# Device-First Feedback: Toward Mobile-Native LLM-Driven Neural Architecture Search

arXiv: https://arxiv.org/abs/2608.00078 · HTML: https://mlsanity.com/paper/2608.00078/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 16 references that are in the map.
- First posted on arXiv 29 Jul 2026.

## Builds on (most cited first, 8 of 8)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications](https://mlsanity.com/paper/1704.04861/) (2017)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference](https://mlsanity.com/paper/1712.05877/) (2017)
- [MnasNet: Platform-Aware Neural Architecture Search for Mobile](https://mlsanity.com/paper/1807.11626/) (2018)
- [ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware](https://mlsanity.com/paper/1812.00332/) (2018)
- [MLPerf Inference Benchmark](https://mlsanity.com/paper/1911.02549/) (2019)
- [NNGPT: Rethinking AutoML with Large Language Models](https://mlsanity.com/paper/2511.20333/) (2025)

## Abstract (arXiv)

Deploying convolutional neural networks generated by large language models (LLMs) on real mobile hardware requires more than GPU validation accuracy: INT8 TensorFlow Lite export, delegate selection, and on-device latency jointly determine whether a model is usable. We present an automated mobile deployment pipeline that closes the loop from QLoRA fine-tuning of an architecture-generating LLM through GPU evaluation, INT8 export, and physical-device benchmarking to gated augmentation of the training corpus. The pipeline is fully scripted and runs cycle-by-cycle without manual intervention, with resume support after interruptions. We evaluate the same frozen protocol on two benchmarks, CIFAR-10 and CIFAR-100, on a Samsung SM-P613 tablet (seed 42, 20 models per cycle, cycles 0-6). On CIFAR-10, cycle 1 is gate-accepted and improves the mobile deployment score approximately 25.6x over the baseline with a mean quantized accuracy of 46.9%; later cycles raise GPU accuracy but fail the non-decreasing mobile gate. On CIFAR-100, the pre-QLoRA baseline retains the best mobile score; iterative rounds improve GPU accuracy (up to 26.2%) yet cannot surpass cycle 0 on-device, and the training pool stalls at 19 examples after the first accepted round. Together, the two studies show that closed-loop GPU fine-tuning does not guarantee monotonic mobile gains, especially on harder classification tasks, and that multi-dataset, on-device measurement is needed to stress-test deployment objectives. We release per-cycle metrics with 95% confidence intervals, all figures, and complete reproduction commands.