Download Emotional Intelligence – Applications Based on Multi

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Agent (The Matrix) wikipedia , lookup

Agent-based model wikipedia , lookup

AI winter wikipedia , lookup

Embodied cognitive science wikipedia , lookup

History of artificial intelligence wikipedia , lookup

Ethics of artificial intelligence wikipedia , lookup

Intelligence explosion wikipedia , lookup

Philosophy of artificial intelligence wikipedia , lookup

Existential risk from artificial general intelligence wikipedia , lookup

Affective computing wikipedia , lookup

Transcript
Emotional Intelligence and Agents –
Survey and Possible Applications
Mirjana Ivanovic, Milos Radovanovic, Zoran
Budimac, Dejan Mitrovic, Vladimir Kurbalija,
Weihui Dai and Weidong Zhao
1University
of Novi Sad, Serbia
2Fudan University, Shanghai, China
Agenda
 Introduction
 Emotional Intelligence and its Influence on Intelligent
Systems
 Multi-agent Intelligent System for Emotion Recognition
 Conclusions
Emotional Intelligence and Agents – Survey and Possible Applications - WIMS'14, June 2-4, 2014, Thessaloniki, Greece
2/14
Introduction
 Emotional Intelligence is a new discipline of knowledge,
dealing with modeling, recognition and control of human
emotions.
 It is an interdisciplinary research related to machine
learning, natural language processing, psychology…
 Cognitive Model and Propagation Model
Emotional Intelligence and Agents – Survey and Possible Applications - WIMS'14, June 2-4, 2014, Thessaloniki, Greece
3/14
Introduction
 Human emotion
- caused by specific situations,
- emotional change can trigger a series of physiological responses,
- form the unique subjective experience expressed like gesture, action,
language …
 Objective of cognitive model
- reason about emotions,
- predict and understand human emotions,
- process emotions and give appropriate response.
 EI provides the key theory and technical basis
- new generation sensor networks,
- intelligent information processing techniques,
- man-machine conversation technologies,
- machine intelligence services,
- …
Emotional Intelligence and Agents – Survey and Possible Applications - WIMS'14, June 2-4, 2014, Thessaloniki, Greece
4/14
Introduction
 Multi-agent systems (MAS) - Highly connected to the
application of EI in real domains.
- Development of logical frameworks for the formal
specification of emotions.
- Exploit logical methods in order to provide a rigorous
specification of how emotions should be implemented
in an artificial agent.
 Recent work:
- Enabling intelligent software agents to detect emotions
via verbal, non-verbal, and textual cues,
- Express emotions through speech and gestures.
Emotional Intelligence and Agents – Survey and Possible Applications - WIMS'14, June 2-4, 2014, Thessaloniki, Greece
5/14
Agenda
 Introduction
 Emotional Intelligence and its Influence on Intelligent
Systems
 Multi-agent Intelligent System for Emotion Recognition
 Conclusions
Emotional Intelligence and Agents – Survey and Possible Applications - WIMS'14, June 2-4, 2014, Thessaloniki, Greece
6/14
Emotional Intelligence and its Influence on IS
1. Usability of EI: Necessary Research Activities
 Today’s users are not satisfied with simple visual and audio
responses of interactive systems.
 Want to employ all their sensory organs in order to receive the
details of a multimedia presentation: require tactile, olfactory,
and other modes of communications.
 Trend: to build interactive systems that are more attractive and
closer to end-users, to be believable and trustable.
Emotional Intelligence and Agents – Survey and Possible Applications - WIMS'14, June 2-4, 2014, Thessaloniki, Greece
7/14
Emotional Intelligence and its Influence on IS
1. Usability of Emotional Intelligence: Necessary Research Activities
 Research challenges and activities:
- Analysis and evaluation of different emotional classification methods:
speech, expression, gesture, action, and electrophysiology information
- Improve the stability and accuracy of emotional information pattern
recognition: use and combine different recognition algorithms,
- Exploiting logical methods in order to provide a rigorous specification of
how emotions should be implemented in an artificial agent.
