Request access to AVI-Personality
We review requests manually (similar to the previous email process).
Access requires agreement to the AVI-Personality User Agreement (see user_agreement.pdf in this repository, or https://apal-seu.github.io/AVI6/user_agreement.html). Academic, non-commercial research only. Do not redistribute videos or attempt to re-identify participants.
Log in or Sign Up to review the conditions and access this dataset content.
AVI-Personality
AVI-Personality: A Trait-Activated Multimodal Dataset for Personality and Competency Assessment in Asynchronous Video Interviews
Tianyi Zhang, Jinwenxi Shang, Antonis Koutsoumpis, Yuan Zong*, Reinout E. de Vries, Wenming Zheng
Paper: arXiv:2608.25316 · Hugging Face Papers
Links
| Resource | Link |
|---|---|
| Paper (arXiv) | https://arxiv.org/abs/2608.25316 |
| Paper (HF Papers) | https://huggingface.co/papers/2608.25316 |
| Project Page | English, 中文 |
| User Agreement | AVI-Personality EULA · user_agreement.pdf |
| Dataset access | Request access on this Hugging Face page (gated). Backup contact: t.zhang@seu.edu.cn |
News
| Date | Update |
|---|---|
| 2026-09 | Full release: annotations + 3,876 interview videos (gated). Paper on arXiv. |
| 2026-08 | Project page and user agreement released. Dataset hosting on Hugging Face (gated). |
Overview
AVI-Personality is a trait-activated multimodal dataset for personality and job-related competency assessment from asynchronous video interviews. The dataset contains 3,876 interview videos from 646 participants who completed a simulated management traineeship application. Participants answered two generic selection questions and four personality-targeted questions designed according to Trait Activation Theory. AVI-Personality provides HEXACO self-reports, observer-rated personality, recruiter-rated competencies, interview performance, and cognitive ability.
Key Features
- Trait-activated AVI protocol — 2 generic questions + 4 personality questions targeting Honesty-Humility, Extraversion, Agreeableness, and Conscientiousness
- Psychometrically grounded labels — HEXACO-100 self-reports and BARS ratings by trained psychologists and professional recruiters
- Dual assessment targets — personality traits and job-related competencies in the same interview setting
- Multimodal release — video, demographics, and an official subject-level split
- Academic access — available after agreeing to the user agreement (gated download on Hugging Face)
Dataset Statistics
| Property | Value |
|---|---|
| Participants | 646 (309 men / 309 women / 28 non-binary) |
| Videos | 3,876 (~85 hours) |
| Age | 36.69 ± 11.92 years |
| Work experience | 15.96 ± 11.36 years |
| Personality labels | HEXACO self-reports + observer ratings |
| Competency labels | Integrity, Collegiality, Social versatility, Development orientation, Interview performance |
| Cognitive ability | ICAR g-score (0–28, mean 14.51) |
| Official split | Subject-level 70 / 10 / 20 (452 / 64 / 130) |
Files in this repository
| Path | Description |
|---|---|
Training.csv |
Training split annotations (n=452) |
Validation.csv |
Validation split annotations (n=64) |
Testing.csv |
Test split annotations (n=130) |
Training/ |
Training videos (452 × 6) |
Validation/ |
Validation videos (64 × 6) |
Testing/ |
Test videos (130 × 6) |
CODEBOOK.md / CODEBOOK_zh.md |
Column definitions (EN / 中文) |
user_agreement.pdf |
End-user license / user agreement |
Video filenames follow {id}_q{1-6}_{generic|personality}.mp4 (q1–q2 generic, q3–q6 personality).
Getting Started
- Read the User Agreement (PDF also in this repo:
user_agreement.pdf). - On this dataset page, click Agree and request access and complete the form. Requests are reviewed manually. A researcher with a permanent position should be listed; students may be listed as participants.
- After approval, load the annotations:
from datasets import load_dataset
ds = load_dataset("APAL-SEU/AVI-Personality", token=True)
print(ds)
print(ds["train"][0])
Videos are under Training/, Validation/, and Testing/. After access is granted you can download them from the Files and versions tab, or fetch individual files via the Hub (e.g. hf://datasets/APAL-SEU/AVI-Personality/Training/{id}_q1_generic.mp4).
Citation
If you find AVI-Personality useful in your research, please cite:
@article{zhang2026AVI-Personality,
author={Zhang, Tianyi and Shang, Jinwenxi and Koutsoumpis, Antonis and Zong, Yuan and de Vries, Reinout E. and Zheng, Wenming},
title={AVI-Personality: A Trait-Activated Multimodal Dataset for Personality and Competency Assessment in Asynchronous Video Interviews},
year={2026},
eprint={2608.25316},
archivePrefix={arXiv},
primaryClass={cs.HC},
url={https://arxiv.org/abs/2608.25316},
}
License
The dataset is released for academic, non-commercial, not-for-profit research. Use of AVI-Personality requires agreeing to the User Agreement.
- Downloads last month
- 4