The projects in this repository build theories, algorithms, and frameworks that can be used to design network-aware machine learning tools aimed at eliciting useful diversity and improving the accuracy of collective forecasting. Researchers in the social and economic sciences know that there is great capacity for collective intelligence to emerge from Web-based systems. Yet herding and homophily effects often restrain the wisdom of crowds, vastly limiting this potential. The research furthers the study of complex systems by introducing a new framework that improves our understanding of the mechanisms that govern decision-making under social influence. Advancing complex systems theory in this way greatly enhances the ability to predict when crowds will provide accurate decision-making support for complex problems and when they will fail miserably. Further, the research aids the development of opinion aggregation mechanisms that efficiently capitalize on diversity. The planned work will result in developments that make collective intelligence detection tools practical by providing early warning signs of shared misconceptions.

To attain these goals, the research will apply a general framework that incorporates (1) network models that help understand the social processes that lead to observed decision patterns; (2) machine learning tools that draw from uncovered processes to identify signals that optimize the accuracy of collective judgment; and (3) evaluation testbeds that use simulation tools in addition to rich high-dimensional, real-world data about the various stages and performance of group decisions. This research will contribute to societally-relevant outcomes, including: (a) understanding decision-making in online investment and lending settings to enhance the economic growth of underserved market segments; (b) generating novel knowledge about the performance benefits of collective judgment, and (c) quantifying the link between limited opinion diversity and crowd misconceptions. The project will connect undergraduate students, including women and under-represented minorities, to authentic practice in science and engineering research.

The following projects are funded by the National Science Foundation (NSF) through a CRII Grant IIS-1755873. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

Crowd financing is a burgeoning phenomenon that promises to improve access to capital by enabling borrowers with limited financial opportunities to receive small contributions from individual lenders towards unsecured loan requests. Faced with information asymmetry about borrowers' credibility, individual lenders bear the entire loss in case of loan default. Predicting loan payment is therefore crucial for lenders and for the sustainability of these platforms. To this end, we examine whether the ''wisdom'' of the lending crowd can provide reliable decision support with respect to projects' long-term success. Using data from Prosper, we investigate the association between the dynamics of lending behaviour and successful loan payment through interpretable classification models. We find evidence for collective intelligence signals in lending behaviour and observe variability in crowd wisdom across loan categories. We find that the wisdom of the lending crowd is most prominent in the auto loan category, but it is statistically significant for all other categories except student debt. Our study contributes new insights on how signals deduced from lending behaviour can improve the efficiency of crowd financing thereby contributing to economic growth and societal development.

Citation: Dambanemuya, H.K. and Horvát, E-Á., 2019. Harnessing Collective Intelligence in P2P Lending. Proceedings of the 10th ACM Conference on Web Science (WebSci), pp.57-64. DOI: https://doi.org/10.1145/3292522.3326040

The growing popularity of online fundraising (aka ``crowdfunding'') has attracted significant research on the subject. In contrast to previous studies that attempt to predict the success of crowdfunded projects based on specific characteristics of the projects and their creators, we present a more general approach that focuses on crowd dynamics and is robust to the particularities of different crowdfunding platforms. We rely on a multi-level analysis to investigate the correlates predictive importance, and quasi-causal effects of features that describe crowd dynamics in determining the success of crowdfunded projects. By applying a multi-level analysis to a study of fundraising in three different online markets, we uncover general crowd dynamics that ultimately decide which projects will succeed. In all levels of analysis and across the three different platforms, we consistently find that funders' behavioural signals (1) are significantly correlated with fundraising success; (2) approximate fundraising outcomes better than the characteristics of projects and their creators such as credit grade, company valuation, and subject domain; and (3) have significant quasi-causal effects on fundraising outcomes while controlling for potentially confounding project variables. By showing that universal features deduced from crowd behaviour are predictive of fundraising success on different crowdfunding platforms, our work provides design-relevant insights about novel types of collective decision-making online. This research inspires thus potential ways to leverage cues from the crowd and catalyses research into crowd-aware system design.

