Skip to main content
LAS Logo Laboratory for Analytic Sciences
Top of Page

Staff

flowchart

Dec 2, 2025

EvalOps: Effective Evaluation in a Changing AI Landscape

Bijesh Shrestha1, Hilson Shrestha1, Karen Bonilla1, Noëlle Rakotndravony1, Mengtian Guo2, Yue (Ray) Wang2*, R. Jordan Crouser3*, Lane T. Harrison1* Affiliations: 1 Worcester Polytechnic Institute, 2 UNC–Chapel Hill, 3 Kenyon College…
AI Generated Image from Adobe Firefly 2025

Dec 2, 2025

LLMming: Validating Models Without Ground Truth

Chris Argenta (Rockfish Research), John Slankas (NCSU LAS), Jascha Swisher (NCSU LAS), and Aaron W. (LAS GOV) Despite all the exciting progress in AI, our understanding of what large language models…
Images and maps posted on a social media platform of the Lilac fire.

Nov 25, 2025

Measuring the Added Information Utility of Visual Media for Summarizing Crises

Synthesizing informative summaries from noisy, diverse, and multimodal streams of documents is a long-standing problem across research domains (e.g., information retrieval, natural language processing, etc.). In this project, we extend…
Workflow diagram with multi-modal decomposition and CLAP-based audio retrieval.

Nov 25, 2025

WOLFSIGHT Looks at Video

Edward S., Liz Richerson, John Nolan, Nicole F. Analysts face the challenge of reviewing vast amounts of complex data that spans multiple modalities. In previous work we sought to understand…

Nov 25, 2025

Meta-Embeddings for Small Object Image Search

Analysts and researchers use image search tools to comb through huge collections of photos. Our team’s work in this space is specifically motivated by our efforts to support the National…
Sunburst data visualization

Dec 13, 2024

CARD: Content-Aware Recommendation of Data

We have been working with LAS to develop CARD, a front-end for our recommendation pipeline ORBS. ORBS is the culmination of our work with LAS in previous years, and with…

Dec 2, 2024

Improving Text-Guided Image Search 

Sometimes you have an image and are looking for similar images. Sometimes that search doesn’t work perfectly. Maybe you don’t have exactly the right image to start with, or maybe…
Man wearing headphones and looking at a computer

Nov 25, 2024

ECHOLAB: Leveraging LLMs to Boost Speech-to-Text Accuracy in Challenging Audio Environments

Speech-to-text (STT) transcriptions are crucial to audio analysis, forming a foundation for retrieval, prioritization, and sensemaking activities within the analyst workflow. The accuracy of STT output is essential; without it,…
An AI-generated image of a woman eating lunch. Errors in the image are circled.

Nov 25, 2024

5 Telltale Signs That a Photo Is AI-generated

While diffusion models, or artificial intelligence systems that create images from text prompts, are a simple way to make realistic-looking images, they still have a long way to go before…

Nov 25, 2024

Automated Workflows to Triage Technical Intelligence

Introduction In today’s rapidly evolving technological landscape, analysts face the daunting task of sifting through large volumes of unstructured data to identify critical trends and insights. Our team recently undertook…