-
Identifying desirable features of emotion theories that make them ideal
blueprints for agent models.
- A natural consequence: build a prototype of a multiagent system capable of
reasoning about emotions, predicting and understanding human emotions,
and processing emotions by reasoning and during the interaction with a
human user.
Emotional Intelligence and Agents – Survey and Possible Applications - WIMS'14, June 2-4, 2014, Thessaloniki, Greece
8/14
Emotional Intelligence and its Influence on IS
2. Emotional Intelligence and Agent Technology
 Apply specific methods for dynamic data analysis and pattern
recognition, to identify /discover new knowledge from available
emotional information.
 The main focus on specific activities:
- Choosing the appropriate analysis methods for large amounts of
dynamic and temporal data (collected from simulations and experiments)
which can be used in explanations and modeling of human emotions.
- To explore possibilities of modeling multi-agent systems which
incorporate emotional intelligence.
- To build a logic that enables specification of complex emotions (e.g.
regret, jealousy, envy, shame, guilt, reproach, admiration, remorse, pride,
embarrassment).
- To perform experiments on numerous appropriate data sets (system
based on FAP – Framework for Analysis and Prediction)
Emotional Intelligence and Agents – Survey and Possible Applications - WIMS'14, June 2-4, 2014, Thessaloniki, Greece
9/14
Agenda
 Introduction
 Emotional Intelligence and its Influence on Intelligent
Systems
 Multi-agent Intelligent System for Emotion Recognition
 Conclusions
Emotional Intelligence and Agents – Survey and Possible Applications - WIMS'14, June 2-4, 2014, Thessaloniki, Greece
10/14
Multi-agent Intelligent System for Emotion Recognition
Concrete research topics - 1. Emotion Detection in Human
Speech
 The voice characteristics and emotion expression of human speech are
detectable in sound data (various aspects of telephone services).
 To improve customer experience, two major problems are faced: customer
emotions are delayed and hard to detect; and the lack of an effective
problem-solving strategy for different cases of customer emotional response.
- The first phase of our research is acquisition and preparation of data in
Chinese and English, and development of a prototype system that can
detect emotional content and intensity in recorded voice data.
- The second problem is based on application of techniques involving
emotional agents: simulation of various scenarios of voice interaction.
Emotional Intelligence and Agents – Survey and Possible Applications - WIMS'14, June 2-4, 2014, Thessaloniki, Greece
11/14
Multi-agent Intelligent System for Emotion Recognition
Concrete research topics - 2. Emotion Detection in Online
Documents
 Agent technology will be employed to perform distributed
emotion detection in online documents: blogs, discussion
forums, social networks, news that report on emergency
events, etc.
 Two types of agents will be developed: harvester and analyst.
 Harvester agents will scan the web in search of documents
that meet the given criteria.
 Analyst agents will perform text analysis in a distributed
setting.
Emotional Intelligence and Agents – Survey and Possible Applications - WIMS'14, June 2-4, 2014, Thessaloniki, Greece
12/14
Agenda
 Introduction
 Emotional Intelligence and its Influence on Intelligent
Systems
 Multi-agent Intelligent System for Emotion Recognition
 Conclusions
Emotional Intelligence and Agents – Survey and Possible Applications - WIMS'14, June 2-4, 2014, Thessaloniki, Greece
13/14
Conclusions
 Emotion modeling is expected to have an interesting and
important role in the next generation man-machine
interactive systems.
 It can be realized by modeling both input and output
parameters of the interactive system.
- Visual expression of the input-user-interface (laughter,
cry, other forms of facial manifestation) can be regarded
as input to the computers.
- Emotional expression of a computer can be realized by
synthesizing emotions either on the video display unit or
on an artificial emotional creature.
Emotional Intelligence and Agents – Survey and Possible Applications - WIMS'14, June 2-4, 2014, Thessaloniki, Greece
14/14