mlsanityData to 14 Oct 2026

Ask for Any Appliance: A Prompt-Programmable Foundation Model for Non-Intrusive Load Monitoring

arXiv 2609.23146 ↗First posted 19 Sep 2026cs.LG eess.SPcited by 0

Not on the map. Why?

Abstract (arXiv)

Non-intrusive load monitoring (NILM) estimates appliance-level consumption from a whole-home meter, but appliance-specific models and fixed output inventories make coverage costly to extend. We present FM4NILM (Foundation Model for NILM), a single prompt-programmable model that estimates a requested appliance's power trajectory from aggregate measurements, a natural-language description, and optional activation exemplars. A lightweight cadence-aware transformer is pretrained by masked reconstruction on 645k sequences from seven public corpora spanning 1-60 s sampling intervals, then aligned with appliance requests using observation-masked losses for partially labeled households. A Bernoulli-lognormal decoder separates activity detection from conditional power estimation. On held-out households and time periods from REDD, UK-DALE, and REFIT, one frozen text-prompted model serves twelve appliance-corpus requests, achieving 0.556 event F1, 0.625 AUPRC, and the lowest active-window MAE (251.8 W) among seven appliance-specific baselines. Streaming score aggregation raises event F1 to 0.582 with a 60 s aggregation delay. In a separate category-held-out evaluation, adding ten activation exemplars raises microwave AUPRC from 0.132 to 0.214 without parameter updates. Input-intervention ablations probe the model's dependence on appliance requests and aggregate measurements. These results demonstrate competitive disaggregation with one shared model and support extending appliance coverage through prompts and examples rather than additional specialist networks.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 16 of its 40 references that are in the map. First posted on arXiv 19 Sep 2026.

Builds on

16 of 40 references.

References · 16

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  2. 2016XGBoost: A Scalable Tree Boosting Systemcited by 5,166
  3. 2023Large Language Models Are Zero-Shot Time Series Forecasterscited by 500
  4. 2024Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Expertscited by 287
  5. 2015Neural NILM: Deep Neural Networks Applied to Energy Disaggregationcited by 54
  6. 2018Subtask Gated Networks for Non-Intrusive Load Monitoringcited by 14
  7. 2022SoundBeam: Target sound extraction conditioned on sound-class labels and enrollment clues for increased performance and continuous learningcited by 14
  8. 2024TLINet: Differentiable Neural Network Temporal Logic Inferencecited by 9
  9. 2025Detect Any Sound: Open-Vocabulary Sound Event Detection with Multi-Modal Queriescited by 9
  10. 2023MSDC: Exploiting Multi-State Power Consumption in Non-intrusive Load Monitoring based on A Dual-CNN Modelcited by 4
  11. 2022Efficient Localness Transformer for Smart Sensor-Based Energy Disaggregationcited by 3
  12. 2025NILMFormer: Non-Intrusive Load Monitoring that Accounts for Non-Stationaritycited by 3
4 more
  1. 2025Prompting Large Language Models for Training-Free Non-Intrusive Load Monitoringcited by 1
  2. 2025Non-Intrusive Load Monitoring Based on Image Load Signatures and Continual Learningcited by 1
  3. 2025Energy Injection Identification enabled Disaggregation with Deep Multi-Task Learningcited by 1
  4. 2026Lightweight and Scalable Transfer Learning Framework for Load Disaggregationcited by 1

Built on by

Not cited yet.