Technology, Summary of Using non-invasive wearables for detecting emotions with intelligent agents.<\/p>\n
Using non-invasive wearables for detecting emotions with intelligent agents
\n2<\/p>\n
Abstract. This paper proposes the use of intelligent wristbands for the auto- matic detection of emotional states in order to develop an application which allows to extract, analyze, represent and manage the social emotion of a group of entities. Nowadays, the detection of the joined emotion of an heterogeneous group of people is still an open issue. Most of the existing approaches are cen- tered in the emotion detection and management of a single entity. Concretely, the application tries to detect how music can influence in a positive or neg- ative way over individuals\u2019 emotional states. The main goal of the proposed system is to play music that encourages the increase of happiness of the overall patrons.
\n1 Introduction
\nOver the last few years, research on computational intelligence is being conducted in order to emulate and\/or detect emotional states [1]. The emulation of emotional states allow machines to represent some human emotions. This artificial representa- tion of emotions is being used by machines to improve the interaction process with humans. In order to create a fluid emotional communication between human and machines, the machines need first to detect the emotion of the human with the final purpose of improving human-computer interactions [2]. To do this it is necessary to use different techniques such as: artificial vision [3], speech recognition [4], body gestures [5], written text [6] and biosignals [7].
\nHuman beings perceive and analyse a wide range of stimuli in different envi- ronments. These stimuli interfere in our commodity levels modifying our emotional states. Before each one of these stimuli, humans generate several type of responses, like varying our face gestures, body movement or bio-electrical impulses. These vari- ations in our emotional states could be used as a very useful information for ma- chines. To do this, machines will require the capability of interpreting correctly such variations. This is the reason for the design of emotional models that interpret and represent the different emotions in a computational way. In this case, emotional models such as Ortony, Clore & Collins model [8] and the PAD (Pleasure-Arousal- Dominance) model [9] are the most used ones to detect or simulate emotional states. Moreover, emotional states are a very valuable information, allowing to develop ap- plications that help to improve the human being quality of life.
\nNowadays, the detection of the joined emotion of an heterogeneous group of people is still an open issue. Most of the existing approaches are centered in the emotion detection and management of a single entity. In this work we propose to detect the social emotion of a group of people in an Ambient Intelligence (AmI) application with the help of wearables. Specifically, we show a system that controls automatically the music which is playing in a bar through the detection of the emo- tions of the patrons with the use of individual wristbands. Thus, the main goal of the proposed system is to play music that encourages the increase of happiness of the overall patrons. Each one of the individuals will have an emotional response according to his musical taste. This response will be detected and transmitted by the wristbands in order to calculate a social emotion of the set of individuals. This social emotion will be used to predict the most appropriated songs to be played in the bar.
\n2 State of the Art
\nThe AmI area is rapidly gaining notoriety due to its usage on complex social envi- ronments like nursing homes and regular homes. By monitoring fragile users (like elderly or mentally challenged people) the available systems pose as an alternative to regular caregiving services while being cost-effective. Despite the several aims AmI projects have they can be clustered in five clusters of operational areas [10]:
\n\u2013 Daily living activities
\n\u2013 Fall and movement detection \u2013 Location tracking
\n\u2013 Medication control
\n\u2013 Medical status monitoring
\nIn terms of daily living activities there is the project Caregiver\u2019s Assistant, which uses RFID and a database with human activities events and a fast inference mechanism that allows the identifications of the actions within a given space [11]. It works by registering the RFID cards that the users carry, which in some cases they have to actively pass them through the readers due to they small communication range. Thus, it is very intrusive to the users of the system as they have to be actively aware of the procedures so that the system is able to correctly access the information.
\nIn terms of fall detection or movement detection, most of the operation methods resort to use cameras to register the visual information and extract information from it, like the projects in [12,13,14]. Although they require no interaction with the users, these systems are very invasive, not to mention the possible loss of privacy, due to the permanent recording of the environment.
