Social interaction opportunity detection method and system

1. A method for improving an interaction process of a user in a social network with at least one other person, the method comprising steps of:

in a training phase,

acquiring sensor data on at least a user state of the at least one other person and interaction data on a social interaction of the user with the at least one other person, wherein the acquired sensor data concerns at least one of user activity and user emotion of the at least one other person,

analyzing the sensor data and the interaction data to generate classifier data from the acquired sensor data and the acquired interaction data,

wherein analyzing the sensor data and the interaction data comprises learning a correlation between the acquired sensor data and the acquired interaction data for generating the classifier data, and the generated classifier data comprises a learned classifier for the user and the at least one other person,

wherein the learned classifier determines an interaction initiation success for the user and the at least one other person, or the learned classifier determines an interaction initiation or an initiation acceptance success for the user and the at least one other person, and

in an application phase,

acquiring current sensor data on at least a current user state of the at least one other person, wherein the acquired current sensor data concerns at least one of user activity and user emotion of the at least one other person,

predicting an interaction score for the user based on the acquired current sensor data and the classifier data,

generating an interaction identifier for the at least one other person for an interaction with the at least one other person in the social network based on the predicted interaction score.

2. The method according to claim 1 , wherein

the acquired sensor data and the acquired current sensor data further comprise user location.

3. The method according to claim 1 , wherein

the acquired interaction data includes at least one of information on an initiated interaction on the social network, acceptance of the initiated interaction on the social network, and rejection of the initiated interaction on the social network.

4. The method according to claim 1 , wherein

acquiring the sensor data comprises acquiring sensor data on a user state of the user and further acquiring sensor data on at least the user state of the at least one other person, and

acquiring the current sensor data comprises acquiring sensor data on a current user state of the user and further acquiring sensor data on the current user state of the at least one other person.

5. The method according to claim 4 , wherein, in the training phase, the method further comprises steps of

generating a generic model,

learning an adaptation function based on the generic model and the acquired sensor data on the user state of the user and the interaction data on the social interaction of the user with the at least one other person, and,

in the step of analyzing the sensor data,

generating the classifier data based on the learned adaptation function and the generic model.

6. The method according to claim 1 , wherein

acquiring the sensor data includes acquiring the sensor data from at least one sensor of at least one of a wearable device, a mobile processing device, a mobile communication device, and of a smart home system.

7. The method according to claim 1 , wherein

the predicted interaction score includes at least a binary prediction value or a confidence value for a predicted interaction to occur.

8. The method according to claim 1 , wherein,

in the training phase,

the acquired sensor data includes further sensor data on the at least one other person's user state and/or global information, and

in the application phase,

the acquired current sensor data includes further current sensor data on the at least one other person's user state and/or global information.

9. The method according to claim 1 , wherein the method is

performing the steps of acquiring sensor data and analyzing the acquired sensor data in the training phase offline, and the method further performs storing the acquired sensor data and the classifier data offline.

10. The method according to claim 1 , wherein the method is

performing the steps of acquiring sensor data and analyzing the acquired sensor data in the training phase online, and the method further performs

updating the classifier data based on at least the acquired current sensor data and current interaction data.

11. The method according to claim 10 , wherein the method is

performing the step of analyzing the acquired current sensor data to generate updated classifier data by unsupervised learning, in particular clustering based learning, or by an active learning method, in particular reinforcement learning.

12. The method according to claim 1 , wherein the method further comprises a step of

predicting a future user state of the user and/or the at least one other person, and wherein the method is

predicting the interaction score for the user based on the acquired current sensor data, the classifier data and further based on the predicted future state.

13. The method according to claim 1 , wherein the method further comprises at least one of steps of

providing, to the user, an interface presenting learned classifier data and accepting a user feedback on the presented learned classifier data via the interface,

providing, to the user, the interface for inputting further classifier data, and

providing, to the user, the interface for specifying the at least one other person for which the interaction identifier is to be generated based on the classifier data.

