Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

HACHIMI: Scalable and Controllable Student Persona Generation via Orchestrated Agents

Yilin Jiang, Fei Tan, Xuanyu Yin, Jing Leng, Aimin Zhou

TL;DR

HACHIMI is introduced, a multi-agent Propose-Validate-Revise framework that generates theory-aligned, quota-controlled personas and provides a standardized synthetic student population for group-level benchmarking and social-science simulations.

Abstract

Student Personas (SPs) are emerging as infrastructure for educational LLMs, yet prior work often relies on ad-hoc prompting or hand-crafted profiles with limited control over educational theory and population distributions. We formalize this as Theory-Aligned and Distribution-Controllable Persona Generation (TAD-PG) and introduce HACHIMI, a multi-agent Propose-Validate-Revise framework that generates theory-aligned, quota-controlled personas. HACHIMI factorizes each persona into a theory-anchored educational schema, enforces developmental and psychological constraints via a neuro-symbolic validator, and combines stratified sampling with semantic deduplication to reduce mode collapse. The resulting HACHIMI-1M corpus comprises 1 million personas for Grades 1-12. Intrinsic evaluation shows near-perfect schema validity, accurate quotas, and substantial diversity, while external evaluation instantiates personas as student agents answering CEPS and PISA 2022 surveys; across 16 cohorts, math and curiosity/growth constructs align strongly between humans and agents, whereas classroom-climate and well-being constructs are only moderately aligned, revealing a fidelity gradient. All personas are generated with Qwen2.5-72B, and HACHIMI provides a standardized synthetic student population for group-level benchmarking and social-science simulations. Resources available at https://github.com/ZeroLoss-Lab/HACHIMI

HACHIMI: Scalable and Controllable Student Persona Generation via Orchestrated Agents

TL;DR

HACHIMI is introduced, a multi-agent Propose-Validate-Revise framework that generates theory-aligned, quota-controlled personas and provides a standardized synthetic student population for group-level benchmarking and social-science simulations.

Abstract

Student Personas (SPs) are emerging as infrastructure for educational LLMs, yet prior work often relies on ad-hoc prompting or hand-crafted profiles with limited control over educational theory and population distributions. We formalize this as Theory-Aligned and Distribution-Controllable Persona Generation (TAD-PG) and introduce HACHIMI, a multi-agent Propose-Validate-Revise framework that generates theory-aligned, quota-controlled personas. HACHIMI factorizes each persona into a theory-anchored educational schema, enforces developmental and psychological constraints via a neuro-symbolic validator, and combines stratified sampling with semantic deduplication to reduce mode collapse. The resulting HACHIMI-1M corpus comprises 1 million personas for Grades 1-12. Intrinsic evaluation shows near-perfect schema validity, accurate quotas, and substantial diversity, while external evaluation instantiates personas as student agents answering CEPS and PISA 2022 surveys; across 16 cohorts, math and curiosity/growth constructs align strongly between humans and agents, whereas classroom-climate and well-being constructs are only moderately aligned, revealing a fidelity gradient. All personas are generated with Qwen2.5-72B, and HACHIMI provides a standardized synthetic student population for group-level benchmarking and social-science simulations. Resources available at https://github.com/ZeroLoss-Lab/HACHIMI
Paper Structure (63 sections, 4 figures, 1 table, 1 algorithm)

This paper contains 63 sections, 4 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Classic Student Persona Modeling
  4. Student Persona Generation via LLMs
  5. Student Persona Datasets
  6. HACHIMI Framework
  7. Problem Formulation
  8. Theory-Anchored Persona Schema
  9. Multi-Agent Generation Architecture
  10. Mechanism I: Modular Generation via Shared Whiteboard.
  11. Mechanism II: Neuro-Symbolic Constraint Satisfaction.
  12. Mechanism III: Stratified Sampling and Diversity Control.
  13. The HACHIMI-1M Corpus
  14. Hybrid Semi-Structured Representation.
  15. Distribution via Stratified Sampling.
  16. ...and 48 more sections

Figures (4)

  • Figure 1: HACHIMI pipeline overview. From target distributions (grade/gender/academic level), steps (1)--(5) produce the HACHIMI-1M corpus.
  • Figure 2: Immersive role-playing prompt template used for HACHIMI student agents when answering CEPS- and PISA-based shadow surveys.
  • Figure 3: Pearson $r$ and Spearman $\rho$ between human and HACHIMI cohort means for each CEPS target.
  • Figure 4: Distribution of Pearson correlations between human and agent group means on PISA 2022, summarized by region.