\nThe location tracking systems like the one presented on [15] use mobile de- vices sensors to provide the current location of the user to an AmI system, more specifically to the caregivers. These systems require the constant monitoring of the localization, thus there is no guarantee of privacy, thus being very intrusive systems. This intrusion is not done directly but by allowing 3rd party users to constantly know the location of another person the system becomes very intrusive [16].
\nMedication control projects consist in systems that help the users to remind the medications that they have to take [17,18]. They play an important role on the users life, as most of the AmI and AAL projects users have some sort of cognitive disability and have trouble in remembering to do activities, such as taking medication. These systems are mostly recommenders and are able to only provide information without being disruptive or actively monitoring the users. Due to the simplicity of the projects premises it is available a large number of simple applications for mobile devices and desktop computers currently.
\nThe medical status monitoring projects like the ones presented in [19,20,21,22] show platforms that are constituted by sensor systems that are directly in contact with the human body. These sensor systems create a body area network and pro- vide information about the carrier vital signs and bodily functions. For instance, all three works presented capture electrocardiograms and reason that information to obtain knowledge about their specific domain. The [21] uses the electrocardiogram information and ballistocardiogram information to assert if the drivers are calm and concentrated or if they are stressed or having some kind of medical issue (as the project is directed to elderly people). The project ALADIN [22] presents a system that manages home lights (brightness and colour) according to the users physical state. The users carry a biosignal reading glove (that captures photoplethysmogra- phy) that sends the readouts to the server and according to their pulse\/heart-rate the lights are dimmed or changed their colour. The aim of this project is to provide comfort and promote a peaceful living, adapting the environment to the user state or preferences. The glove has to be put by the users and limits its use to a confined range of actions (it cannot be wetted or be used to manage heavy objects) thus being quite invasive, possibly undermining the results as the users become actively aware of their status, thus allowing them to manipulate the system. The issue with these systems is that they require users to attach sensors on their own body (the case of [19,20]) or that the users are in a very controlled environment like the [21].
\nThese projects are a small representation of the plethora of the existent projects and show the current lines of development. One common problem revealed of these projects is that they are interested in the implementation and the execution of their components but do not reveal any particular interest on the opinion of the users towards the devices that they are using. Only recently the theme of invasiveness has been considered due to the high reluctance of the users towards clumsy and complex apparatus [23].
\nIn the line of ideas presented by [22], we are aiming to produce a system that increases the comfort level of the users (by managing the current music) through the use of a non-invasive wearable bracelet that performs medical status monitoring to attain the users\u2019 emotional status.
\n3 Problem description
\nThis application example is based on how music can influence in a positive or neg- ative way over emotional states [24], [25], [26]. The application example is devel- oped in a bar, where there is a DJ agent in charge of playing music and a specific
\nnumber of individuals listening to the music. The main goal of the DJ is to play mu- sic making that all individuals within the bar are mostly as happy as possible. To get this, it is necessary to detect the human emotions, and there exist many techniques to do it. But in our case, we decided to use the bio-signals to detect the emotional change. This emotional change is used by the DJ agent to change the musical genre and try that the clients of the bar are mostly as happy as possible. Each one of the individuals will have an emotional response according to its musical taste. This re- sponse is reflected in a variation of the bio-signals [27][28]. This variation allows us to calculate the social emotion [29] of the people within the bar. Based on the metric of the social emotion, the DJ agent could change the music genre to move the social emotion to a target emotion. To capture the bio-signals we designed a prototype of an Emotional Smart Wristband (Figure 1), that will be explained with more detail in Section 4.
\nFig. 1: Prototype of an Emotional Smart Wristband.
