# Data-Local Autonomous LLM-Guided Neural Architecture Search for Multiclass Multimodal Time-Series Classification

arXiv: https://arxiv.org/abs/2603.15939 · HTML: https://mlsanity.com/paper/2603.15939/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 14 of its 32 references that are in the map.
- First posted on arXiv 16 Mar 2026.

## Builds on (most cited first, 14 of 14)

- [Neural Architecture Search with Reinforcement Learning](https://mlsanity.com/paper/1611.01578/) (2016)
- [Multimodal Machine Learning: A Survey and Taxonomy](https://mlsanity.com/paper/1705.09406/) (2017)
- [AutoGluon-Tabular: Robust and Accurate AutoML for Structured Data](https://mlsanity.com/paper/2003.06505/) (2020)
- [A Comprehensive Survey of Neural Architecture Search: Challenges and Solutions](https://mlsanity.com/paper/2006.02903/) (2020)
- [Auto-Sklearn 2.0: Hands-free AutoML via Meta-Learning](https://mlsanity.com/paper/2007.04074/) (2020)
- [Benchmarking Multivariate Time Series Classification Algorithms](https://mlsanity.com/paper/2007.13156/) (2020)
- [AutoML-Agent: A Multi-Agent LLM Framework for Full-Pipeline AutoML](https://mlsanity.com/paper/2410.02958/) (2024)
- [AutoML to Date and Beyond: Challenges and Opportunities](https://mlsanity.com/paper/2010.10777/) (2020)
- [A deep learning architecture for temporal sleep stage classification using multivariate and multimodal time series](https://mlsanity.com/paper/1707.03321/) (2017)
- [Improving Position Encoding of Transformers for Multivariate Time Series Classification](https://mlsanity.com/paper/2305.16642/) (2023)
- [AutoML in the Age of Large Language Models: Current Challenges, Future Opportunities and Risks](https://mlsanity.com/paper/2306.08107/) (2023)
- [Multi-Modal Deep Learning for Credit Rating Prediction Using Text and Numerical Data Streams](https://mlsanity.com/paper/2304.10740/) (2023)
- [Look Into the LITE in Deep Learning for Time Series Classification](https://mlsanity.com/paper/2409.02869/) (2024)
- [Multi-Stage Prototype Learning for Interpretable Time Series Classification](https://mlsanity.com/paper/2106.09636/) (2021)

## Built on by (most cited first, 1 of 1)

- [Staging by the Book: Automatic Sleep Stage Classification Using Scoring Rules](https://mlsanity.com/paper/2605.22859/) (2026)

## Abstract (arXiv)

Applying machine learning to sensitive time-series data is often bottlenecked by the iteration loop: Performance depends strongly on preprocessing and architecture, yet training often has to run on-premise under strict data-local constraints. This is a common problem in healthcare and other privacy-constrained domains (e.g., a hospital developing deep learning models on patient EEG). This bottleneck is particularly challenging in multimodal fusion, where sensor modalities must be individually preprocessed and then combined. LLM-guided neural architecture search (NAS) can automate this exploration, but most existing workflows assume cloud execution or access to data-derived artifacts that cannot be exposed. We present a novel data-local, LLM-guided search framework that handles candidate pipelines remotely while executing all training and evaluation locally under a fixed protocol. The controller observes only trial-level summaries, such as pipeline descriptors, metrics, learning-curve statistics, and failure logs, without ever accessing raw samples or intermediate feature representations. Our framework targets multiclass, multimodal learning via one-vs-rest binary experts per class and modality, a lightweight fusion MLP, and joint search over expert architectures and modality-specific preprocessing. We evaluate our method on two regimes: UEA30 (public multivariate time-series classification dataset) and SleepEDFx sleep staging (heterogeneous clinical modalities such as EEG, EOG, and EMG). The results show that the modular baseline model is strong, and the LLM-guided NAS further improves it. Notably, our method finds models that perform within published ranges across most benchmark datasets. Across both settings, our method reduces manual intervention by enabling unattended architecture search while keeping sensitive data on-premise.