aiqbench analysis
This research introduces AOSNet, a novel machine learning framework that addresses non-rigid periodicity in time series data by adaptively aligning local oscillatory states rather than relying on fixed templates. By leveraging Hilbert-guided descriptors and learnable priors, the model achieves high accuracy with low latency for critical applications like cloud capacity planning and autoscaling. The approach offers a significant advancement over traditional spectral methods that fail when cycle magnitude or duration varies dynamically.
Adaptive Oscillatory-State Alignment for Time Series Forecasting
arXiv:2606.06010v2 Announce Type: replace Abstract: Long-term time series forecasting benefits from inductive biases that expose recurring temporal structure. Existing periodic forecasting methods typically model recurrence through predefined periods, global spectral components, or fixed learnable templates. However, real-world temporal dynamics are rarely rigidly periodic: around a nominal cycle, oscillatory behavior often exhibits \emph{non-rigid periodicity} (NRP), where cycle magnitude, cycle alignment, and local cycle duration vary over time. Under these conditions, fixed-template periodic modeling can become fundamentally mismatched to the underlying temporal states. We propose AOSNet, a Hilbert-guided forecasting framework that reformulates periodic forecasting from fixed template matching to adaptive oscillatory-state alignment. AOSNet extracts analytic-signal descriptors from both the observed sequence and a learnable global oscillatory prior, then adaptively aligns local states through a descriptor-conditioned gate that selectively preserves reliable observations while softly correcting mismatched regions. The learned prior serves not as a rigid repeated template but as a flexible oscillatory reference interpreted through local state dynamics. Experiments on eight public benchmarks and two cloud workload traces demonstrate leading or highly competitive accuracy with a compact model size and low inference latency, supporting repeated forecasting settings such as capacity planning and autoscaling. Controlled synthetic studies that isolate cycle-magnitude and cycle-alignment variation and combine them with cycle-duration changes show that the advantage of oscillatory-state alignment increases as NRP intensifies.
Abstract:Long-term time series forecasting benefits from inductive biases that expose recurring temporal structure. Existing periodic forecasting methods typically model recurrence through predefined periods, global spectral components, or fixed learnable templates. However, real-world temporal dynamics are rarely rigidly periodic: around a nominal cycle, oscillatory behavior often exhibits \emph{non-rigid periodicity} (NRP), where cycle magnitude, cycle alignment, and local cycle duration vary over time. Under these conditions, fixed-template periodic modeling can become fundamentally mismatched to the underlying temporal states. We propose AOSNet, a Hilbert-guided forecasting framework that reformulates periodic forecasting from fixed template matching to adaptive oscillatory-state alignment. AOSNet extracts analytic-signal descriptors from both the observed sequence and a learnable global oscillatory prior, then adaptively aligns local states through a descriptor-conditioned gate that selectively preserves reliable observations while softly correcting mismatched regions. The learned prior serves not as a rigid repeated template but as a flexible oscillatory reference interpreted through local state dynamics. Experiments on eight public benchmarks and two cloud workload traces demonstrate leading or highly competitive accuracy with a compact model size and low inference latency, supporting repeated forecasting settings such as capacity planning and autoscaling. Controlled synthetic studies that isolate cycle-magnitude and cycle-alignment variation and combine them with cycle-duration changes show that the advantage of oscillatory-state alignment increases as NRP intensifies.
Machine Learning (cs.LG); Databases (cs.DB)
arXiv:2606.06010 [cs.LG]
(or arXiv:2606.06010v2 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2606.06010 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 305 words
Related coverage
NeST: Neighborhood-aware semantic alignment and temporal modulation for LLM based time series forecasting
arXiv · Aug 25, 2026
Adapting LLMs to Time Series Forecasting via Temporal Heterogeneity Modeling and Representation Alignment
arXiv · Aug 18, 2026
InA-Probe: Instruction-Aware Active Probing for Time Series Forecasting with LLMs
arXiv · Jun 9, 2026
DiTS: Multimodal Diffusion Transformers Are Time Series Forecasters
arXiv · Oct 8, 2026