Back to AI News
arxiv

LEMMO: A Large Electromagnetic Model leveraging Pretrained LLMs for Million-Sample Signal Understanding

LEMMO is a large electromagnetic model for instruction-driven recognition and description of long raw IQ recordings. It combines a linear-attention EM encoder with anti-aliased temporal resampling to process up to eight million complex samples, plus an EM-text interaction module that conditions learnable queries on instructions. The work also releases a 1.38-TB corpus of real and synthetic EM signals, including LEMMO-ShortQA and LEMMO-LongDesc. Compared with MERLIN-LoRA, it reports higher mean EM-Bench perception accuracy, strategy-generation ROUGE-L F1, ShortQA accuracy, and LongDesc fact F1 while preserving its language backbone's MMLU performance.

AI Brief

It provides a concrete architecture and large released dataset for applying pretrained LLMs to long raw electromagnetic signal understanding, backed by benchmark comparisons.

arXiv:2611.09476v1 Announce Type: new Abstract: Large language models have expanded beyond text to vision, audio, and other modalities, but understanding raw electromagnetic (EM) signals remains challenging. Wideband recordings can contain millions of samples, whereas existing waveform-based EM models are primarily evaluated on short in-phase and quadrature (IQ) segments, leaving long-recording understanding and detailed description insufficiently explored. We introduce LEMMO, a large EM model for instruction-driven recognition and description of long raw IQ recordings. LEMMO combines a linear-attention EM encoder with anti-aliased temporal resampling to process up to eight million complex samples. Its EM-text interaction module conditions learnable queries on instructions before retrieving signal evidence. Compact queries support recognition, while structured queries with auxiliary factual supervision organize scene-level properties, signal attributes, and cross-signal relationships for detailed description. We also construct and release a unified 1.38-TB corpus of real-world and synthetic EM signals, including LEMMO-ShortQA and LEMMO-LongDesc. Compared with MERLIN-LoRA, LEMMO improves mean EM-Bench perception accuracy by 15.77 %, mean strategy-generation ROUGE-L F1 by 4.38 points, and mean LEMMO-ShortQA accuracy by 14.31 %. It also improves categorical and numerical fact F1 on both LEMMO-LongDesc tasks while preserving its language backbone's MMLU performance. Our code and datasets are available at https://github.com/N1rv4n4/LEMMO and https://huggingface.co/datasets/DaftJun/RFLAS, respectively.
Be the first to rate this article.

Comments & Questions (0)

Captcha
Please be respectful — let's keep the conversation friendly.

No comments yet

Be the first to comment!