Behavior Latticing: Inferring User Motivations from Unstructured Interactions DOI: XXXXXXX.XXXXXXX CCS: Human-centered computing Interactive systems and tools CCS: Computing methodologies Natural language processing

Dora Zhao, Michelle S. Lam, Diyi Yang, Michael S. Bernstein · 2026

Read Behavior Latticing: Inferring User Motivations from Unstructured Interactions DOI: XXXXXXX.XXXXXXX CCS: Human-centered computing Interactive systems and tools CCS: Computing methodologies Natural language processing by Dora Zhao, Michelle S. Lam, Diyi Yang, Michael S. Bernstein (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-07629.txt.

Description

Zhao and three colleagues argue personal AI should grasp why users act. Today's systems tune or repeat old acts, such as having ChatGPT finish homework, even when that hurts deeper needs. Their design builds user grasp through behavior latticing: linking scattered acts into insights about hidden motives over long spans of data. This yields new powers, including reading needs rather than tasks. In tests, behavior latticing produces accurate insights with far greater depth than top approaches, and an agent steered by user insights serves needs better.

Catalog record

CREATORDora Zhao, Michelle S. Lam, Diyi Yang, Michael S. Bernstein
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE92916
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.07629

Browse the full library · About the Tower of Babel