Reducing graphical text analysis using physiological priors

• US 10,417,572 B2
• Filed: 06/18/2015
• Issued: 09/17/2019
• Est. Priority Date: 03/26/2015
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Cecchi, Guillermo A., Kozloski, James R., Pickover, Clifford A., Rish, Irina
• Primary Examiner(s): Rifkin, Ben M
• Application Number:
  US14/742,861
• Publication Number:
  US 20160283855A1
• Time in Patent Office: 1,552 Days
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks

Abstract

Embodiments relate to facilitating a meeting. A method for reducing an amount of communications to analyze in order to determine a cognitive state of an entity is provided. The method determines a first likelihood of an entity to have a particular cognitive state based on a set of physiological measures of the entity. The method receives communications from the entity. The method generates a graph of communications of the entity. The method performs a graphical text analysis on the graph to determine a second likelihood of the entity to have the particular cognitive state. The method determines whether the entity has the particular cognitive state based on the first likelihood and the second likelihood.

Claims

1. A computer-implemented method for a method of reducing an amount of communications to analyze in order to determine a cognitive state of an entity, comprising:

receiving, by a computer comprising one more processing circuits, a set of physiological measures of the entity that are taken by a wearable device;

determining, by the computer, a first likelihood of an entity to have a particular cognitive state based on the set of physiological measures of the entity that are received from the wearable device;

receiving, by the computer, electronic communications that are conducted by the entity, wherein the electronic communications includes audible communications comprising audio signals that is transmitted from the entity to at least one other entity;

transcribing, by the computer, the audible communications into text via an automated speech-to-text technique that converts the audio signals into the text;

generating, by the computer, a graph of the electronic communications of the entity, wherein generating the graph of the electronic communications of the entity includes;

extracting, by the computer, syntactic features from the text of the received electronic communications of the entity;

converting, by the computer, the extracted features into syntactic vectors;

generating, by the computer, semantic vectors from the text of the received electronic communications of the entity; and

generating, by the computer, the graph of the electronic communications of the entity based at least in part on the syntactic vectors and semantic vectors, wherein the generated graph of the electronic communications of the entity includes nodes representing tokens and edges representing temporal precedence in the electronic communications of the entity, wherein each node of the generated graph of the electronic communication of the entity comprises a feature vector that is generated based at least in part on a combination of a syntactic and semantic vector of the syntactic and semantic vectors;

in response to determining that the first likelihood of the entity having the particular cognitive state falls below a threshold confidence level, starting performing, by the computer, a graphical text analysis on the graph of the electronic communications of the entity to determine a second likelihood of the entity to have the particular cognitive state, wherein the graphical text analysis includes comparing the graph of the electronic communications of the entity with clusters of previously generated graphs stored in a clusters repository, wherein comparing the graph of the electronic communications of the entity with clusters of previously generated graphs stored in a clusters repository includes;

plotting, by the by the computer, in a multi-dimensional text feature space, feature vectors of previously generated graphs having known cognitive states;

forming, by the by the computer, clusters from the multi-dimensional text feature space;

plotting, by the by the computer, the feature vectors of the generated graph of the electronic communications of the entity in the multi-dimensional text feature space; and

comparing, by the computer, the plot of the feature vectors of the previously generated graphs having the known cognitive states with the plot of the feature vectors of the generated graph of the electronic communications of the entity;

determining, by the computer, whether the entity has the particular cognitive state based on the first likelihood and the second likelihood; and

in response to determining that the entity has the particular cognitive state;

stopping performing, by the computer, the graphical text analysis on the graph before the graphical text analysis is completed;

generating a notification that indicates the particular cognitive state and a confidence level of the entity having the particular cognitive state; and

transmitting the notification to at least a third party, wherein the third party comprise an a entity that did not participate in the electronic communications.

2. The method of claim 1, wherein the determining whether the entity has the particular cognitive state comprises:

combining the first likelihood and the second likelihood using Bayes'"'"' rule to determine a combined likelihood; and

determining that the entity has the particular cognitive state if the combined likelihood exceeds a threshold likelihood, wherein the confidence level of the entity having the particular cognitive state is based on the combined likelihood.

3. The method of claim 2, wherein the threshold likelihood is calibratable.

4. The method of claim 1, wherein the set of physiological measures of the entity comprises at least one of a galvanic skin response of the entity, an Electroencephalography (EEG) of the entity, a heart rate of the entity, a behavior analysis of video of the entity, and an observer assessment of the entity.

5. The method of claim 1, further comprising, in response to determining that the entity has the particular cognitive state, triggering an action to help the entity to transition to another cognitive state from the particular cognitive state.