# State Space Models are Effective Sign Language Learners: Exploiting Phonological Compositionality for Vocabulary-Scale Recognition

arXiv: https://arxiv.org/abs/2604.08761 · HTML: https://mlsanity.com/paper/2604.08761/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 23 of its 37 references that are in the map.
- First posted on arXiv 9 Apr 2026.

## Builds on (most cited first, 23 of 23)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Semi-Supervised Classification with Graph Convolutional Networks](https://mlsanity.com/paper/1609.02907/) (2016)
- [Graph Attention Networks](https://mlsanity.com/paper/1710.10903/) (2017)
- [Rethinking the Inception Architecture for Computer Vision](https://mlsanity.com/paper/1512.00567/) (2015)
- [Mamba: Linear-Time Sequence Modeling with Selective State Spaces](https://mlsanity.com/paper/2312.00752/) (2023)
- [Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset](https://mlsanity.com/paper/1705.07750/) (2017)
- [Prototypical Networks for Few-shot Learning](https://mlsanity.com/paper/1703.05175/) (2017)
- [Efficiently Modeling Long Sequences with Structured State Spaces](https://mlsanity.com/paper/2111.00396/) (2021)
- [SlowFast Networks for Video Recognition](https://mlsanity.com/paper/1812.03982/) (2018)
- [Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognition](https://mlsanity.com/paper/1801.07455/) (2018)
- [TSM: Temporal Shift Module for Efficient Video Understanding](https://mlsanity.com/paper/1811.08383/) (2018)
- [HiPPO: Recurrent Memory with Optimal Polynomial Projections](https://mlsanity.com/paper/2008.07669/) (2020)
- [MediaPipe: A Framework for Building Perception Pipelines](https://mlsanity.com/paper/1906.08172/) (2019)
- [Measuring Compositional Generalization: A Comprehensive Method on Realistic Data](https://mlsanity.com/paper/1912.09713/) (2019)
- [Sign Language Transformers: Joint End-to-end Sign Language Recognition and Translation](https://mlsanity.com/paper/2003.13830/) (2020)
- [Word-level Deep Sign Language Recognition from Video: A New Large-scale Dataset and Methods Comparison](https://mlsanity.com/paper/1910.11006/) (2019)
- [Skeleton Aware Multi-modal Sign Language Recognition](https://mlsanity.com/paper/2103.08833/) (2021)
- [Sign Language Recognition, Generation, and Translation: An Interdisciplinary Perspective](https://mlsanity.com/paper/1908.08597/) (2019)
- [PYSKL: Towards Good Practices for Skeleton Action Recognition](https://mlsanity.com/paper/2205.09443/) (2022)
- [SignBERT: Pre-Training of Hand-Model-Aware Representation for Sign Language Recognition](https://mlsanity.com/paper/2110.05382/) (2021)
- [ASL Citizen: A Community-Sourced Dataset for Advancing Isolated Sign Language Recognition](https://mlsanity.com/paper/2304.05934/) (2023)
- [Fingerspelling recognition in the wild with iterative visual attention](https://mlsanity.com/paper/1908.10546/) (2019)
- [Dynamic Spatial-Temporal Aggregation for Skeleton-Aware Sign Language Recognition](https://mlsanity.com/paper/2403.12519/) (2024)

## Abstract (arXiv)

Sign language recognition suffers from catastrophic scaling failure: models achieving high accuracy on small vocabularies collapse at realistic sizes. Existing architectures treat signs as atomic visual patterns, learning flat representations that cannot exploit the compositional structure of sign languages-systematically organized from discrete phonological parameters (handshape, location, movement, orientation) reused across the vocabulary. We introduce PHONSSM, enforcing phonological decomposition through anatomically-grounded graph attention, explicit factorization into orthogonal subspaces, and prototypical classification enabling few-shot transfer. Using skeleton data alone on the largest ASL dataset ever assembled (5,565 signs), PHONSSM achieves 72.1% on WLASL2000 (+18.4pp over skeleton SOTA), surpassing most RGB methods without video input. Gains are most dramatic in the few-shot regime (+225% relative), and the model transfers zero-shot to ASL Citizen, exceeding supervised RGB baselines. The vocabulary scaling bottleneck is fundamentally a representation learning problem, solvable through compositional inductive biases mirroring linguistic structure.