README.md

---
license: cc-by-4.0
task_categories:
  - video-classification
  - video-text-to-text
  - object-detection
tags:
  - egocentric-video
  - mistake-detection
  - temporal-localization
  - video-language-grounding
  - hand-object-interaction
  - action-recognition
  - procedural-activities
  - semantic-role-labeling
  - ego4d
  - epic-kitchens
  - point-of-no-return
  - cvpr2026
pretty_name: MATT-Bench
size_categories:
  - 100K<n<1M
---

# Mistake Attribution: Fine-Grained Mistake Understanding in Egocentric Videos

**CVPR 2026**

[Yayuan Li](https://www.linkedin.com/in/yayuan-li-148659272/)<sup>1</sup>, [Aadit Jain](https://www.linkedin.com/in/jain-aadit/)<sup>1</sup>, [Filippos Bellos](https://www.linkedin.com/in/filippos-bellos-168595156/)<sup>1</sup>, [Jason J. Corso](https://www.linkedin.com/in/jason-corso/)<sup>1,2</sup>

<sup>1</sup>University of Michigan, <sup>2</sup>Voxel51

[[Paper](https://arxiv.org/abs/2511.20525)] [[Code](https://github.com/yayuanli/MATT)] [[Project Page](https://yayuanli.github.io/MATT/)]

---

> **Dataset coming soon.** We are preparing the data for public release. Stay tuned!

## MATT-Bench Overview

MATT-Bench provides two large-scale benchmarks for **Mistake Attribution (MATT)** — a task that goes beyond binary mistake detection to attribute *what* semantic role was violated, *when* the mistake became irreversible (Point-of-No-Return), and *where* the mistake occurred in the frame.

The benchmarks are constructed by **MisEngine**, a data engine that automatically creates mistake samples with attribution-rich annotations from existing egocentric action datasets:

| Dataset | Samples | Instruction Texts | Semantic | Temporal | Spatial |
|---|---|---|---|---|---|
| **Ego4D-M** | 257,584 | 16,099 | ✓ | ✓ | ✓ |
| **EPIC-KITCHENS-M** | 221,094 | 12,283 | ✓ | — | — |

These are at least **two orders of magnitude larger** than any existing mistake dataset.

## Annotations

Each sample consists of an instruction text and an attempt video, annotated with:

- **Semantic Attribution**: Which semantic role (predicate, object) in the instruction is violated in the attempt video
- **Temporal Attribution**: The Point-of-No-Return (PNR) frame where the mistake becomes irreversible (Ego4D-M)
- **Spatial Attribution**: Bounding box localizing the mistake region in the PNR frame (Ego4D-M)

## Citation

```bibtex
@inproceedings{li2026mistakeattribution,
  title     = {Mistake Attribution: Fine-Grained Mistake Understanding in Egocentric Videos},
  author    = {Li, Yayuan and Jain, Aadit and Bellos, Filippos and Corso, Jason J.},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year      = {2026},
}
```