Aoming Liu

Aoming Liu

Auto-generated profile, based on publicly available information

Aoming Liu

OPPO US Research Center

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications

14 October 2024 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
FZFeichi Zhong
Feichi Zhong
ZSZihan Shan
Zihan Shan
GSGangquan Si
Gangquan Si
ALAoming Liu
Aoming Liu
GZGerui Zhao
Gerui Zhao
BLBo Li
Bo Li
Non-intrusive load monitoring (NILM) benefits the planning of energy consumption and time-of-use pricing through disaggregating appliance-level electrical information. However, its widespread adoption and rapid application face significant restrictions and challenges. Variations in energy consumption backgrounds, like user habits and appliance brands, result in substantial distribution disparities in load data, which significantly deteriorate the performance of trained models when applied to new scenarios. Moreover, concerns regarding user privacy and costs further impede the collection of load data when transfer training for adaptability is necessary. To address these issues, we propose a source-free domain adaptation (SFDA) method for NILM to enhance the generalization performance under conditions of severely limited data acquisition. We design a self-supervised subnetwork based on a sequence masking-restoration task to learn domain-invariant features of appliances without the utilization of source-domain dataset and target-domain label data. Furthermore, the entropy minimization and Representation Subspace Distance (RSD) are introduced to align the feature spaces of different domains and mitigate the feature scaling effect on model performance. The cross-house and a cross-dataset adaptation experiment are conducted on four publicly available datasets. The proposed method achieves an average 6.6% improvement in MAE and 7.1% in F1 score over the baseline and performs well compared to other state-of-the-art models using additional training data, which proves the great potential of the proposed method to enhance the generalization with data restrictions.

5 August 2024 in arXiv
Verified
Preprint
Open access
PDF file
ALAoming Liu
Aoming Liu
ZLZhong Li
Zhong Li
ZCZhang Chen
Zhang Chen
NLNannan Li
Nannan Li
YXYi Xu
Yi Xu
BPBryan A. Plummer
Bryan A. Plummer
Immersive scene generation, notably panorama creation, benefits significantly from the adaptation of large pre-trained text-to-image (T2I) models for multi-view image generation. Due to the high cost of acquiring multi-view images, tuning-free generation is preferred. However, existing methods are either limited to simple correspondences or require extensive fine-tuning to capture complex ones. We present PanoFree, a novel method for tuning-free multi-view image generation that supports an extensive array of correspondences. PanoFree sequentially generates multi-view images using iterative warping and inpainting, addressing the key issues of inconsistency and artifacts from error accumulation without the need for fine-tuning. It improves error accumulation by enhancing cross-view awareness and refines the warping and inpainting processes via cross-view guidance, risky area estimation and erasing, and symmetric bidirectional guided generation for loop closure, alongside guidance-based semantic and density control for scene structure preservation. In experiments on Planar, 360°, and Full Spherical Panoramas, PanoFree demonstrates significant error reduction, improves global consistency, and boosts image quality without extra fine-tuning. Compared to existing methods, PanoFree is up to 5x more efficient in time and 3x more efficient in GPU memory usage, and maintains superior diversity of results (2x better in our user study). PanoFree offers a viable alternative to costly fine-tuning or the use of additional pre-trained models. Project website at https://panofree.github.io/.