Table 1.
Characteristics of the conversational agents evaluated in the included studies
First author, year | Type of communication technology; type of conversational agenta | Dialogue managementl | Dialogue initiative | Input | Output | Task-oriented |
---|---|---|---|---|---|---|
Fitzpatrick et al. 201719 | Platform independent appb | Frame-based | Mixed | Written | Written | No |
Tanaka et al., 201720 | Windows computer app; ECAc | Finite-state | System | Spoken | Spoken, written, visual | Yes |
Miner et al., 201626 | Mobile device app | Agent-based | User | Spoken | Spoken, written | No |
Ireland et al., 201621 | Mobile device app; chatbotd | Frame-based | Mixed | Spoken | Spoken, written | No |
Rhee et al., 201428 | SMSe | Frame-based | Mixed | Written | Written | No |
Hudlicka, 201324 | Web browser app; ECAf | Frame-based | Mixed | Written | Written | No |
Crutzen et al., 201127 | Windows computer app; chatbotg | Frame-based | Mixed | Written | Written | No |
Philip et al., 201722 | Windows computer app; ECA | Finite-state | System | Spoken | Spoken | Yes |
Lucas et al., 201735 | Multimodal platform; ECAh | Finite-state | System | Spoken | Spoken | Yes |
Philip et al., 201423 | Windows computer app; ECA | Finite-state | System | Spoken | Spoken | Yes |
Beveridge and Fox, 200625 | Telephone and web browser appi | Frame-based | Mixed | Spoken | Spoken | Yes |
Black et al. 2005,29 Harper et al. 2008,30 Griol et al., 201331 | Telephonej | Finite-state | System | Spoken | Spoken | Yes |
Levin and Levin, 200632 | Telephonek | Finite-state | System | Spoken | Spoken | Yes |
Giorgino et al. 2005,33 Azzini et al. 200334 | Telephonei | Frame-based | Mixed | Spoken | Spoken | Yes |
Abbreviations: app: application; ECA: Embodied Conversational Agent; SMS, Short Message Service
Type of conversational agent considered unspecific, where not ECA nor chatbot;
Woebot, Woebot Labs: instant messenger app, platform independent;
Automated skills trainer developed from MMDAgent (http://www.mmdagent.jp);
Harlie the Chatbot (http://www.itee.uq.edu.au/cis/harlie);
mASMAA, an extension of TRIPS (The Rochester Interactive Panning System);
Virtual Mindfulness Coach;
Bzz Dutch chatbot for Windows Live Messenger;
SimSensei Virtual Agent, based on the MultiSense perception system, a multimodal sensing platform which fuses information from web cameras, the Microsoft Kinect and audio capture, and processing hardware (http://multicomp.ict.usc.edu/? p=1799);
HOMEY project – home monitoring through an intelligent dialogue system (http://www.openclinical.org/dm_homey.html#);
DI@L-log: although the system allows for dual tone multi frequency input this is rarely used, as all interactions can occur via spoken language;
Pain Monitoring Voice Diary, developed by Spacegate, Inc;
Not objectively reported in the paper, but inferred from descriptions of the CA, sample dialogues, or other published material on the system