More Like this

1. Swarm Intelligence Amplifies the IQ of Collaborating Teams

   https://doi.org/10.1109/AI4I.2019.00036  

   Rosenberg, Louis; Willcox, Gregg (September 2019, ai4i 2019)

   In the natural world, Swarm Intelligence (SI) is a well-known phenomenon that enables groups of organisms to make collective decisions with significantly greater accuracy than the individuals could do on their own. In recent years, a new AI technology called Artificial Swarm Intelligence (ASI) has been developed that enables similar benefits for human teams. It works by connecting networked teams into real-time systems modeled on natural swarms. Referred to commonly as “human swarms” or “hive minds,” these closed-loop systems have been shown to amplify group performance across a wide range of tasks, from financial forecasting to strategic decision-making. The current study explores the ability of ASI technology to amplify the IQ of small teams. Five small teams answered a series of questions from a commonly used intelligence test known as the Raven’s Standard Progressive Matrices (RSPM) test. Participants took the test first as individuals, and then as groups moderated by swarming algorithms (i.e. “swarms”). The average individual achieved 53.7% correct, while the average swarm achieved 76.7% correct, corresponding to an estimated IQ increase of 14 points. When the individual responses were aggregated by majority vote, the groups scored 56.7% correct, still 12 IQ points less than the real-time swarming method. 

   more » « less
2. Artificial Swarm Intelligence employed to Amplify Diagnostic Accuracy in Radiology

   https://doi.org/10.1109/IEMCON.2018.8614883  

   Rosenberg, Louis; Lungren, Matthew; Halabi, Safwan; Willcox, Gregg; Baltaxe, David; Lyons, Mimi (November 2018, IEMCON 2018)

   Swarm Intelligence (SI) is a biological phenomenon in which groups of organisms amplify their combined intelligence by forming real-time systems. It has been studied for decades in fish schools, bird flocks, and bee swarms. Recent advances in networking and AI technologies have enabled distributed human groups to form closed-loop systems modeled after natural swarms. The process is referred to as Artificial Swarm Intelligence (ASI) and has been shown to significantly amplify group intelligence. The present research applies ASI technology to the field of medicine, exploring if small groups of networked radiologists can improve their diagnostic accuracy when reviewing chest X-rays for the presence of pneumonia by “thinking together” as an ASI system. Data was collected for individual diagnoses as well as for diagnoses made by the group working as a real-time ASI system. Diagnoses were also collected using a state-of-the-art deep learning system developed by Stanford University School of Medicine. Results showed that a small group of networked radiologists, when working as a real-time closed-loop ASI system, was significantly more accurate than the individuals on their own, reducing errors by 33%, as well as significantly more accurate (22%) than a state- of-the-art software-only solution using deep learning. 

   more » « less
3. Sales Forecasting, Polls vs Swarms

   https://doi.org/10.2139/ssrn.3390043  

   Willcox, Gregg; Rosenberg, Louis; Schumann, Hans (March 2019, SSRN Electronic Journal)

   Sales forecasts are critical to businesses of all sizes, enabling teams to project revenue, prioritize marketing, plan distribution, and scale inventory levels. To date, however, sales forecasts of new products have been shown to be highly inaccurate, due in large part to the lack of data about each new product and the subjective judgements required to compensate for this lack of data. The present study explores product sales forecasting performed by human groups and compares the accuracy of group forecasts generated by traditional polls to those made using Artificial Swarm Intelligence (ASI), a technique which has been shown to amplify the forecasting accuracy of groups in a wide range of fields. In collaboration with a major fashion retailer and a major fashion publisher, groups of fashion-conscious millennial women predicted the relative sales volumes of eight sweaters promoted during the 2018 holiday season, first by ranking each sweater’s sales in an online poll, and then using Swarm software to form an ASI system. The Swarm-based forecast was significantly more accurate than the poll. In fact, the top four sweaters ranked by swarm sold 23.7% more units, or $600,000 worth of sweaters during the target period, as compared to the top four sweaters as ranked by survey, (p = 0.0497), indicating that swarms of small consumer groups can be used to forecast sales with significantly higher accuracy than a traditional poll. 

   more » « less
4. Can We Obtain Fairness For Free?

   https://doi.org/10.1145/3461702.3462614  

   Islam, Rashidul; Pan, Shimei; Foulds, James R. (May 2021, AIES '21: Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society 2021)

   null (Ed.)

   There is growing awareness that AI and machine learning systems can in some cases learn to behave in unfair and discriminatory ways with harmful consequences. However, despite an enormous amount of research, techniques for ensuring AI fairness have yet to see widespread deployment in real systems. One of the main barriers is the conventional wisdom that fairness brings a cost in predictive performance metrics such as accuracy which could affect an organization's bottom-line. In this paper we take a closer look at this concern. Clearly fairness/performance trade-offs exist, but are they inevitable? In contrast to the conventional wisdom, we find that it is frequently possible, indeed straightforward, to improve on a trained model's fairness without sacrificing predictive performance. We systematically study the behavior of fair learning algorithms on a range of benchmark datasets, showing that it is possible to improve fairness to some degree with no loss (or even an improvement) in predictive performance via a sensible hyper-parameter selection strategy. Our results reveal a pathway toward increasing the deployment of fair AI methods, with potentially substantial positive real-world impacts. 

   more » « less
5. Impact of Student/Team Characteristics on Design Project Outcomes in Senior Design Courses

   Godbole, Hrushikesh; DeBartolo, Elizabeth; Takai, Shun; Esterman, Marcos (June 2023, ASEE Annual Conference proceedings)

   Ability to effectively work in teams is one of the desired outcomes of engineering and engineering technology programs. Unfortunately, working in teams is still challenging for many students. Rather than contributing to team projects, some students resort to social loafing. Social loafing tends to destroy both teamwork performance and individual learning, especially in solving ill-structured problems, such as design. Furthermore, a bad experience on a past team is a significant concern as it could generate negative feelings toward future team projects. Formation of collaborative teams is a critical first step in team-project-based design courses as team composition directly affects not only teamwork processes and outcomes, but also teamwork skills and experience. This NSF-IUSE sponsored project aims to enhance students’ teamwork experiences and teamwork learning through 1) understanding how to form better student design teams and 2) identifying exercises that will effectively improve team member collaboration. We do this by comparing student characteristics and design task characteristics with the quality of the design team outcome, and examining the resulting correlations. Student characteristics cover six categories: 1) background information, 2) work structure preferences, 3) personality, 4) ability, 5) motivation, and 6) attitude. Task characteristics and design team outcomes are characterized using the Creative Product Semantic Scale. In this article, we present correlations between student/team characteristics and design project outcome, and correlations between task characteristics and design project outcome for 2020-2021 senior design teams at two institutions. For both institutions, we will present correlations between individual student characteristics and team outcome. For one institution, we will also present correlation between team-level characteristics and team outcomes. 

   more » « less