Chain-of-Thought Enhanced Shallow Transformers for Wireless Symbol Detection
Li Fan, Peng Wang, Jing Yang, Cong Shen

TL;DR
This paper introduces CHOOSE, a shallow Transformer framework with autoregressive reasoning that enhances wireless symbol detection, achieving deep-model performance with significantly reduced computational and storage requirements.
Contribution
The paper proposes a novel CoT-enhanced shallow Transformer architecture for wireless detection, enabling high performance with minimal layers and resource efficiency.
Findings
Outperforms conventional shallow Transformers in detection accuracy.
Achieves performance comparable to deep Transformers.
Maintains low computational and storage costs.
Abstract
Transformers have shown potential in solving wireless communication problems, particularly via in-context learning (ICL), where models adapt to new tasks through prompts without requiring model updates. However, prior ICL-based Transformer models rely on deep architectures with many layers to achieve satisfactory performance, resulting in substantial storage and computational costs. In this work, we propose CHain Of thOught Symbol dEtection (CHOOSE), a CoT-enhanced shallow Transformer framework for wireless symbol detection. By introducing autoregressive latent reasoning steps within the hidden space, CHOOSE significantly improves the reasoning capacity of shallow models (1-2 layers) without increasing model depth. This design enables lightweight Transformers to achieve detection performance comparable to much deeper models, making them well-suited for deployment on resource-constrained…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Memory and Neural Computing · Quantum-Dot Cellular Automata · Wireless Signal Modulation Classification
MethodsLayer Normalization · Dropout · Absolute Position Encodings · Dense Connections · Byte Pair Encoding · Softmax · Label Smoothing · Transformer
