A study of trust development in virtual human-agent teamwork without explicit coordination
Başlık çevirisi mevcut değil.
- Tez No: 402942
- Danışmanlar: DR. SANDIP SEN
- Tez Türü: Doktora
- Konular: Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrol, Computer Engineering and Computer Science and Control
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 2015
- Dil: İngilizce
- Üniversite: The Unıversıty Of Tulsa
- Enstitü: Yurtdışı Enstitü
- Ana Bilim Dalı: Belirtilmemiş.
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: Belirtilmemiş.
Özet
Özet yok.
Özet (Çeviri)
With the advancement in agent technology and social recognition and acceptance of autonomous online services we foresee increasing use of agents in social contexts and, in particular, in human-agent virtual and ad hoc team applications. For such virtual ad hoc teams, particularly those where explicit coordination to determine task allocations are infeasible due to communication and time constraints, to be e ective, agents need to quickly develop an understanding of the expectation of human team members and be able to quickly adapt accordingly. This dissertation empirically investigates the e ects of agent performance, past experience of humans, and agent reputation on human trust in and behavior towards agent teammates. We also study the di erences in growth of humans' trust attitudes towards human and agent teammates during initial interactions while achieving team goals. Ad hoc collaboration of humans and agents without pre-planning is a novel humanagent teamwork concept that enables diverse, high-impact applications including emergency response, group buying, online social movements, and next generation crowd-sourcing. In spite of prior work on the in uence of various factors on human trust in automation, very little is known about human trust in \peer" agent teammates. This is the rst study to investigate these factors in virtual human-agent teamwork without pre-planning or explicit coordination. This dissertation introduces a repeated team coordination game, the Game of Trust (GoT), in which two players repeatedly cooperate to complete team tasks without prior assignment of subtasks. We systematically evaluate the e ects of agent performance, past experience, and agent reputation on human trust in the GoT by performing an extensive set of controlled experiments with subjects recruited from Amazon Mechanical Turk, a crowdsourcing Internet marketplace. We collect both teamwork performance data as well as surveys to gauge participants' trust in their agent partners. A learning agent player is developed by training on collected game data to predict future task choices of humans and is shown to increase social utility by reducing redundant work and increasing team goal achievement rates. The empirical results indicate that participants were able to respond appropriately to changes in an agent's behavior in terms of trust and contributions to teamwork, that positive (negative) past experience and positive (negative) reputation increase (decrease) human trust in agent teammates, and humans' trust attitudes towards human and agent teammates di er. The results also show that past experience can a ect three antecedents of trust: emotional state, game expertise, and expectation. The interplay between these three elements signi cantly a ected the in uence of agent performance and reputation on human trust. The detailed ndings and the recommendations derived from them in this dissertation will allow agent developers to set appropriate context for more e cient and successful ad hoc collaboration between human and agent peers in virtual teams. This dissertation enhances our understanding of the changes in human trust in peer level agent teammates with respect to agent trustworthiness, past experience, and agent reputation. The empirical ndings provide clear and signi cant evidence of the in uence of key factors on human trust in virtual agent teammates. We also derive several guidelines for agents to follow in practice.
Benzer Tezler
- Göçmen öğrencilerin okula uyum sürecinde kullanılacak 3B sanal öğrenme ortamının geliştirilmesi ve etkililiğinin incelenmesi
Developing and examining the effectiveness of the 3D virtual learning environment to be used in the school adaptation process of immigrant students
İBRAHİM ÖZKAL
Yüksek Lisans
Türkçe
2021
Eğitim ve ÖğretimTokat Gaziosmanpaşa ÜniversitesiBilgisayar ve Öğretim Teknolojileri Eğitimi Ana Bilim Dalı
DOÇ. DR. KEREM KILIÇER
- İnternetten alışverişte marka imajı ve marka güveni ilişkisi üzerine bir araştırma
A research on the relationship of brand image and brandsecurity in online shopping
MUHAMMED ENES SAĞ
Yüksek Lisans
Türkçe
2020
İşletmeBahçeşehir Üniversitesiİşletme Bilim Dalı
YRD. DOÇ. CEYDA AYSUNA TÜRKYILMAZ
- Sanal market uygulamalarının kullanımını etkileyen faktörler üzerine bir alan çalışması
A field study on the factors affecting the use of the online market applications
İPEK ÖZGAZİ
- E-hizmet kalitesi, web sitesine duyulan güven ve algılanan riskin müşteri memnuniyeti ve müşterilerin olumlu davranışlarına olan etkisi
The effect of e-service quality, trust in website and perceived risk on customer satisfaction and positive behaviour of customers
MEHMET DEMİRDÖĞMEZ
- Süpermarketler açısından e-ticarette itibarın sürdürülebilirliği üzerine bir araştırma
A survey on the sustainability of reputation in e-commerce for supermarkets
SEDA MUTLU
Yüksek Lisans
Türkçe
2018
Halkla İlişkilerSakarya ÜniversitesiHalkla İlişkiler ve Reklamcılık Ana Bilim Dalı
DOÇ. CENGİZ ERDAL