Optimization of Connectionist Knowledge
An Idea for Improving Organisational Learning

Gunter Heim, Aachen
July 2004

Abstract

There is a large potential for increasing the intelligence of large organisations such as, for example, commercial companies. The basic idea to be presented here rests on three contributary concepts.

  1. The notion of alternating task-oriented activation patterns
  2. Constant generation of suggestions on how to be more successful
  3. Constant evaluation of the desired result i. e. the organisation`s success

By constantly generating and testing hypotheses about who whould communicate with whom or access what information when organisational learning could be turned into a semi-autonomous, software supported mechanism.

1.The notion of alternating
task-oriented activation patterns

In most organisations, such as companies for example, most people are involved in different projects or tasks at different times. The picture below shows 3 people, one intranet site and some drawers containing some folders.
On Monday morning the company wants to have a rough financial planning of a new production plant to be done. To get the job done the people and the other sources of information are of different relevance. Mrs. Miller, who is an expert on the subject and the folders with archived versions of financial planning processes are of great relevance. In the communicational activity that goes hand in hand with the task, these two play a prominent role. Mrs. Miller will send a lot of mails to coordinate her colleagues, she will phone a lot and she will have to access many documents. The folders are more passive, but they will be accessed many times by many of the people involved.
On Monday afternoon, however, the company wants a rough technical design of the new production plant to be done. This time it is Hugh Cox from the Machine Department and the intranet site on strategic planning that are most important. Hugh will need to know about all the other people`s activites and the intranet site will be accessed by many people many times as it provides much useful information on the process.
This is what is meant by alternating task-oriented activation patterns. The picture below shows an imaginary view of the communicational activity accumulated over a few hours. The thicker the line connecting the people or sources of information the more emails, data transfers, phone calls etc. have taken place.

The idea is that such activation patterns may actively be supported and improved by way of a communication software. The technical realisation would be nothing new: By placing links to websites or contact dates more prominent on a browser interface the user will be more likely to follow the link or contact a person. By hiding a link or a contact information so that the user whould have to circumstantially search it with many clicks and scrolls the user is more likely not to follow the link or contact the person.That is one way of directing information flows. The picture below shows the browser setting of Jane Ball while helping with the financial planning of the future production plant.

Goto another site: The notion of communicational synapses
The names and documents that are placed most prominently on her computer interface are quite likely to be given much attention by her. Names and documents that she whoud have to access by circumstantial navigation have a chance of being neglected.
Therefore, by adapting the placing of names, documents and other such links to further information on a browser interface one may influence the activation patterns within a company according to the current task to be done. This, in short, is the notion of task-oriented activation patterns.

2. Constant generation of suggestions
on how to be more successful
Let`s assume that within a large, international company, projecting small production plants worldwide is a standard business process that is carried out once or twice a week. In order to compete best on the global market, the company constantly considers where it could best produce, and many alternatives are suggested all the time. A small department is responsible for quickly assessing the financial aspects of such suggestions. This department, headed by Mrs. Miller, usually has a week or two time to write a short report on the financial aspects of a suggested new locality.
The communication that characterizes this business process could now be made the object of suggestions on how to improve it. It is useful to categorize such suggestions along two criteria: a) who makes the suggestions and b) are the suggestions concrete or abstract. Here are some examples to illustrate the kinds of suggestions that might be formulated to improve the communication in the case of financial planning of projected construction plants. Note that there is a shift from concrete knowledge to general learning rules as the level of abstraction increases.

Insiders making quite concrete suggestions:

  • Mr. Branson of the technical department should be informed about new projects as early as possible. He has a lot of experience that he is very willing to communicate.
  • If liquidity over the coming two years is being discussed as critical then the Investment Department should be involved.
  • Mails from Mr. Burlington from the strategic planning department are scarce but usually of high relevance. They should enjoy a high priority!

