# Video Question Answering on AGQA (novel compositions)

Source: [Video Question Answering on AGQA (novel compositions)](<https://www.sota2.com/research/sota/video-question-answering-on-agqa-novel-compositions>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T05:37:48.703Z

Evaluation of model performance on the novel compositions subset of the AGQA dataset for Video Question Answering, reporting Binary, Open, and Overall Accuracy.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Video Question Answering |
| Dataset (including any stated split) | AGQA (novel compositions) |
| Metric | Binary Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 5 of 5 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Binary Accuracy](<https://www.sota2.com/research/sota/video-question-answering-on-agqa-novel-compositions.md?metric=Binary%20Accuracy>): Higher is better; unit: PERCENT.

- [Open Accuracy](<https://www.sota2.com/research/sota/video-question-answering-on-agqa-novel-compositions.md?metric=Open%20Accuracy>): Higher is better; unit: PERCENT.

- [Overall Accuracy](<https://www.sota2.com/research/sota/video-question-answering-on-agqa-novel-compositions.md?metric=Overall%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Binary Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| SHG-VQA | 49.27 | Not recorded | 68fa6f4b1e209131cea44d66 |
| SHG-VQA | 46.55 | Not recorded | 68fa6f4b1e209131cea44d66 |
| PSAC | 46.49 | Not recorded | 68fa6f4b1e209131cea44d66 |
| HME | 45.42 | Not recorded | 68fa6f4b1e209131cea44d66 |
| HCRN | 44.88 | Not recorded | 68fa6f4b1e209131cea44d66 |

### SHG-VQA

Binary Accuracy: 49.27. Result ID: 6990b52a3a6c314849029f9c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f4b1e209131cea44d66.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f4b1e209131cea44d66.md>)

[Implementation](<https://github.com/aurooj/SHG-VQA>)

Recorded evaluation configuration: {"training data size":"1.4M"}. This record does not assert independent reproduction.

### SHG-VQA

Binary Accuracy: 46.55. Result ID: 6990b52a3a6c314849029f9b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f4b1e209131cea44d66.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f4b1e209131cea44d66.md>)

[Implementation](<https://github.com/aurooj/SHG-VQA>)

Recorded evaluation configuration: {"training data size":"100K"}. This record does not assert independent reproduction.

### PSAC

Binary Accuracy: 46.49. Result ID: 6990b52a3a6c314849029f98.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f4b1e209131cea44d66.md>)

Recorded evaluation configuration: {"training data size":"1.6M"}. This record does not assert independent reproduction.

### HME

Binary Accuracy: 45.42. Result ID: 6990b52a3a6c314849029f99.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f4b1e209131cea44d66.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65a31e209131cea0b77c.md>)

[Implementation](<https://github.com/fanchenyou/HME-VideoQA>)

Recorded evaluation configuration: {"training data size":"1.6M"}. This record does not assert independent reproduction.

### HCRN

Binary Accuracy: 44.88. Result ID: 6990b52a3a6c314849029f9a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f4b1e209131cea44d66.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67351e209131cea14bfd.md>)

[Implementation](<https://github.com/thaolmk54/hcrn-videoqa>)

Recorded evaluation configuration: {"training data size":"1.6M"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.