\n4 System Proposal
\nThis section explains the different components that constitute the multi-agent sys- tem which describes a way to detect emotions based on bio-signals through wearable devices. The main problem in the detection of human emotions is the information capture. This information is normally obtained using image processing, text analysis or voice analysis. These ways are invasive and in some of them is necessary to have the consent of the person. In currently years the use of wearable devices has been growing, devices such as Samsung3 with the Gear Fit, Gear S2 or Apple4 with the Apple Watch are only some examples. These devices can measure heart rate beat or hand movement using the IMU (Inertial Measurement Unit). Based on these devices and using the currently technology in embedded systems, it is possible to create new smart bracelets which include other type of measures, such as the EEG, GSR, Photo- plethysmogram or ECG allowing the acquisition of biosignals that can help for the detection of the human\u2019s emotions. Using signals of this kind along with the incor- poration of complex algorithms based on machine learning techniques, it is possible to recognise how humans change their emotional states.
\n3 http:\/\/www.samsung.com 4 http:\/\/www.apple.com<\/p>\n
The proposed multi-agent system is formed by three types of agents. These agents are: the Wristband agent, the Social Emotion Agent, and the DJ agent. The Wristband agent is mainly in charge of: (i) capture some emotional information from the envi- ronment and specifically from a specific individual, this is done by interacting with the real world through the employed wristband. The agent captures the different bio-signals, that will be used to detect the emotion of a human being; and (ii) pre- dict the emotional state of the individual from the processed biosignals. In order to analyze these changes and predict emotional states, the Wristband agent employs a classifier algorithm that will be later explained. Once the emotion has been ob- tained, it is sent to the agent which is in charge of calculating the social emotion of the agent group. This agent is called Social Emotion Agent or SEtA. The main goal of this agent is to receive the calculated emotions from all the Wristband agents and, us- ing this information, generate a social emotional state for the agent\u2019s group (details of how this social emotion is calculated can be seen in [29]). Once this social emo- tion is obtained, the SEtA can calculate the distance between the social emotion and a possible target emotion (in this case the target emotion is happiness). This allows to know how far is the agent\u2019s group of the target emotion. This can be used by the system to try to reduce that distance modifying the environment. This modification of the environment is made by the DJ agent. This agent uses this social emotional value to calculate what is the next song to be played. After different executions, the DJ agent can evaluate the effect that the song has had over the audience. This will help the DJ to decide whether to continue with the same musical genre or not in order to improve the emotional state of the group of people.
\nDue to the limits of the paper, we only describe in detail the processes made by the Wristband agent which are the data acquisition process and the emotion recog- nition. Moreover, the physical components of the wristband prototype are also de- scribed.
\n4.1 Data Acquisition Process
\nThis process made by the Wristband agent is responsible to capture the different needed bio-signals. To do this, the Wristband agent uses different sensors. The sen- sors used are: GSR and Photoplethysmogram (Figure2). The GSR measures the gal- vanic skin response. The measurement is performed by passing through the skin a very low current, and storing small variations in voltage. On the other hand, the Photoplethysmogram is a process of applying a light source measuring the light re- flected by the skin. The received signal consists of pulses that reflect the change in vascular blood volume with each cardiac beat. The information captured by each one of these sensors is subsequently preprocessed. This last process allows to convert the measure captured for each sensor in the corresponding units. The GSR sensor converts the measurement of the skin conductance in Ohm and the Photoplethys- mogram returns raw data that can be easily processed.
\n(a) Heart Rate Sensor (b) Skin Resistance Sensor Fig. 2: View of the employed sensors.
\n4.2 Emotion Recognition
\nOnce the data has been obtained, it is necessary to implement a machine learning algorithm in order to identify the human emotions. To do this, the process has been divided into two subprocesses. The first one employs a Fuzzy logic algorithm in order to obtain the qualitative value (the name of the emotion) of the emotions stored in the employed dataset (which is below explained). The second process employs a Neural Network in order to classify new bio-signals inputs from the wristband into emotional values.