Citation: Dambanemuya, H.K. and Horvát, E-Á. 2021. A Multi-Platform Study of Crowd Signals Associated with Successful Online Fundraising. Proceedings of the ACM (PACM): Computer-Supported Cooperative Work. Proc. ACM Hum.-Comput. Interact.5, CSCW1, Article 115. DOI: https://doi.org/10.1145/3449189

Extensive literature argues that crowds possess essential collective intelligence benefits that allow superior decision-making by untrained individuals working in low-information environments. Classic wisdom of crowds theory is based on evidence gathered from studying large groups of diverse and independent decisionmakers. Yet, most human decisions are reached in online settings of interconnected like-minded people that challenge these criteria. This observation raises a key question: Are there surprising expressions of collective intelligence online? Here, we explore whether crowds furnish collective intelligence benefits in crowdfunding systems. Crowdfunding has grown and diversified quickly over the past decade, expanding from funding aspirant creative works and supplying pro-social donations to enabling large citizen-funded urban projects and providing commercial interest-based unsecured loans. Using nearly 10 million loan contributions from a marketdominant lending platform, we find evidence for collective intelligence indicators in crowdfunding. Our results, which are based on a two-stage Heckman selection model, indicate that opinion diversity and the speed at which funds are contributed predict who gets funded and who repays, even after accounting for traditional measures of creditworthiness. Moreover, crowds work consistently well in correctly assessing the outcome of high-risk projects. Finally, diversity and speed serve as early warning signals when inferring fundraising based solely on the initial part of the campaign. Our f indings broaden the field of crowd-aware system design and inform discussions about the augmentation of traditional financing systems with tech innovations.

Citation: Horvát, E-Á., Dambanemuya, H.K., Uparna, J.S., and Uzzi, B., 2023. Hidden Indicators of Collective Intelligence in Crowdfunding. Proceedings of the ACM Web Conference (WWW). DOI: https://doi.org/10.1145/3543507.3583414

Crowdfunding is increasingly transforming financing for many people across the globe. We conduct two studies of 𝑁 = 500 and 𝑁 = 750 participants involved in crowdfunding to investigate the effect of characteristics of prior contributions (“crowd signals”) on their funding decisions. First, we detect the presence of signaling and then demonstrate its importance in a naturalistic context. We find that contributions of varying amounts arriving at heterogeneous time intervals are 19.7% more likely to be selected than homogeneous contribution amounts and times. Although the impact of signaling is strongest among participants who are susceptible to social influence, the effect is remarkably general and typically unrecognized by participants who attribute their decisions to nonexistent differences in project descriptions. These findings hold across different project types, fundraising goals, interest levels in the projects, and participants’ altruistic attitudes. Our results underscore the importance of social signaling in crowdfunding, providing novel insights on how, when, and why signaling between funders is impacting funding outcomes.

Citation: Dambanemuya, H.K., Choi, E., Gergle, D., and Horvát, E-Á. Beyond Words: An Experimental Study of Signaling in Crowdfunding. Pre-print available at https://arxiv.org/abs/2206.07210

Investigations of social influence in collective decision-making have become possible due to recent technologies and platforms that record interactions in far larger groups than could be studied before. Herding and its impact on decision-making are critical areas of practical interest and research study. However, despite theoretical work suggesting that it matters whether individuals choose who to imitate based on cues such as experience or whether they herd at random, there is little empirical analysis of this distinction. To demonstrate the distinction between what the literature calls "rational" and "irrational" herding, we use data on tens of thousands of loans from a well-established online peer-to-peer (p2p) lending platform. First, we employ an empirical measure of memory in complex systems to measure herding in lending. Then, we illustrate a network-based approach to visualize herding. Finally, we model the impact of herding on collective outcomes. Our study reveals that loan performance is not solely determined by whether the lenders engage in herding or not. Instead, the interplay between herding and the imitated lenders' prior success on the platform predicts loan outcomes. In short, herds led by expert lenders tend to pick loans that do not default. We discuss the implications of this under-explored aspect of herding for platform designers, borrowers, and lenders. Our study advances collective intelligence theories based on a case of high-stakes group decision-making online.

Citation: Dambanemuya, H.K., Wachs, J., and Horvát, E-Á. Understanding (Ir)rational Herding Online. Pre-print available at (coming soon)