# Large Language Models Can Perform Automatic Modulation Classification via Discretized Self-supervised Candidate Retrieval

arXiv: https://arxiv.org/abs/2510.00316 · HTML: https://mlsanity.com/paper/2510.00316/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 16 of its 28 references that are in the map.
- First posted on arXiv 30 Sep 2025.

## Builds on (most cited first, 16 of 16)

- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities](https://mlsanity.com/paper/2507.06261/) (2025)
- [What Makes Good In-Context Examples for GPT-3?](https://mlsanity.com/paper/2101.06804/) (2021)
- [GLM-4.5V and GLM-4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning](https://mlsanity.com/paper/2507.01006/) (2025)
- [The Faiss library](https://mlsanity.com/paper/2401.08281/) (2024)
- [Over the Air Deep Learning Based Radio Signal Classification](https://mlsanity.com/paper/1712.04578/) (2017)
- [Deeper Insights Without Updates: The Power of In-Context Learning Over Fine-Tuning](https://mlsanity.com/paper/2410.04691/) (2024)
- [Auto-ICL: In-Context Learning without Human Supervision](https://mlsanity.com/paper/2311.09263/) (2023)
- [Deep Learning Based Automatic Modulation Recognition: Models, Datasets, and Challenges](https://mlsanity.com/paper/2207.09647/) (2022)
- [Large Language Models in Wireless Application Design: In-Context Learning-enhanced Automatic Network Intrusion Detection](https://mlsanity.com/paper/2405.11002/) (2024)
- [Sample Efficient Demonstration Selection for In-Context Learning](https://mlsanity.com/paper/2506.08607/) (2025)
- [Towards a Wireless Physical-Layer Foundation Model: Challenges and Strategies](https://mlsanity.com/paper/2403.12065/) (2024)
- [NMformer: A Transformer for Noisy Modulation Classification in Wireless Communication](https://mlsanity.com/paper/2411.02428/) (2024)
- [DenoMAE: A Multimodal Autoencoder for Denoising Modulation Signals](https://mlsanity.com/paper/2501.11538/) (2025)
- [DenoMAE2.0: Improving Denoising Masked Autoencoders by Classifying Local Patches](https://mlsanity.com/paper/2502.18202/) (2025)
- [Plug-and-Play AMC: Context Is King in Training-Free, Open-Set Modulation with LLMs](https://mlsanity.com/paper/2505.03112/) (2025)

## Abstract (arXiv)

Identifying wireless modulation schemes is essential for cognitive radio, but standard supervised models often degrade under distribution shift, and training domain-specific wireless foundation models from scratch is computationally prohibitive. Large Language Models (LLMs) offer a promising training-free alternative via in-context learning, yet feeding raw floating-point signal statistics into LLMs overwhelms models with numerical noise and exhausts token budgets. We introduce DiSC-AMC, a framework that reformulates Automatic Modulation Classification (AMC) as an LLM reasoning task by combining aggressive feature discretization with nearest-neighbor retrieval over self-supervised embeddings. By mapping continuous features to coarse symbolic tokens, DiSC-AMC aligns abstract signal patterns with LLM reasoning capabilities and reduces prompt length by over 50%. Simultaneously, utilizing a DINOv2 visual encoder to retrieve the kNN most similar labeled exemplars provides highly relevant, query-specific context rather than generic class averages. On a 10-class benchmark, a fine-tuned 7B-parameter LLM using DiSC-AMC achieves 83.0% in-distribution accuracy (-10,to,+10,dB) and 82.50% out-of-distribution (OOD) accuracy (-11,to,-15,dB), outperforming supervised baselines. Comprehensive ablations on vanilla LLMs demonstrate the token efficiency of DiSC-AMC. A training-free 7B LLM achieves 71% accuracy using only 0.5,K-token prompt,surpassing a 200B-parameter baseline that relies on a 2.9K-token prompt. Furthermore, similarity-based exemplar retrieval outperforms naive class-average selection by over 20%. Finally, we identify a fundamental limitation of this pipeline. At extreme OOD noise levels (-30,dB), the underlying self-supervised representations collapse, degrading retrieval quality and reducing classification to random chance.