\nThe dataset used to train and validate this model is the DEAPdataset [7]. The dataset contains physiological signals of 32 participants (and frontal face video of 22 participants), where each participant watched and rated their emotional response to 40 music videos along the scales of arousal, valence, and dominance, as well as their liking of and familiarity with the videos. This dataset integrates different bio- signals as: EEG, GSR, EOG, among other signals. All these signals are associated to the emotional changes using musical videos. Specifically, the authors identified 16 different emotions, which are the following:
\n1. Pride
\n2. Elation
\n3. Joy
\n4. Satisfaction 5. Relief
\n6. Hope
\n7. Interest
\n8. Surprise
\n9. Sadness 10. Fear
\n11. Shame 12. Guilt
\n13. Envy
\n14. Disgust 15. Contempt 16. Anger
\nThese emotions are represented following the circumflex emotional model [30][31][32]. This emotional model represents the emotions using three components Valence, Arousal, Dominance. In addition, the model evaluates every emotion in a trivalent
\nscale: light, medium, and strong. As a result, the model has 36 possible emotional
\nstates.
\nFig. 3: Circle of emotions (arousal and valence).[31]
\nThis model locates the emotions in twelve sub-quadrants, where each sub-quadrant
\nis discretized in ranges of 30 degrees. The intensity of the emotion is the module of
\n#\u00bb
\nthe vector composed by E (Ag) = [Arousal, Valence]. The representation of emo- tions is done using a polar coordinate plane, where one takes into account the angle and the magnitude of the vector (see Equation 1 and Equation 2).
\nr = pArousal2 + +Valencd2 (1) 8<arctan(valence) if=”” arousal=””>0
\n\u2713= \u21e1 Arousal if Arousal=0 (2) : 2 Valence
\narctan(Arousal)+\u21e1if Arousal<0
\nTherefore, the emotion is represented as a tuple composed by the radius (r) and the angle (\u2713) E(Ag) = {r,\u2713} (all angles are in radians). Based on these data we em- ploy a set of fuzzy logic rules in order to estimate the name of the emotion according to the input values stored in the dataset. These rules allow us to change a quantita- tive response to a qualitative response. This qualitative response is calculated and stored in the database for all the available registers. Once this has been calculated, the next step is the definition of a neural network, which allow us to identify the human emotions using only the data obtained from the GSR and Photoplethysmo- gram.
\nIt is necessary to remark that, as each channel of the GSR and the Photoplethys- mogram is formed by 8064 different values, it is impossible to build a neural network with these number of inputs. For this reason each channel was sub-sampled, convert- ing each channel in an array of 252 values. Therefore, the neural network has 504 inputs (252 per each channel). The network has also has five hidden neurons and sixteen outputs (each output corresponds with a specific emotion). The architecture of our neural network is shown in Figure 4.
\nFig. 4: Neural Network Architecture.
\nThe ANN was trained using a supervised trained methodology, since the objective of the network is to classify the human emotion. Concretely, the training process employed a dataset composed by 1280 entries per channel. As before commented, this information was extracted from the DEAPdataset. From the selected dataset, the 20% was used to test and the 80% was used to train.
\n4.3 Wristband Prototype
\nThis section describes the design of the physical wristband where the Wristband agent is executed. The wristband device was programmed in Python and it was em- bedded in the Intel Edison 5 computer-on-module. The Intel Edison (Figure 5) is a new technology designed by Intel which contains a Dual Core IA-32 @ 500 MHz, a 32-bit CPU specially designed for Internet of Things (IoT) applications and wearable computing products.
\n5 http:\/\/www.intel.la\/content\/www\/xl\/es\/do-it-yourself\/edison.html<\/p>\n
Fig. 5: Intel Edison Processor.
\nThe Intel Edison supports Yocto Linux 6 and incorporates the SPADE7 platform [33], which is a multi-agent system platform based on Python. As before commented, the prototype has been designed as a wristband in which is deployed the wristband agent. The bio-signals captured in the wristband are passed by an Analog to Digi- tal Conversion or ADC allowing the discretization of the analogue signals. Figure 6 shows the different components of our wristband prototype which are the following.
\nFig. 6: Components of the Smart Wristband Prototype.
\n1. Display and Control Module: This module is employed to show information of our emotional state through the LCD screen, and also controls the wristband.
