Subjective Bayesian Networks and Human-in-the-Loop Situational Understanding

Abstract

In this paper we present a methodology to exploit human-machine coalitions for situational understanding. Situational understanding refers to the ability to relate relevant information and form logical conclusions, as well as identify gaps in information. This process for comprehension of the meaning information requires the ability to reason inductively, for which we will exploit the machines’ ability to ‘learn’ from data. However, important phenomena are often rare in occurrence with high degrees of uncertainty, thus severely limiting the availability of instance data for training, and hence the applicability of many machine learning approaches. Therefore, we present the benefits of Subjective Bayesian Networks—i.e., Bayesian Networks with imprecise probabilities—for situational understanding, and the role of conversational interfaces for supporting decision makers in the evolution of situational understanding.

This research was sponsored by the U.S. Army Research Laboratory and the U.K. Ministry of Defence under Agreement Number W911NF-16-3-0001. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the U.S. Army Research Laboratory, the U.S. Government, the U.K. Ministry of Defence or the U.K. Government. The U.S. and U.K. Governments are authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation hereon.

A Briefing Received by the Participants

A computer analysed the data of the German Stock Market Börse Frankfurt related to nine companies:

• Bayer, a pharmaceutical company

• Beiersdorf, a cosmetic company

• Henkel, a cosmetic company

• BMW, an automotive manufacturer

• Daimler, an automotive manufacturer

• Porsche, an automotive manufacturer

• Volkswagen, an automotive manufacturer

• Continental, a tyre manufacturer

In particular, the computer was programmed only to consider whether the closing value of a stock price was significantly different from same stock price at the closing time of the day before (±0.5%). And then the computer automatically derived possible dependencies between stocks.

Example. The Bayer stock value at the closing time on 7th December 2016 was 90.10; at the closing time on 8th December 2017 it was 93.17, thus with a significant change of 3.4%.

Similarly, the computer also analyses the changes of all the other companies considered in this study, thus producing a large table like the following:

Company	07/12/16	08/12/16	09/12/16	...
Bayer	Stable	Changed	Changed	...
Beiersdorf	Stable	Changed	Stable	...
Henkel	Stable	Stable	Stable	...
...	...	...	...	...

On the basis of such a large table, and by employing Machine Learning procedures, the computer identifies dependencies between companies’ stock values. An example of such a dependencies can be:

When Bayer stock price changes, there is low confidence that Henkel stock price is unlikely to change.