Artificial Intelligence/Machine Learning Research at IARPA
As part of its mission to address some of the most difficult challenges in the Intelligence Community, IARPA sponsors research programs and challenges that either leverage or improve Artificial Intelligence/Machine Learning (AI/ML), including:
- Aladdin Video, pioneered machine learning techniques in video by combining the state-of-the-art in video and audio extraction, knowledge representation, and search technologies to create a fast, accurate, robust, and extensible video search capability;
- Better Extraction from Text Towards Enhanced Retrieval (BETTER), develop AI/ML-based methods for extracting increasingly fine-grained semantic information, with a focus of events in the form of who-did-what-to-whom-whenwhere, across multiple languages and problem domains.
- Cyber-attack Automated Unconventional Sensor Environment (CAUSE), applies AI/ML-based models to develop novel, automated methods for event-based detection and prediction of cyber-attacks significantly earlier than existing approaches. Forecasting cyber-attack events with actionable details advances the state-of-the-art by enabling threat-specific cyber incident response and defense measures;
- Creation of Operationally Realistic 3D Environment (CORE3D), uses machine learning and deep learning techniques to develop methods for the construction of a fully automated high fidelity 3D model of the world using remote sensing data;
- Deep Intermodal Video Analytics (DIVA), leverages machine learning techniques to develop robust automatic activity detection in streaming video across multiple cameras;
- Finding Engineering-Linked Indicators (FELIX), uses AI for detection of engineering signatures across multiple biological organisms. The goal is to distinguish natural organisms from those that have been engineered;
- Functional Map of the World Challenge, developed algorithms that would quickly and accurately classify 63 classes of buildings and regions in satellite imagery. All the top participants used various forms of deep learning;
- Functional Genomic and Computational Assessment of Threats (Fun GCAT), develops AI/ML-based approaches to learn and classify genetic (e.g., DNA) sequence data by genetic taxonomy, sequence function, and threat potential;
- Mercury Challenge, asked challenge participants to make use of AI/ML approaches to forecast a variety of political events in the Middle East and North Africa region, such as non-violent civil unrest and military activity;
- Machine Intelligence from Cortical Networks (MICrONS), aims to revolutionize machine learning by reverse-engineering the algorithms of the brain. The program is expressly designed as a dialogue between data science and neuroscience;
- Machine Translation for English Retrieval of Information in Any Language (MATERIAL), develops machine learning methods to identify foreign language information from speech and text relevant to English queries, and providing evidence of relevance of the retrieved information in English in a meaningful way. Algorithms will be developed under low resource conditions and without foreign language expertise;
- Modeling of Reflectance Given Only Transmission of High-concentration Spectra for Chemical Recognition Over Widely-varying eNvironments (MORGOTH’S CROWN) Challenge, solicited new approaches for predicting the influence of effects such as substrate, loading, and deposition characteristics on the infrared spectra of trace chemicals on surfaces. Top participants used machine learning techniques;
- Multimodal Objective Sensing to Assess Individuals with Context (MOSAIC), extracts contextually-meaningful data from a variety of individual, environmental and social sensing data streams and uses machine learning and artificial intelligence-based models to estimate and predict psychological, cognitive, and physiological constructs, as well as estimates of overall job performance;
- Multi-View Stereo 3D Mapping (MSV) Challenge, required groups to develop algorithms that generated high resolution 3D point clouds of the physical world. It also created research opportunities to a new segment of the computer vision community by providing a baseline reconstruction algorithm for satellite imagery;
- OpenCLIR Challenge, developed machine learning methods in a low training data condition to retrieve Swahili speech and text documents relevant to English queries;
- Rapid Analysis of Various Emerging Nanoelectronics (RAVEN), uses AI/ML to accelerate the speed and accuracy of image processing for state-of-the-art integrated circuits;
- Strengthening Human Adaptive Reasoning and Problem-Solving (SHARP), used machine learning models of behavioral (e.g., test scores, cognitive tasks, self-report measures) and brain–based measures (e.g., MRI, EEG) to predict intelligence scores, responsiveness to interventions, and baseline neurophysiological and cognitive correlates of intelligence;
- Virtuous User Environment (VirtUE), uses adaptive learning algorithms to build analytic tools and technologies that would identify and respond to deviations in normal computer user activity which could prevent zero day attacks.
IARPA is always seeking novel ideas aligned with our mission. If you are interested in working with IARPA through one of our existing solicitations, prize challenges, requests for information, or other mechanisms, please see this link for more details.
Related Article(s)
- IARPA VirtUE focuses on user roles for security
- Spy Agencies Need AI to Get Word on Street, Predict Events
- DNA cops make sure deadly viruses don’t get rebuilt
- IARPA Launches Satellite Imagery Program
- IARPA Wants to Use Machine Learning to Help Prevent Terrorist Attacks
- Predictive Cybersecurity: Prepare for Attackers Before They’re at Your Door
- Big Data Researchers Aim to Make Sense of Video