\n6 https:\/\/www.yoctoproject.org\/
\n7 https:\/\/github.com\/javipalanca\/spade<\/p>\n
2. Power Supply and Battery Charger: This is the power supply system. The wrist- band uses a 3.7 volt battery.
\n3. Sensors: Sensors are responsible for carrying out the acquisition of the signals of GSR and Photoplethysmogram.
\n4. AnaloguetoDigitalConvert:Thismoduleisresponsiblefordigitizingthesignals captured by the sensors.
\n5. Intel Edison Processor: This is the microprocessor where the agent is located. It is responsible for performing the processes of emotion recognition and commu- nication with other agents (through the built-in Wi-Fi).
\n5 Conclusions and future work
\nThis paper presents how to integrate non-invasive biosignals for the detection of human emotional states through an agent-based application. The identification and detection of human emotional states allow the enhancement of the decision-making process of intelligent agents. The proposed application allows extracting (in a non- invasive way) the social emotion of a group of persons by means of wearables facili- tating the decision-making in order to change the emotional state of the individuals. As commented before, the application incorporates automatic emotion recognition using biosignals and machine learning techniques, which are easily included in the proposed system. The flexibility and dynamism of the proposed application allow the integration of new sensors or signals in future stages of the project. Moreover, as future work, we want to apply this system to other application domains, specifi- cally the proposed framework fits with the industrial one, for instance representing production lines including the individuals and their emotional states as yet another elements to be considered in the production line.
\n6 Acknowledgements
\nThis work is partially supported by the MINECO\/FEDER TIN2015-65515-C4-1-R and the FPI grant AP2013-01276 awarded to Jaime-Andres Rincon. This work is supported by COMPETE: POCI-01-0145-FEDER-007043 and FCT \u2013 Fundac\u0327a\u0303o para a Cie\u0302ncia e Tecnologia within the projects UID\/CEC\/00319\/2013 and Post-Doc schol- arship SFRH\/BPD\/102696\/2014 (A. Costa)
\nReferences
\n1. Jonathan Gratch and Stacy Marsella. Tears and fears: Modeling emotions and emotional behaviors in synthetic agents. In Proceedings of the fifth international conference on Au- tonomous agents, pages 278\u2013285. ACM, 2001.
\n2. C Maaoui and a Pruski. Emotion recognition through physiological signals for human- machine communication. Cutting Edge Robotics, pages 317\u2013333, 2010.
\n3. M Karthigayan, M Rizon, R Nagarajan, and Sazali Yaacob. Genetic algorithm and neural network for face emotion recognition. Affective Computing, pages 57\u201368, 2008.
\n4. Kun Han, Dong Yu, and Ivan Tashev. Speech Emotion Recognition Using Deep Neu- ral Network and Extreme Learning Machine. Fifteenth Annual Conference of …, (September):223\u2013227, 2014.
\n5. Stefano Piana, Francesca Odone, Alessandro Verri, and Antonio Camurri. Real-time Au- tomatic Emotion Recognition from Body Gestures. arXiv preprint arXiv:1402.5047, pages 1\u20137, 2014.
\n6. Z J Chuang and Chung-hsien Wu. Multi-modal emotion recognition from speech and text. Journal of Computational Linguistics and Chinese, 9(2):45\u201362, 2004.
\n7. SanderKoelstra,ChristianMu\u0308hl,MohammadSoleymani,JongSeokLee,AshkanYazdani, Touradj Ebrahimi, Thierry Pun, Anton Nijholt, and Ioannis Patras. DEAP: A database for emotion analysis; Using physiological signals. IEEE Transactions on Affective Computing, 3(1):18\u201331, 2012.
\n8. B. N. Colby, Andrew Ortony, Gerald L. Clore, and Allan Collins. The Cognitive Structure of Emotions, volume 18. Cambridge University Press, 1989.
\n9. a Mehrabian. Analysis of affiliation-related traits in terms of the PAD Temperament Model. The Journal of psychology, 131(1):101\u2013117, 1997.
