Shunyuan Zhang’s research while affiliated with Harvard Medical School and other places

... In comparison, while projects like CryptoPunks have innovated in terms of artistic style, Yuan Yuan, Xiao Liu, Shunyuan Zhang, and Kannan Srinivasan have noted that female CryptoPunks typically trade at a 36.8% lower price than male CryptoPunks, and Black CryptoPunks at a 30.7% lower price than their White counterparts [3]. This contrast offers a unique perspective on how digital art challenges social norms. ...

Reference:

The Shaping of Digital Collectibles: Digital Imprints of Gender and Cultural Representation in the Age of the Non-Fungible Tokens Boom

... Lastly, Bertolini, Clevert, and Montanari introduced an aggregation method that generalizes attribution maps between any two convolutional layers of a neural network (Bertolini, Clevert, and Montanari 2023). R-XAI has also been used in many applications, for instance in healthcare Weinberger, Lin, and Lee 2023) and business (Feng, Li, and Zhang 2023). ...

Reference:

REPEAT: Improving Uncertainty Estimation in Representation Learning Explainability

... The home-sharing economy involves uncertainty and information asymmetry between peer providers and customers, and photos can reduce the uncertainty by heightening social presence and the ability to visualize the experience (Ert, Fleischer, & Magen, 2016). Customers in the P2P context rely heavily on photos to make purchase decisions (Zhang, Mehta, Singh, & Srinivasan, 2019). Prior studies have tested several effects involving the photos of peer providers and properties. ...

Reference:

The impact of photo verification service on sales performance in the peer-to-peer economy: Moderating role of customer uncertainty

... This support can come from family recommendations or observed use by peers. Social influences are crucial as they affect how older adults think about and feel toward technology, influencing their digital engagement (Zhang S Y, et al., 2023). Despite the clear importance of these social aspects, research on leveraging this influence to promote elderly Fintech engagement is limited. ...

Reference:

Exploring the mechanisms driving elderly Fintech engagement: the role of social influence and the elderly’s digital literacy

... As both of these levers can further enhance generative text-to-image models' effectiveness (Jansen et al., 2024;Rombach et al., 2022), our results likely represent a lower bound for the performance of AI-generated marketing imagery. The emergence of future generative AI models will likely improve the synthetic images' perceptual ratings and realworld effectiveness, especially when combined with task-specific data for model calibration (Feng et al., 2023). ...

Reference:

The power of generative marketing: Can generative AI create superhuman visual marketing content?

... Other publications directly predict relevant Marketing Outcomes from images. Examples include return rate prediction (Dzyabura et al. 2019), inference of design aesthetics (Burnap et al. 2021), or other subjective variables such as number of likes, or visual potential of being a celebrity (Feng et al. 2022) (e.g., for selecting suitable endorsers for campaign collaboration). Lastly, articles on Content & Context identify product content on images or classify the product usage or consumption context. ...

Reference:

Performance of automated image classification

... Pioneering models such as Convolutional Neural Networks (CNNs) [28], Vision Transformers (ViT) [14], and Diffusion Models [56,57] have demonstrated efficacy across di-* Equal contribution. verse applications, including medical imaging, retail, and e-commerce [9,19,35,40,58,70]. These models can be applied as tools for image editing tasks, significantly enhancing image customization for specific needs. ...

Reference:

AI Tailoring: Evaluating Influence of Image Features on Fashion Product Popularity

... Concentration is typically identified as congestion in the urban context [17,28,31,32,68,83,[117][118][119][120][121]. Segregation is a situation in which groups of users are set apart from each other [122,123]. Some outcomes emerge at the individual or model level only. ...

Reference:

Human-AI coevolution

... Race significantly shapes users' perceptions and trust in AI (M. K. Lee & Rich, 2021;S. Zhang et al., 2021). We argue that Non-White users in America are more likely to experience algorithmic aversion or automation bias compared to White users. ...

Reference:

Communicating and combating algorithmic bias: effects of data diversity, labeler diversity, performance bias, and user feedback on AI trust

... Given their importance, profile photos have long caught the attention of scholars, who started studying them in the first decade of the twentyfirst century from different fields such as psychology (e.g., Hudson & Gore, 2017;Leary & Allen, 2011), communication (Bond, 2009;Gibbs et al., 2006;Hancock & Toma, 2009;Mesch & Beker, 2010, among many others), or marketing and consumer behaviour (Bente et al., 2012;Fagerstrøm et al., 2017;Jang, 2022;Tussyadiah & Park, 2018;Zhang et al., 2020). ...

Reference:

The Profile Picture