{"slug": "fm-llm-a-frequency-enhanced-mixture-of-experts-framework-for-adapting-llms-to", "title": "FM-LLM: A frequency-enhanced mixture-of-experts framework for adapting LLMs to time series forecasting", "summary": "Researchers propose FM-LLM, a frequency-enhanced mixture-of-experts framework that adapts frozen large language models to time-series forecasting without textual prompts, achieving state-of-the-art performance on 59 of 78 evaluation metrics across eleven public benchmarks. Compared to the strongest autoregressive LLM-based baseline, FM-LLM delivers average improvements of 5.3% in MSE and 5.6% in MAE, with maximum gains of 8.0% and 8.4%, respectively, and maintains superior performance in 10% few-shot and zero-shot scenarios.", "body_md": "arXiv:2608.11623v1 Announce Type: new\nAbstract: Recent advances in Large Language Models (LLMs) have spurred cross-modal solutions for time-series forecasting. However, existing methods rely heavily on textual prompts for modality alignment-introducing nontrivial computational overhead and failing to leverage the rich spectral dynamics inherent in time-series data. To enable prompt-free, frequency-aware adaptation of frozen LLMs, we propose FM-LLM (Frequency-Enhanced Mixture-of-Experts for adapting LLMs to Time Series Forecasting), an autoregressive framework grounded in constrained asymmetric coupling. A Fourier Analysis Network (FAN)-based spectral token aligner injects structured harmonic representations directly into the frozen LLM with numerical compatibility. An asymmetric Mixture-of-Experts (MoE) decoder enforces role separation: shared experts with lightweight FAN layers reconstruct the global periodic backbone, while routed experts-restricted to standard FFNs-specialize in modeling non-periodic residual dynamics. A time-frequency hybrid loss function jointly optimizes temporal accuracy and spectral consistency, mitigating error accumulation during long-horizon autoregressive rollouts. Evaluated across eleven public benchmarks, FM-LLM achieves state-of-the-art performance on 59 out of 78 evaluation metrics. Compared to the strongest autoregressive LLM-based baseline, it delivers average improvements of 5.3% in MSE and 5.6% in MAE, with maximum gains reaching 8.0% for MSE and 8.4% for MAE. FM-LLM also demonstrates robust transferability, maintaining superior performance in 10% few-shot and zero-shot forecasting scenarios.", "url": "https://wpnews.pro/news/fm-llm-a-frequency-enhanced-mixture-of-experts-framework-for-adapting-llms-to", "canonical_source": "https://www.machinebrief.com/news/fm-llm-a-frequency-enhanced-mixture-of-experts-framework-for-qld1", "published_at": "2026-08-13 04:00:00+00:00", "updated_at": "2026-08-13 04:41:36.146917+00:00", "lang": "en", "topics": ["large-language-models", "machine-learning", "artificial-intelligence"], "entities": ["FM-LLM", "arXiv"], "alternates": {"html": "https://wpnews.pro/news/fm-llm-a-frequency-enhanced-mixture-of-experts-framework-for-adapting-llms-to", "markdown": "https://wpnews.pro/news/fm-llm-a-frequency-enhanced-mixture-of-experts-framework-for-adapting-llms-to.md", "text": "https://wpnews.pro/news/fm-llm-a-frequency-enhanced-mixture-of-experts-framework-for-adapting-llms-to.txt", "jsonld": "https://wpnews.pro/news/fm-llm-a-frequency-enhanced-mixture-of-experts-framework-for-adapting-llms-to.jsonld"}}