is, but it looks like robots

can recognize the elegance of TCM formulae.

Check this out:

J Ethnopharmacol. 2006 Jun 29; [Epub ahead of print]

Usefulness of traditionally defined herbal properties for distinguishing

prescriptions of traditional Chinese medicine from non-prescription recipes.

Ung CY, Li H, Kong CY, Wang JF, Chen YZ.

Bioinformatics and Drug Design Group, Department of Pharmacy, National

University of Singapore, Blk SOC1, Level 7, 3 Science Drive 2, Singapore

117543, Singapore; Bioinformatics and Drug Design Group, Department of

Computational Science, National University of Singapore, Blk SOC1, Level 7,

3 Science Drive 2, Singapore 117543, Singapore; Department of Biochemistry,

The Yong Loo Lin School of Medicine, National University of Singapore, Blk

MD7, #02-03, 8 Medical Drive, Singapore 117597, Singapore.

Traditional Chinese medicine (TCM) has been widely practiced and is

considered as an attractive to conventional medicine. Multi-herb recipes

have been routinely used in TCM. These have been formulated by using

TCM-defined herbal properties (TCM-HPs), the scientific basis of which is

unclear. The usefulness of TCM-HPs was evaluated by analyzing the

distribution pattern of TCM-HPs of the constituent herbs in 1161 classical

TCM prescriptions, which shows patterns of multi-herb correlation. Two

artificial intelligence (AI) methods were used to examine whether TCM-HPs

are capable of distinguishing TCM prescriptions from non-TCM recipes. Two AI

systems were trained and tested by using 1161 TCM prescriptions, 11,202

non-TCM recipes, and two separate evaluation methods. These systems

correctly classified 83.1-97.3% of the TCM prescriptions, 90.8-92.3% of the

non-TCM recipes. These results suggest that TCM-HPs are capable of

separating TCM prescriptions from non-TCM recipes, which are useful for

formulating TCM prescriptions and consistent with the expected correlation

between TCM-HPs and the physicochemical properties of herbal ingredients

responsible for producing the collective pharmacological and other effects

of specific TCM prescriptions.

PMID: 16884871 [PubMed - as supplied by publisher]

August 5, 2006

Just as importantly, what is the practical use of this technology that can

accurately distinguish between classic formulas and random lists of herbs?

Is this research about artificial intelligence or TCM?

-al.

On 8/5/06, Tom Verhaeghe <tom.verhaeghe wrote:

>

> I have no idea how valid this study actually is, but it looks like

> robots

> can recognize the elegance of TCM formulae.

>

> Check this out:

>

> J Ethnopharmacol. 2006 Jun 29; [Epub ahead of print]

>

--

Pain is inevitable, suffering is optional.

August 5, 2006

This is the application of a supervised learning algorithm to a new

domain (TCM formulas) for the sake of seeing how " learnable " the TCM

formula domain is ( " learnability " here has nothing to do with human

learning). These results would be publishable in the U.S. as an

original application of known AI techniques, but it would not (as

described) qualify as a contribution to AI since this type of

learning algorithm has been around for about 15 years now.

As far as practical applications go, an herbal pharmacy that fills

prescriptions for a large number of herbalists could conceivably use

the program to " red flag " prescriptions that were not recognized as

feasible TCM-formulas in order to bring the prescription to the

attention of an experienced herbalist. That person could then either

override the program (because they know the prescription is fine as

it stands) or contact the author of the prescription to make sure

there are no errors in the prescription. So such a thing could

conceivably add some quality control to the process of filling

herbal prescriptions - although I don't know if prescription errors

are enough of an ongoing concern to justify the extra effort of

typing all incoming prescriptions into a computer program.

With additional work, a program along these lines might also be

useful as an automated tutor for students who are getting started

writing their own formulas. But students need useful feedback for a

formula that does not look sound: it's not enough to just give it a

thumbs up/down. And this system, as it stands, would not be able to

say anything useful/instructional about a suspect formula. So it's

an interesting little result which could be pushed in more useful

directions if someone wanted to keep working on it. Useful things

often start out as baby steps in a vague direction.

Wendy Lehnert (Professor)

Department of Computer Science

University of Massachusetts - Amherst

>

> Just as importantly, what is the practical use of this technology

that can

> accurately distinguish between classic formulas and random lists

of herbs?

> Is this research about artificial intelligence or TCM?

>

> -al.

>

> On 8/5/06, Tom Verhaeghe <tom.verhaeghe wrote:

> >

> > I have no idea how valid this study actually is, but it looks

like

> > robots

> > can recognize the elegance of TCM formulae.

> >

> > Check this out:

> >

> > J Ethnopharmacol. 2006 Jun 29; [Epub ahead of print]

> >

>

> --

>

> Pain is inevitable, suffering is optional.

>

August 5, 2006