Complex Systems Discussion Group

 Forgot Password
 Register now

QQ登录

只需一步,快速开始

Search
View: 7796|Reply: 7

Understanding the Traditional Chinese Medicine by Artificial Neuron Network

[Copy Link]

43

Threads

186

Posts

524

Credits

Moderator


IP  123.117.20.52

Posted at 2013-9-15 14:07:34 | All floors |Read mode
Here is the project I propsed for the PKU-Heb brain storming: Understanding the Traditional Chinese Medicine by Artificial Neuron Networks

Background: While there are many efforts trying to identifyeffective compounds from herbs being used by Traditional Chinese Medicines (TCM)and understand the molecular targets of these effective compounds, little hasbeen done in understanding the theories of TCM. The old theories of TCM areindeed very elusive and contain the interactions of “five element” and thebalance between “Yin” and “Yang”, which can not be found in the realisticbiological systems.

Hypothesis: However, this theories are often helpful in guidingdoctors to prescribe effective treatment to the patient given the symptom. Here,I suspect that these theories are actually an “effective mapping”—an artificialneuron network formed by thousands of years’ experience, which maps symptoms tothe right combination of herbs. These fancy “five element” and other conceptmay not be the real existence, but the virtual node inside the artificial neuron network that maps input to output.

Aritifical Neuron Network:


TCM theory in my view:


Method: The method to testify this hypothesis is tore-construct the artificial neuron network from literature. There are databasefor thousands of prescriptions given symptoms. We need to digitalize these symptomsand prescriptions, and obtain the possible neuron network structures that bestmap the symptom to prescriptions, using in-silicon evolution or other globaloptimization methods. The obtained neuron networks will be compared to thecurrent structure of TCM theories,  andreveal more insights into the essence of TCM.

This artificial neuron network may be used to predict novel therapeuticstrategies for newly appeared disease like cancer or HIV.

This post contains more resources

You need to Login Before they can download or view, Not have an account?Register now

x
给我一块二向箔,清理用

2

Threads

7

Posts

41

Credits

Moderator

QQ

IP  162.105.160.12

Posted at 2013-9-18 11:24:36 | All floors
It's a good point for TCM. I think the important concept in TCM is "Yinyang", means balance, which lead to system control. I think the different network of TCM between natural network is the network of TCM may need a probability interpretation. The input may have different interaction type of hidden node by different probability, what we need to find is the interaction that has the maximum probability

0

Threads

1

Posts

12

Credits

Newbie Member


IP  115.27.39.49

Posted at 2013-9-18 14:13:14 | All floors
It is quite interesting to link the apprehensible neural systems with the complicated TCM; but from a evolution perspective, the neural network is evolving with no direction by optimizing the structures to adapt to the environment, while the complex connections between the symptoms and priscriptions may be static or evolve in other directions.

1

Threads

6

Posts

146

Credits

Registered member


IP  114.249.136.119

Posted at 2013-9-23 14:44:36 | All floors
It's a great idea but might be a bit dangerous. The “five element” and the balance between “Yin” and “Yang” are too simplified as hidden node which will be difficult to find relationships to either symptoms or prescriptions. For instance, from one's face, we could hardly tell whether it is "wood" or "water", but it can have some relationships to either "kidney" or "liver", which will show clues to the right prescription. However, the artificial neuron network is a good method to simplify the relationships between symptoms and prescriptions (which is actually how doctors are thinking while working), but I'm suggesting whether it will to better to make the organs (such as heart, liver, spleen, lung, and kidney) as hidden node instead of “five element”. However, if "Yin" and "Yang" are the key points we want to discuss, we would have to go a bit further into the traditional medical literature.

1

Threads

7

Posts

29

Credits

Newbie Member


IP  162.105.60.102

Posted at 2013-9-23 15:18:59 | All floors
I'm wondering how the database is like. But when it refers to mapping the symptom to prescriptions, the critical point is the quantity.  Because the prescription is a combination of different TCMs, the ratio between the medicines cannot be ignored. (As is shown in Prof.Lai's paper)  

However, it might be interesting to use this concept to analyze the symptom. Say, to figure out the most accurate way to diagnose a disease.

Besides, I think at that time, people believed everything can be clustered into "five elements", and we might get some clues from those books written in classical Chinese Like <黄帝内经>
Working makes you tired; thinking drives you crazy!

1

Threads

8

Posts

30

Credits

Newbie Member


IP  162.105.248.61

Posted at 2013-9-23 16:54:29 | All floors
I think it seems that it is easy to use TCM theory to find the correlation between external symptoms and hidden node. The actual difficult part is understanding the prescriptions.
The TCM prescription is not only a simple combinations of several herbs, but a mixture of some therapeutic compounds. Because the active compounds come from different sources, such as a certain herb, the reaction between different herbs and "Pao Zhi", we cannot briefly digitalize the prescription.  

3

Threads

27

Posts

107

Credits

Registered member


IP  162.105.248.165

Posted at 2013-9-24 09:26:59 | All floors
Traditional science is based on experience, not only medicine. Notional connections to 阴阳五行 are common in other fields such as building for theory support. These experience perhaps implicate complex connection between organs. But I don't think insist on notion like 阴阳 is essential.

43

Threads

186

Posts

524

Credits

Moderator


IP  169.230.219.58

 Author| Posted at 2013-9-28 20:47:38 | All floors
wuhongkun Posted at 2013-9-23 16:54
I think it seems that it is easy to use TCM theory to find the correlation between external symptoms ...

I think for prescription, one can digitalize by the form and amount of herbs, and there should be already database about herbs


给我一块二向箔,清理用
You need to log in before you can reply Login | Register now

This forum Credits Rules

Mobile ver.|Darkroom|Complex Systems Discussion Group

GMT+8, 2018-12-13 12:04 , Processed in 0.082272 second(s), 23 queries .

Powered by Discuz! X3 English ver.

© 2001-2013 Comsenz Inc.

Quick Reply Back to top Back to list