Category: News & Events

Robert Sammarco and Jane Greenberg presented “Human-in-the-Loop and AI: Crowdsourcing
Metadata Vocabulary for Materials Science” at the Metadata and Semantics Research Conference held in Thessaloniki, Greece, December 15-19. They presented findings on a recent proof of concept study which investigated the functionality of MatSci-YAMZ application. The research team looked at how human researchers defined terms in comparison to those produced by an LLM, Gemma3:27b. Slides from their presentation can be seen here. The conference paper with the same title as their presentation can be found here. AI World featured it as one of the 5 notable papers of the week.

Dave Breen presented “Metadata and 3D Shape Similarity: Establishing a Ground Truth Dataset.” Slides from his presentation can be found here, and a copy of the accompanying paper can be accessed here. The abstract for the paper is as follows:

We have developed a shape-based object retrieval method which is trained using computational feature metadata that encode 3D shape descriptors. A shape similarity metric is critical for identifying 3D objects according to their geometric and structural features, rather than relying solely on type descriptions. Since existing 3D model datasets lack ground truth that captures geometric/shape properties, we have developed an approach for establishing robust ground truth metadata for 3D shape similarity. Due to the absence of a suitable 3D object dataset that includes shape characteristics, we created our own ground truth collection using an image-based retrieval method from Azure AI Vision Studio. The newly defined computational metadata for 3D shape have been utilized to produce a rotation-invariant shape-based object retrieval capability.

The ID4 All Hands meeting was held at the Colorado School of Mines on October 13-14, 2025. The MRC was represented by Kio Polson, Scott McClellan, Xintong Zhao, and Dave Breen.

Scott McClellan and Colton Gerber (Toberer Group) presented on MatSci-YAMZ, an LLM-augmented version of Yet Another Metadata Zoo (YAMZ). MatSci-YAMZ is dedicated to studying vocabulary development in the materials science community and investigating the role large language models might play. Their presentation engaged audience members to test the new system by entering terms and definitions as well as commenting and voting on definitions produced by the LLM. Slides from their presentation can be seen here.

Prof. David Breen gave the presentation “AI-Ready Data: Knowledge Extraction from Chemistry Lab Notebooks’.  The talk summarized the MRC’s research on converting hand-written chemistry lab notebooks into a structured digital form, making their data AI-ready, i.e. amenable for downstream analysis and model training. The three steps in the conversion process are: 1) automatic segmentation of the notebook pages’ components, 2) extraction of structured data from the components, and 3) error analysis and correction of the data. Slides from Dr. Breen’s presentation can be found here.

MRC member Tim Gorichanaz is organizing this year’s annual meeting of the Document Academy (Docam ’24) taking place from September 18-20. This year’s theme is “Documents from the Future.”

Documents involve technology, meaning they change with the times. The ancient Latin root of “document” referred to an oral teaching, and centuries later the most common form of document was a piece of paper. Now, with computing widespread, perhaps most documents are digital. What will tomorrow’s documents bring, particularly in light of generative AI? What might change, and what might stay the same?

http://documentacademy.org/?2024

Docam ’24 will feature an array of presentations exploring the social, cultural, and technological conceptualizations and effects of documents and their possible future iterations. Among the presenters at this year’s meeting will be the MRC members Chris Rauch, Mat Kelly, and Hyung Wook Choi. The conference agenda, including a full list of presenters and presentations can be found here.

The Document Academy was founded in 2001 by Maribeth Back and Niels Windfeld Lund. The organization is dedicated to exploring documents and documentation through a variety of media and means. For more information about the Document Academy, please visit their website: https://documentacademy.org/

This summer, the MRC hosted three NSF REUs.

Elizabeth (Lizzie) Jones, Northeastern University (Project: AI-ready data: Knowledge Extraction from Laboratory Notebooks). Lizzie pursued document segmentation, optical character recognition, and text tokenization to extract research protocols and results from digitized lab notebooks produced by members of the Reticular Synthesis Laboratory led by Fernando Uribe-Romo, University of Central Florida. Project aim: To make archival laboratory notebook data AI-ready. (Mentors include Drexel Joel Pepper, David Breen, and Jane Greenberg.)

Robert Sammarco, Drexel University(Project: Developing YAMZ (Yet Another Metadata Zoo) for Materials Science Terminology). Robert extended the YAMZ foundation, and developed a materials science terminology portal. Project aim: To achieve better data interoperability, support the FAIR data principles, and help materials scientists better communicate with-in and across subdomains. (Mentors: Christopher Rauch, John Kunze, and Mat Kelly)

Rob Fleur, University of Michigan (Project: Knowledge Graph Implementation for Materials Science). Rob worked on automatic knowledge graph generation, drawing from an extensive collection of materials science research literature. Project aim: To help researchers more expediently extract knowledge from research literature. Rob will continue his work over September 2024. (Mentor: Alex Kalinowski)

Lizzie and Robert S. also participated in the ID4 REU end-of-summer event at Northwestern University in August 2024, and they each presented posters on their work. Kudos to all our REUs and their awesome accomplishments!

*all ID4 mentors are affiliated with Drexel’s MRC

On June 11, 2024, Dave Breen gave a presentation titled “Image Informatics for Metadata Extraction and Verification of Museum Specimen Images” at the Advances in Digital Media Workshop Series at the Yale Peabody Museum. The series is part of the Integrated Digital Biocollections (iDigBio) project and sought to answer the question, “How can we use media technologies to position biodiversity collections for even greater relevance to science, society, and Earth’s biota in the future?”

Dr. Breen’s slides can be found here.

Please join the MRC on April 17 from 11AM-12PM in Room 928 of the College of Computing and Informatics for “Accelerating Artificial Intelligence for Data-Driven Discovery” a talk delivered by Shih-Chieh Hsu, PhD, University of Washington.

Abstract

As scientific datasets become progressively larger, algorithms to process this data quickly become more complex. In response, Artificial Intelligence (AI) has emerged as a solution to efficiently analyze these massive datasets. Emerging processor technologies such as graphics processing units (GPUs) and field-programmable gate arrays (FPGAs) allow AI algorithms to be greatly accelerated. The Accelerated AI Algorithms for Data-Driven Discovery (A3D3) Institute sponsored by the National Science Foundation under the Harnessing the Data Revolution program is established to enable real-time AI at scale for broad applications. In this talk, Hsu will give an overview about the challenges of high energy physics, multi-messenger astrophysics and neuroscience regarding AI across latency and throughput regimes. He will introduce various techniques for model compression using state-of-the-art techniques such as pruning and quantization for edge computing. He will demonstrate that acceleration of AI inference as a web service represents a heterogeneous computing solution. Finally Hsu will discuss how A3D3 can bring together disparate communities that are threaded by common data-intensive grand challenges to accelerate discovery in science and engineering.

Biography

Shih-Chieh Hsu, PhD is a professor in physics and adjunct professor in electrical and computer engineering at University of Washington (UW), and director of NSF HDR Institute: Accelerated Artificial Intelligence Algorithms for Data-Driven Discovery. He earned the BS/MS in physics from National Taiwan University and the PhD in Physics from University of California San Diego. He is working on experimental particle physics using proton-proton collision data from the Large Hadron Collider. His research interests range from dark matter searches with the ATLAS experiment neutrino cross-section measurements with the FASER experiment innovative artificial intelligence algorithms for data-intensive discovery and accelerated machine learning with heterogeneous computing.