\n10. HandeAlemdarandCemErsoy.Wirelesssensornetworksforhealthcare:Asurvey.Com- puter Networks, 54(15):2688\u20132710, 2010.
\n11. Kenneth P. Fishkin, Bing Jiang, Matthai Philipose, and Sumit Roy. I sense a disturbance in the force: Unobtrusive detection of interactions with RFID-tagged objects. In UbiComp 2004: Ubiquitous Computing, pages 268\u2013282. Springer Berlin Heidelberg, 2004.
\n12. Jose\u0301 Carlos Castillo, Juan Serrano-Cuerda, Antonio Ferna\u0301ndez-Caballero, and Arturo Marti\u0301nez-Rodrigo. Hierarchical architecture for robust people detection by fusion of in- frared and visible video. In Intelligent Distributed Computing IX, pages 343\u2013351. Springer Berlin Heidelberg, 2015.
\n13. Jose\u0301 Carlos Castillo, Antonio Ferna\u0301ndez-Caballero, A\u0301lvaro Castro-Gonza\u0301lez, Miguel A. Salichs, and Mari\u0301a T. Lo\u0301pez. A framework for recognizing and regulating emotions in the elderly. In Ambient Assisted Living and Daily Activities, pages 320\u2013327. Springer Berlin Heidelberg, 2014.
\n14. Chung-HsienKuo,Chun-TzuChen,Ting-ShuoChen,andYu-ChenKuo.Awirelesssensor network approach for rehabilitation data collections. In 2011 IEEE International Con- ference on Systems, Man, and Cybernetics. Institute of Electrical & Electronics Engineers (IEEE), 2011.
\n15. Joa\u0303oRamos,TiagoOliveira,KenSatoh,Jose\u0301Neves,andPauloNovais.Orientationsystem based on speculative computation and trajectory mining. In Highlights of Practical Appli- cations of Scalable Multi-Agent Systems. The PAAMS Collection, pages 250\u2013261. Springer Nature, 2016.
\n16. PaulHert,SergeGutwirth,AnnaMoscibroda,DavidWright,andGloriaGonza\u0301lezFuster. Legal safeguards for privacy and data protection in ambient intelligence. Personal and Ubiquitous Computing, 13(6):435\u2013444, 2008.
\n17. Katarzyna Stawarz, Anna L. Cox, and Ann Blandford. Don\u2019t forget your pill! In Proceed- ings of the 32nd annual ACM conference on Human factors in computing systems – CHI\u201914. Association for Computing Machinery (ACM), 2014.
\n18. NancyTran,JanetM.Coffman,KaharuSumino,andMichaelD.Cabana.Patientreminder systems and asthma medication adherence: a systematic review. Journal of Asthma, 51(5):536\u2013543, 2014.
\n19. Mirza Mansoor Baig, Hamid GholamHosseini, Martin J. Connolly, and Ghodsi Kashfi. Real-time vital signs monitoring and interpretation system for early detection of multiple physical signs in older adults. In IEEE-EMBS International Conference on Biomedical and Health Informatics. IEEE, 2014.
\n20. Anna Hristoskova, Vangelis Sakkalis, Giorgos Zacharioudakis, Manolis Tsiknakis, and Filip De Turck. Ontology-driven monitoring of patient\u2019s vital signs enabling personal- ized medical detection and alert. Sensors, 14(1):1598\u20131628, 2014.
\n21. Marian Walter, Benjamin Eilebrecht, Tobias Wartzek, and Steffen Leonhardt. The smart car seat: personalized monitoring of vital signs in automotive applications. Pers Ubiquit Comput, 15(7):707\u2013715, 2011.
\n22. Edith Maier and Guido Kempter. ALADIN – a magic lamp for the elderly? In Handbook of Ambient Intelligence and Smart Environments, pages 1201\u20131227. Springer Berlin Hei- delberg, 2010.
