# Conditional Collapse in Sign Language Production: A Diagnostic and a Scaling Argument

arXiv: https://arxiv.org/abs/2606.01643 · HTML: https://mlsanity.com/paper/2606.01643/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 17 of its 40 references that are in the map.
- First posted on arXiv 1 Jun 2026.

## Builds on (most cited first, 17 of 17)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Expressive Body Capture: 3D Hands, Face, and Body from a Single Image](https://mlsanity.com/paper/1904.05866/) (2019)
- [Human Motion Diffusion Model](https://mlsanity.com/paper/2209.14916/) (2022)
- [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)
- [The Role of ImageNet Classes in Fr'echet Inception Distance](https://mlsanity.com/paper/2203.06026/) (2022)
- [How2Sign: A Large-scale Multimodal Dataset for Continuous American Sign Language](https://mlsanity.com/paper/2008.08143/) (2020)
- [Progressive Transformers for End-to-End Sign Language Production](https://mlsanity.com/paper/2004.14874/) (2020)
- [YouTube-ASL: A Large-Scale, Open-Domain American Sign Language-English Parallel Corpus](https://mlsanity.com/paper/2306.15162/) (2023)
- [Neural Sign Actors: A diffusion model for 3D sign language production from text](https://mlsanity.com/paper/2312.02702/) (2023)
- [Sign Language Translation from Instructional Videos](https://mlsanity.com/paper/2304.06371/) (2023)
- [G2P-DDM: Generating Sign Pose Sequence from Gloss Sequence with Discrete Diffusion Model](https://mlsanity.com/paper/2208.09141/) (2022)
- [T2S-GPT: Dynamic Vector Quantization for Autoregressive Sign Language Production from Text](https://mlsanity.com/paper/2406.07119/) (2024)
- [SignAvatar: Sign Language 3D Motion Reconstruction and Generation](https://mlsanity.com/paper/2405.07974/) (2024)
- [Meaningful Pose-Based Sign Language Evaluation](https://mlsanity.com/paper/2510.07453/) (2025)
- [Toward Phonology-Guided Sign Language Motion Generation: A Diffusion Baseline and Conditioning Analysis](https://mlsanity.com/paper/2603.17388/) (2026)

## Abstract (arXiv)

Sign Language Production (SLP) is the task of generating avatar sign language motion from natural language text. The quality of the generated motion is typically evaluated by a motion-space Fr'echet distance (FID) and back-translation (BT) BLEU score on benchmarks such as How2Sign. Both metrics can improve substantially while the underlying generator fails to faithfully represent the sign language gestures. In this work we propose to evaluate the generated motion at three independent levels: (τ1) initial-pose conditioning, (τ2) output diversity, and (τ3) target faithfulness. We compute these as pairwise-distance ratios using latent representations of a frozen motion autoencoder (MoAE). We evaluate 14 SLP model checkpoints on the How2Sign dataset, including a re-implemented Neural Sign Actors (NSA), and show that τ3 faithfulness is never attained, while FID varies by nearly two orders of magnitude and is uncorrelated with faithfulness. We show that on the isolated gloss dataset ASL3DWord favorable τ3 can be attained, hence isolating the size of the sentence-level paired-dataset as the bottleneck.