Staff

Nov 25, 2024

Large Language Models for Intelligence Analysis

The Challenge: Can we show that LLMs can be useful for intelligence analysts? This post describes observations made at the Laboratory for Analytic Sciences under an umbrella of connected efforts…

Nov 25, 2024

Structured Annotation of Audio Data

Overview Analysis of audio data is an important activity in the intelligence community.  Typically, operators transcribe audio data into a text file.  While they transcribe, they may make annotations called…

Nov 25, 2024

Maturing the LAS Model Deployment Service

A machine learning (ML) model deployment architecture is a foundational component which is essential to effective use of ML in an enterprise. Establishing a set of consistent, repeatable and iterable…

Nov 25, 2024

Next-Gen Prompt Engineering: Provably Better

The rise of Large Language Models (LLMs) has fueled their rapid adoption across diverse fields, especially in software engineering, where their code generation skills and deep technical knowledge shine. While…

Nov 25, 2024

SANDGLOW

Challenge: Can open source multimodal LLMs be helpful for automated intelligence analysis of video? Advances in multimodal language models have revolutionized the field of video summarization, enabling AI systems to…

Nov 25, 2024

Generative AI and Intelligence Analysis

THE BIG PICTURE What Are the Implications of Generative AI on Intelligence Analysis? Over the past decade, Generative AI technologies across every modality (e.g. images, video, audio, and text) have…

Nov 25, 2024

Enhancing Technical Document Triage: Large Vision Language Models for Image-Text Information Extraction

Technical intelligence analysts are responsible for the collection, evaluation, analysis, and interpretation of foreign scientific and technical information. This work often involves scouring massive volumes of documents for crucial details,…

Nov 22, 2024

“Enough is Enough”: Exploring Information Satiety in Knowledge Workers

1.Overview Today’s knowledge workers face a common dilemma: weighing the desire for more information against the natural constraints of time and other limited resources. Striking a balance between comprehensive analysis…

Nov 22, 2024

EYEGLASS: Adapting Computer Vision Models to Drifting Data with Foundation Model Finetuning

Developing computer vision models for tasks like object detection and activity recognition typically require a significant amount of human-curated labeled data. This issue continues after model development and into production,…

Nov 22, 2024

Personalize User Workflow and Improve Technological Engagement

This project focuses on the “Insider Threat” evidence-finding game, developed by a team at the University of Kentucky with LAS funding. Christine Brugh was involved in this project. In the…