\n23. Rune Fensli, P. E. Pedersen, T. Gundersen, and O. Hejlesen. Sensor acceptance model \u2013 measuring patient acceptance of wearable sensors. Methods Inf Med, 2008.
\n24. BrianWhitmanandPSmaragdis.CombiningMusicalandCulturalFeaturesforIntelligent Style Detection. In Ismir, pages 5\u201310, Paris, France, 2002.
\n25. M. D. van der Zwaag, J. H. D. M. Westerink, and E. L. van den Broek. Emotional and psychophysiological responses to tempo, mode, and percussiveness. Musicae Scientiae, 15(2):250\u2013269, 2011.
\n26. Klaus R Scherer and Marcel R Zentner. Emotional Effects of Music : Production Rules. Music and emotion: Theory and research, pages 361\u2013392, 2001.
\n27. Mari\u0301a Viqueira Villarejo, Begon\u0303a Garci\u0301a Zapirain, and Amaia Me\u0301ndez Zorrilla. A stress sensor based on galvanic skin response (GSR) controlled by ZigBee. Sensors (Switzer- land), 12(5):6075\u20136101, 2012.
\n28. Eduardo Coutinho and Angelo Cangelosi. Musical emotions: Predicting second-by- second subjective feelings of emotion from low-level psychoacoustic features and physi- ological measurements. Emotion (Washington, D.C.), 11(4):921\u2013937, 2011.
\n29. J.A. Rincon, V. Julian, and C. Carrascosa. Social Emotional Model. In 13th International Conference on Practical Applications of Agents and Multi-Agent Systems, volume 9086 of LNAI, pages 199\u2013210, 2015.
\n30. Michelle Yik, James A Russell, and James H Steiger. A 12-point circumplex structure of core affect. Emotion, 11(4):705\u2013731, 2011.
\n31. R.E.Thayer.Thebiopsychologyofmoodandarousal.OxfordUniversityPress,USA,1989.
\n32. Jose L. Salmeron. Fuzzy cognitive maps for artificial emotions forecasting. Applied Soft
\nComputing Journal, 12(12):3704\u20133710, 2012.
\n33. M.Escriva,J.Palanca,G.Aranda,A.Garci\u0301a-Fornes,V.Julian,andV.Botti.AJabber-based
\nMulti-Agent System Platform. In Proceedings of the Fifth International Joint Conference on Autonomous Agents and Multiagent Systems (AAMAS06), pages 1282\u20131284. Association for Computing Machinery, Inc. (ACM Press), 2006.<\/p>\n
1) Read the paper carefully. You may need to read it more than once.
\n2) Concentrate on the sections discussed in the classroom, such the abstract, introduction, literature review and discussion.
\nSummarize the paper by describing the important points and given examples. Then, use your critical eye to analyze the study, in terms of the study importance, completeness, richness (level of details), timeliness, etc. This part (to be around 1500 words)
\n– use simple english<\/arctan(valence)><\/p>\n","protected":false},"excerpt":{"rendered":"
Technology, Summary of Using non-invasive wearables for detecting emotions with intelligent agents. Using non-invasive wearables for detecting emotions with intelligent agents 2 Abstract. This paper proposes the use of intelligent wristbands for the auto- matic detection of emotional states in Read More …<\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-426400","post","type-post","status-publish","format-standard","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/www.benedictsol.com\/blogs\/wp-json\/wp\/v2\/posts\/426400","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.benedictsol.com\/blogs\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.benedictsol.com\/blogs\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.benedictsol.com\/blogs\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.benedictsol.com\/blogs\/wp-json\/wp\/v2\/comments?post=426400"}],"version-history":[{"count":0,"href":"https:\/\/www.benedictsol.com\/blogs\/wp-json\/wp\/v2\/posts\/426400\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.benedictsol.com\/blogs\/wp-json\/wp\/v2\/media?parent=426400"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.benedictsol.com\/blogs\/wp-json\/wp\/v2\/categories?post=426400"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.benedictsol.com\/blogs\/wp-json\/wp\/v2\/tags?post=426400"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}