Insiders making rather abstract suggestions:

  • Intranet sites are generally poorly updated. They are ususally of little value and more annoying than helpful.
  • Towards the end of an assessment process communication with people or information from outside the Financial Department becomes rapidly less relevant.
  • Intranet sites that hardly change their contents over time and that are still accessed often by many people seem to be of particularly high relevance.

Outsiders making very abstract suggestions:

  • Recurring paths of spreading email-activity should be enforced by the software system to reduce time lags.
  • Communicational paths that differ from the standard activation patterns may be highy relevant.
  • If department A and department B show a rapid increase in internal communicational activity the chances are high that an unexpected problem as regards the expected liquidity of a projected plant has arisen.
The kind of suggestion that is characterized by the examples above does, however, address temporal cirucumstances that are very likely to change over time. People who are relevant today may move to another department and thereby become irrelevant. The overall relevance of intranet sites might change dramatically as the team in charge of them is allocated more money and personnel. And new technological developments may result in the electrical infrastructure becoming far less cost-relevant as it used to be.
The sources of such suggestion can be manifold. People working within a company may wish to formulate them, external experts may be paid to make them and even pieces of software may look for them using heuristic methods.
But all such suggestions on communictional improvements must be critically tested. The changeability of circumstances is one reason. Another reason and very obvious reason why such suggestions must be critically tested is the simple fact that they may be wrong. Not anyone who has to say something says something useful. A third reason why suggestions have to be tested is the sheer number of possible suggestions that could be made.

3. Constant evaluation of the desired result
If a certain task is carried out a number of times and if the communicational patterns as well as the final success of the job are logfiled then statistical correlations between the communication patterns and the success can (cautiously) be found out. Each line in the picture below stands for one round of a financial assessment carried out by a team of workers.
At the beginning, in July 2002, the task was to assess the financial risks and chances of a new production plant to be erected somewhere in Brazil. The sketch of the communicational pattern shows a rather democratic communication structure. Many of the nodes of the network (coloured dots) are interconnected by strong lines. In reality that would mean that everybody ranks high on everybody else`s computer interface presentation. However, after the job had been done, the quality of the work was rated as poor. By the way, it doesn`t matter here how the rating is done. It may be carried out by the team itself, by some superiors or another department. In November of the same year, the team has slighly changed its pattern to reduce information overload. Still the rating remains poor. Then someone suggests that the following time it should be tried to have install the green person as manager of the process and make her report to the red person. The red person should only evaluate important internal results and adivse on the general line. When tried out in April 2003 on the planning of a production plant for the Mozambiquan port of Quelimane the success of the team rose dramatically. Shortly afterwards, in May 2003, the team is to assess a plant in Durban. As the red person has many other jobs to do, it cannot give much attention to the planning process. As expected, the success of the team drops again. When in June 2003 the team has to assess a plant to be built in Swaziland, the red person again gives much attention to the process. The success is back again.
This episode should have illustrated what is meant by an evaluation of the desired result.

Conclusion

In summary, it is being suggested to install the following routine in companies to constantly adapt their communicational reality to ever changing circumstances:

Some related sites in English:

Constructed Brain: A test environment A bio-inspired constructivist approach to company intelligence
Communicational Synapses The bio-inspired idea of a synaptic browser
The neural company Some rudimentary ideas on analogies between companies and mechanisms of human intelligence

Some related websites in German:

Skatbewertung: The contribution of individual people to teams Lifelong statical scanning of individual people
E-Controlling: Typical episodes in the life of companies In industry and commerce, certain situations are not uncommon.
Mailwatcher: Kommunikationsmuster im Mailverkehr Kann man Mailing-Muster kategorisieren?
Die Idee des Synaptic Browsers Mensch-Maschinen-Schnittstellen als Synapsen
Unternehmens als neuronal-psychische Intelligenzform: spekulative Gedanken Private Homepage über Quantenphysik, Religion, Weltprozess, neuronale Unternehmen, evolutionäre Ökonomie etc.
Neuronale Unternehmen Hobby-Philosophie

E-Mail Adresse // Last edited: July 6th, 2004
The Learning Organisation - How to make it more systematic

Optimization of Connectionist Knowledge
An Idea for Improving Organisational Learning

Gunter Heim, Aachen
July 2004

Abstract

There is a large potential for increasing the intelligence of large organisations such as, for example, commercial companies. The basic idea to be presented here rests on three contributary concepts.

