2026-10-28 ワシントン大学（UW）
＜関連情報＞
- https://www.washington.edu/news/2026/10/28/october-research-highlights-cellular-organization-downsized-llms-migrant-fishers-and-burrowing-shrimp/
- https://arxiv.org/abs/2604.19642
- https://homes.cs.washington.edu/~gshyam/Papers/ulm.pdf
マイクロ言語モデルにより、即時応答が可能になる Micro Language Models Enable Instant Responses
Wen Cheng, Tuochao Chen, Karim Helwani, Sriram Srinivasan, Luke Zettlemoyer, Shyamnath Gollakota
arXiv Submitted on 21 Apr 2026
DOI:https://doi.org/10.48550/arXiv.2604.19642
Abstract
Edge devices such as smartwatches and smart glasses cannot continuously run even the smallest 100M-1B parameter language models due to power and compute constraints, yet cloud inference introduces multi-second latencies that break the illusion of a responsive assistant. We introduce micro language models (μLMs): ultra-compact models (8M-30M parameters) that instantly generate the first 4-8 words of a contextually grounded response on-device, while a cloud model completes it; thus, masking the cloud latency. We show that useful language generation survives at this extreme scale with our models matching several 70M-256M-class existing models. We design a collaborative generation framework that reframes the cloud model as a continuator rather than a respondent, achieving seamless mid-sentence handoffs and structured graceful recovery via three error correction methods when the local opener goes wrong. Empirical results show that μLMs can initiate responses that larger models complete seamlessly, demonstrating that orders-of-magnitude asymmetric collaboration is achievable and unlocking responsive AI for extremely resource-constrained devices. The model checkpoint and demo are available at this https URL.
