Intent representation
Discover latent states that explain how a person’s goal evolves across a behavioral trajectory.
MAPPING HUMAN INTENT THROUGH ADVERTISING
ATLAS asks whether sequential behavior can reveal the decision state beneath it—before the outcome is explicit.
Enter the study ↓THE MISSING MIDDLE
Search, exploration, comparison, hesitation and commitment are visible. The intent and uncertainty driving them are not. ATLAS investigates whether those hidden decision states can be represented computationally.
TRAJECTORY STUDY / CONCEPTUAL INTERFACE
OBSERVATION ≠ UNDERSTANDINGAlternatives are being weighed against a live constraint.
The same visible move can come from very different decision states. This is the distinction under study.
RESEARCH, NOT PERSONAL PROFILINGCORE RESEARCH QUESTION
Not a fixed label attached to a person—a changing state that can account for what they see, reconsider and ultimately decide.
ACTIVE DIRECTIONS
ATLAS / 001—003Discover latent states that explain how a person’s goal evolves across a behavioral trajectory.
Model movement between exploration, comparison, uncertainty and commitment.
Study how new information changes a decision state—and when an AI should stay out of the way.
WHY ADVERTISING?
Advertising creates repeated encounters between people, information, alternatives and decisions. That makes it a useful environment for studying how a decision changes after each new piece of information.
RESEARCH BOUNDARY
ATLAS is a research program, not a personal profiling product. Any investigation must be grounded in consent, data governance, privacy-preserving methods and clear limits on what behavioral evidence can support.
Explore the lab’s other research ↗Interested in the research?
Talk to the ATLAS team. ↗