14. The method according to claim 1 , wherein the method further comprises a step of

outputting, to the user, an interaction recommendation based on the generated interaction identifier, or

initiating directly an interaction between the user and the at least one other person on the social network based on the generated interaction identifier.

15. The method according to claim 1 , wherein the method further comprises a step of

estimating a current user state based on the acquired current sensor data at a local processing device or at a remote processing device linked to the local processing device via a cloud service.

16. The method according to claim 15 , wherein

the classifier data is stored in a memory of the local processing device of the user, in a memory of a processing device of the at least one other person, or in a server memory of the remote processing device.

17. The method according to claim 16 , wherein

the classifier data differentiates different types of interactions in the social network,

wherein the different types of interactions include at least one of interaction duration, interaction content, interaction target, and a communication service.

18. The method according to claim 1 , wherein

the steps of predicting the interaction score and of generating the interaction identifier are performed locally at the user's processing device and provided to a processing device associated with the at least one other person for initiating the interaction, or

the steps of predicting the interaction score and of generating the interaction identifier are performed by a remote processing device and provided to the processing device associated with the user for initiating the interaction, or

the steps of predicting the interaction score and of generating the interaction identifier are performed locally at the user's processing device based on sensor data provided by the processing device associated with the at least one other person.

19. A non-transitory computer-readable recording medium having computer-readable program stored thereon which, when executed, cause a computer or digital signal processor to perform a method for improving an interaction process of a user in a social network with at least one other person, the method comprising steps of:

in a training phase,

acquiring sensor data on at least a user state of the at least one other person and interaction data on a social interaction of the user with the at least one other person, wherein the acquired sensor data concerns at least one of user activity and user emotion of the at least one other person,

analyzing the sensor data and the interaction data to generate classifier data from the acquired sensor data and the acquired interaction data,

wherein analyzing the sensor data and the interaction data comprises learning a correlation between the acquired sensor data and the acquired interaction data for generating the classifier data, and the generated classifier data comprises a learned classifier for the user and the at least one other person,

wherein the learned classifier determines an interaction initiation success for the user and the at least one other person, or the learned classifier determines an interaction initiation or an initiation acceptance success for the user and the at least one other person, and

in an application phase,

acquiring current sensor data on at least a current user state of the at least one other person, wherein the acquired current sensor data concerns at least one of user activity and user emotion of the at least one other person,

predicting an interaction score for the user based on the acquired current sensor data and the classifier data,

generating an interaction identifier for the at least one other person for an interaction with the at least one other person in the social network based on the predicted interaction score.

20. A system for improving an interaction process of a user in a social network with at least one other person, the system comprising:

a processor;

a memory;

a human interaction interface; and

a data acquisition interface for acquiring sensor data from at least one sensor;

wherein the data acquisition interface is configured to acquire the sensor data on at least a user state of the at least one other person and interaction data on a social interaction of the user with the at least one other person, wherein the acquired sensor data concerns at least one of user activity and user emotion of the at least one other person;

the processor is configured to analyze the acquired sensor data and the acquired interaction data to generate classifier data from the acquired sensor data and the acquired interaction data, and to store the classifier data in the memory, wherein analyzing the sensor data and the acquired interaction data comprises learning a correlation between the acquired sensor data and the acquired interaction data for generating the classifier data, and the generated classifier data comprises a learned classifier for the user and the at least one other person,

wherein the learned classifier determines an interaction initiation success for the user and the at least one other person, or the learned classifier determines an interaction initiation or an initiation acceptance success for the user and the at least one other person; and

the processor is further configured to acquire current sensor data on at least a current user state of the at least one other person, wherein the acquired current sensor data concerns at least one of user activity and user emotion of the at least one other person;

to predict an interaction score for the user based on the acquired current sensor data and the classifier data stored in the memory; and

to generate an interaction identifier for the at least one other person for an interaction with the at least one other person in the social network based on the predicted interaction score.