  1. The notion of alternating task-oriented activation patterns
  2. Constant generation of suggestions on how to be more successful
  3. Constant evaluation of the desired result i. e. the organisation`s success

By constantly generating and testing hypotheses about who whould communicate with whom or access what information when organisational learning could be turned into a semi-autonomous, software supported mechanism.

1.The notion of alternating
task-oriented activation patterns

In most organisations, such as companies for example, most people are involved in different projects or tasks at different times. The picture below shows 3 people, one intranet site and some drawers containing some folders.
On Monday morning the company wants to have a rough financial planning of a new production plant to be done. To get the job done the people and the other sources of information are of different relevance. Mrs. Miller, who is an expert on the subject and the folders with archived versions of financial planning processes are of great relevance. In the communicational activity that goes hand in hand with the task, these two play a prominent role. Mrs. Miller will send a lot of mails to coordinate her colleagues, she will phone a lot and she will have to access many documents. The folders are more passive, but they will be accessed many times by many of the people involved.
On Monday afternoon, however, the company wants a rough technical design of the new production plant to be done. This time it is Hugh Cox from the Machine Department and the intranet site on strategic planning that are most important. Hugh will need to know about all the other people`s activites and the intranet site will be accessed by many people many times as it provides much useful information on the process.
This is what is meant by alternating task-oriented activation patterns. The picture below shows an imaginary view of the communicational activity accumulated over a few hours. The thicker the line connecting the people or sources of information the more emails, data transfers, phone calls etc. have taken place.

The idea is that such activation patterns may actively be supported and improved by way of a communication software. The technical realisation would be nothing new: By placing links to websites or contact dates more prominent on a browser interface the user will be more likely to follow the link or contact a person. By hiding a link or a contact information so that the user whould have to circumstantially search it with many clicks and scrolls the user is more likely not to follow the link or contact the person.That is one way of directing information flows. The picture below shows the browser setting of Jane Ball while helping with the financial planning of the future production plant.

Goto another site: The notion of communicational synapses
The names and documents that are placed most prominently on her computer interface are quite likely to be given much attention by her. Names and documents that she whoud have to access by circumstantial navigation have a chance of being neglected.
Therefore, by adapting the placing of names, documents and other such links to further information on a browser interface one may influence the activation patterns within a company according to the current task to be done. This, in short, is the notion of task-oriented activation patterns.

2. Constant generation of suggestions
on how to be more successful
Let`s assume that within a large, international company, projecting small production plants worldwide is a standard business process that is carried out once or twice a week. In order to compete best on the global market, the company constantly considers where it could best produce, and many alternatives are suggested all the time. A small department is responsible for quickly assessing the financial aspects of such suggestions. This department, headed by Mrs. Miller, usually has a week or two time to write a short report on the financial aspects of a suggested new locality.
The communication that characterizes this business process could now be made the object of suggestions on how to improve it. It is useful to categorize such suggestions along two criteria: a) who makes the suggestions and b) are the suggestions concrete or abstract. Here are some examples to illustrate the kinds of suggestions that might be formulated to improve the communication in the case of financial planning of projected construction plants. Note that there is a shift from concrete knowledge to general learning rules as the level of abstraction increases.

Insiders making quite concrete suggestions:

  • Mr. Branson of the technical department should be informed about new projects as early as possible. He has a lot of experience that he is very willing to communicate.
  • If liquidity over the coming two years is being discussed as critical then the Investment Department should be involved.
  • Mails from Mr. Burlington from the strategic planning department are scarce but usually of high relevance. They should enjoy a high priority!

