# Video Question Answering on AGQA Balanced novel compositions 2.0 (test) Source: [Video Question Answering on AGQA Balanced novel compositions 2.0 (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T06:04:18.059Z Evaluation of Video Question Answering performance on the test split of the AGQA Balanced novel compositions 2.0 dataset, reporting granular accuracies across Sequencing, Superlative, Duration, and Object-relation question types (Binary, Open, and All). 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 Balanced novel compositions 2.0 (test) | | Metric | Sequencing (Binary) Acc | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 4 of 4 | | 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 - [Sequencing (Binary) Acc](): Higher is better; unit: PERCENT. - [Sequencing (Open) Acc](): Higher is better; unit: PERCENT. - [Sequencing (All) Acc](): Higher is better; unit: PERCENT. - [Superlative (Binary) Acc](): Higher is better; unit: PERCENT. - [Superlative (Open) Acc](): Higher is better; unit: PERCENT. - [Superlative (All) Acc](): Higher is better; unit: PERCENT. - [Duration (Binary) Acc](): Higher is better; unit: PERCENT. - [Duration (Open) Acc](): Higher is better; unit: PERCENT. - [Duration (All) Acc](): Higher is better; unit: PERCENT. - [Obj-relation (Binary) Acc](): Higher is better; unit: PERCENT. - [Obj-relation (Open) Acc](): Higher is better; unit: PERCENT. - [Obj-relation (All) Acc](): Higher is better; unit: PERCENT. - [Overall (Binary) Acc](): Higher is better; unit: PERCENT. - [Overall (Open) Acc](): Higher is better; unit: PERCENT. - [Overall (All) Acc](): Higher is better; unit: PERCENT. ## Recorded results | Method | Sequencing (Binary) Acc | Result date | Report paper | | --- | --- | --- | --- | | SHG-VQA | 50.88 | Not recorded | 68fa6f4b1e209131cea44d66 | | HME | 49.33 | Not recorded | 68fa6f4b1e209131cea44d66 | | PSAC | 49.19 | Not recorded | 68fa6f4b1e209131cea44d66 | | HCRN | 48.31 | Not recorded | 68fa6f4b1e209131cea44d66 | ### SHG-VQA Sequencing (Binary) Acc: 50.88. Result ID: 6990b52a3a6c314849029fa5. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"CNN","Resolution":"224x224","Clip length":16}. This record does not assert independent reproduction. ### HME Sequencing (Binary) Acc: 49.33. Result ID: 6990b52a3a6c314849029fa3. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### PSAC Sequencing (Binary) Acc: 49.19. Result ID: 6990b52a3a6c314849029fa2. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### HCRN Sequencing (Binary) Acc: 48.31. Result ID: 6990b52a3a6c314849029fa4. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.