Virginia Tech® home

Spotlight on the Crowd Intelligence Lab led by Kurt Luther

November 27, 2023

Vikram Mohanty (second from the right) and Kurt Luther (right) receiving the Best Poster/Demo Award at AAAI HCOMP 2023.
Vikram Mohanty (second from the right) and Kurt Luther (right) receiving the Best Poster/Demo Award at AAAI HCOMP 2023.

The Crowd Intelligence Lab at VT, directed by Kurt Luther, researches the complementary strengths of crowdsourced human intelligence and artificial intelligence (AI) in domains like journalism, history, national security, and creativity. Luther is an associate professor in the VT Department of Computer Science and Associate Director for Research in the Center for Human-Computer Interaction (CHCI).

Recently, members of the Crowd Intelligence Lab presented three papers at the joint AAAI Conference on Human Computation and Crowdsourcing (HCOMP) and ACM Collective Intelligence Conference (CI) in Delft, Netherlands, from November 6-9, 2023. 

HCOMP is the premier venue for disseminating the latest research findings on human computation and crowdsourcing. It contributes to human-centered qualitative studies and HCI design, social computing, artificial intelligence, economics, computational social science, digital humanities, policy, and ethics. CI brings together researchers from academia, business, nonprofits, governments, and the world at large to share insights and ideas relevant to understanding and designing collective intelligence in its many forms.

Both Luther and co-author Vikram Mohanty, who successfully defended his Computer Science Ph.D. from VT earlier this year, were involved as conference organizers, attended in person, and chaired paper sessions. Mohanty gave the paper presentations and demo, and served as a reviewer on the CI program committee. Luther was co-chair of the Works-in-Progress and Demos venue.

A full paper and demonstration paper accepted to HCOMP were co-authored by VT CS/CHCI MS alumni Jude Lim and Terryl Dodson, Mohanty, and Luther. The papers are about BackTrace, a new software tool built to use AI to match painted backdrops in historical photos, which can be used to help identify the subjects of the photos. The authors demonstrated the software at the conference, where it won the Best Poster/Demo Award! This was the third HCOMP Best Poster/Demo Award won by the Crowd Lab, and the second for Mohanty.

Additionally, a full paper by Mohanty and Luther about Photo Steward was accepted at the Collective Intelligence conference. Photo Steward provides an enhancement of the lab’s Civil War Photo Sleuth website that combines visualization and collaboration to improve the accuracy of photo identifications.

A visual representation of BackTrace’s Workflow
BackTrace’s Workflow

Current projects of the Crowd Intelligence Lab focus on improving open source intelligence (OSINT) analysis, combating disinformation and misinformation, identifying unknown people and places in historical and modern photos, and understanding real-world crowdsourced investigations. With new funding from the National Endowment for the Humanities (NEH) Program on Digital Projects for the Public (2023-25), members of the lab are also part of an interdisciplinary team creating a prototype of an Augmented Reality application for visitors to Pamplin Historical Park in Petersburg, Virginia. Using mobile devices, visitors will learn about less familiar Civil War era topics: interconnections between the environment and military affairs; the war’s transformative impact on African Americans and civilians; and the benefits of reading both wartime documents and material artifacts with a historian’s eye. 

Going forward, Luther plans to continue pursuing research opportunities at the intersection of HCI and digital history and cultural heritage, including an upcoming workshop on that topic he and Mohanty are co-organizing at the ACM Intelligent User Interfaces (IUI) 2024 conference with colleagues at the Library of Congress and Utrecht University. They will also be continuing research on the Civil War Photo Sleuth website in collaboration with Crowd Lab members and other VT colleagues. “Virginia Tech has an amazing community of researchers exploring how HCI can illuminate our understanding of the past,” Luther said. “I’m grateful to be part of it and excited to see what’s next.”