Insiders making rather abstract suggestions:

  • Intranet sites are generally poorly updated. They are ususally of little value and more annoying than helpful.
  • Towards the end of an assessment process communication with people or information from outside the Financial Department becomes rapidly less relevant.
  • Intranet sites that hardly change their contents over time and that are still accessed often by many people seem to be of particularly high relevance.

Outsiders making very abstract suggestions:

  • Recurring paths of spreading email-activity should be enforced by the software system to reduce time lags.
  • Communicational paths that differ from the standard activation patterns may be highy relevant.
  • If department A and department B show a rapid increase in internal communicational activity the chances are high that an unexpected problem as regards the expected liquidity of a projected plant has arisen.
The kind of suggestion that is characterized by the examples above does, however, address temporal cirucumstances that are very likely to change over time. People who are relevant today may move to another department and thereby become irrelevant. The overall relevance of intranet sites might change dramatically as the team in charge of them is allocated more money and personnel. And new technological developments may result in the electrical infrastructure becoming far less cost-relevant as it used to be.
The sources of such suggestion can be manifold. People working within a company may wish to formulate them, external experts may be paid to make them and even pieces of software may look for them using heuristic methods.
But all such suggestions on communictional improvements must be critically tested. The changeability of circumstances is one reason. Another reason and very obvious reason why such suggestions must be critically tested is the simple fact that they may be wrong. Not anyone who has to say something says something useful. A third reason why suggestions have to be tested is the sheer number of possible suggestions that could be made.

3. Constant evaluation of the desired result
If a certain task is carried out a number of times and if the communicational patterns as well as the final success of the job are logfiled then statistical correlations between the communication patterns and the success can (cautiously) be found out. Each line in the picture below stands for one round of a financial assessment carried out by a team of workers.
At the beginning, in July 2002, the task was to assess the financial risks and chances of a new production plant to be erected somewhere in Brazil. The sketch of the communicational pattern shows a rather democratic communication structure. Many of the nodes of the network (coloured dots) are interconnected by strong lines. In reality that would mean that everybody ranks high on everybody else`s computer interface presentation. However, after the job had been done, the quality of the work was rated as poor. By the way, it doesn`t matter here how the rating is done. It may be carried out by the team itself, by some superiors or another department. In November of the same year, the team has slighly changed its pattern to reduce information overload. Still the rating remains poor. Then someone suggests that the following time it should be tried to have install the green person as manager of the process and make her report to the red person. The red person should only evaluate important internal results and adivse on the general line. When tried out in April 2003 on the planning of a production plant for the Mozambiquan port of Quelimane the success of the team rose dramatically. Shortly afterwards, in May 2003, the team is to assess a plant in Durban. As the red person has many other jobs to do, it cannot give much attention to the planning process. As expected, the success of the team drops again. When in June 2003 the team has to assess a plant to be built in Swaziland, the red person again gives much attention to the process. The success is back again.
This episode should have illustrated what is meant by an evaluation of the desired result.

Conclusion

In summary, it is being suggested to install the following routine in companies to constantly adapt their communicational reality to ever changing circumstances:

Some related sites in English:

Constructed Brain: A test environment A bio-inspired constructivist approach to company intelligence
Communicational Synapses The bio-inspired idea of a synaptic browser
The neural company Some rudimentary ideas on analogies between companies and mechanisms of human intelligence

Some related websites in German:

Skatbewertung: The contribution of individual people to teams Lifelong statical scanning of individual people
E-Controlling: Typical episodes in the life of companies In industry and commerce, certain situations are not uncommon.
Mailwatcher: Kommunikationsmuster im Mailverkehr Kann man Mailing-Muster kategorisieren?
Die Idee des Synaptic Browsers Mensch-Maschinen-Schnittstellen als Synapsen
Unternehmens als neuronal-psychische Intelligenzform: spekulative Gedanken Private Homepage über Quantenphysik, Religion, Weltprozess, neuronale Unternehmen, evolutionäre Ökonomie etc.
Neuronale Unternehmen Hobby-Philosophie

E-Mail Adresse // Last edited: July 6th, 2004