# Hierarchical GRU with Input-Conditioned Slot Queries for Ball Action Anticipation

arXiv: https://arxiv.org/abs/2606.14730 · HTML: https://mlsanity.com/paper/2606.14730/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 3 of its 6 references that are in the map.
- First posted on arXiv 2 Jun 2026.

## Builds on (most cited first, 3 of 3)

- [Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling](https://mlsanity.com/paper/1412.3555/) (2014)
- [EfficientNetV2: Smaller Models and Faster Training](https://mlsanity.com/paper/2104.00298/) (2021)
- [Action Anticipation from SoccerNet Football Video Broadcasts](https://mlsanity.com/paper/2504.12021/) (2025)

## Abstract (arXiv)

We present a hierarchical model for ball action anticipation in football broadcast video. Given a 30-second observation window, the system predicts actions occurring in the subsequent 5-second window across 10 classes. A shared local Transformer encodes clip-level features within each 5-second sub-window; a GRU then aggregates temporal context across all sub-windows; finally, a Transformer decoder with K input-conditioned event slots decodes the anticipation target via three decoupled heads (objectness, class, temporal offset). We introduce frequency-reweighted Hungarian matching that systematically favours rare action classes, and Gaussian soft targets for temporal bin supervision. On the SoccerNet Ball Action Anticipation benchmark, our method achieves 17.91% mAP on the test server.