# STAP: A Shuffle-Tokenized App Predictor with Ultra Long Context for Vocabulary-Free Mobile App Prediction

arXiv: https://arxiv.org/abs/2605.29863 · HTML: https://mlsanity.com/paper/2605.29863/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 10 of its 23 references that are in the map.
- First posted on arXiv 28 May 2026.

## Builds on (most cited first, 10 of 10)

- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [GLU Variants Improve Transformer](https://mlsanity.com/paper/2002.05202/) (2020)
- [Root Mean Square Layer Normalization](https://mlsanity.com/paper/1910.07467/) (2019)
- [On Layer Normalization in the Transformer Architecture](https://mlsanity.com/paper/2002.04745/) (2020)
- [Smartphone App Usage Prediction Using Points of Interest](https://mlsanity.com/paper/1711.09337/) (2017)
- [Context-Aware Target Apps Selection and Recommendation for Enhancing Personal Mobile Assistants](https://mlsanity.com/paper/2101.03394/) (2021)
- [CoSEM: Contextual and Semantic Embedding for App Usage Prediction](https://mlsanity.com/paper/2108.11561/) (2021)
- [On the Feature Discovery for App Usage Prediction in Smartphones](https://mlsanity.com/paper/1309.7982/) (2013)
- [TGT: A Temporal Gating Transformer for Smartphone App Usage Prediction](https://mlsanity.com/paper/2502.16957/) (2025)

## Abstract (arXiv)

Predicting the next mobile application a user will launch is essential for intelligent device resource management and proactive assistance. Existing models rely on fixed app vocabularies, which prevents them from generalizing across different app ecosystems. Many also depend on user-specific knowledge, which complicates deployment in cold start scenarios. We propose STAP, a Transformer-based model that eliminates the need for a fixed vocabulary. STAP replaces true app identities with randomly reassigned virtual indices via a shuffle mechanism, and compensates for discarded semantic information by processing behavioral sequences with an ultra-long context design. A theoretical analysis shows that, given a sufficiently long context, the predicted distribution converges to the correct one despite the anonymity of the mapping. Experiments on two datasets from different continents demonstrate that STAP achieves strong cross-dataset zero-shot prediction accuracy -- a setting where all existing fixed-vocabulary methods are inherently inapplicable -- while its cold start performance within each dataset remains competitive with leading models. Furthermore, we introduce a deployment strategy that enables the model to retain a sufficiently long context during continuous inference while keeping latency within acceptable bounds.