Presents SE-WaveNet with weight-tied dilated convolutions plus wavelet and spectral components that reproduces empirical scaling collapse on financial returns while using L times fewer convolutional parameters.
Informer: Beyond efficient transformer for long sequence time-series forecasting
6 Pith papers cite this work. Polarity classification is still indexing.
6
Pith papers citing it
fields
cs.LG 6years
2026 6representative citing papers
CastFlow introduces a role-specialized agentic workflow with memory retrieval and multi-view toolkit for iterative ensemble time series forecasting, using two-stage SFT+RLVR training on a domain-specific LLM to outperform static baselines.
ROAM freezes specialist models and uses LLM priors plus online evidence in a 5-D semantic latent space to cut major-shift MAE by over 20% with under 0.02 ms overhead.
NoRIN replaces affine reversible normalization with a Johnson S_U non-linear transform whose shape parameters are initialized by quantile fitting and refined by Bayesian optimization on validation data, yielding backbone-dependent optima that differ from the linear limit.
FTimeXer improves power-grid carbon intensity forecasts by combining an FFT frequency branch with gated fusion and stochastic exogenous masking plus consistency regularization, showing gains on three real datasets.
TSF converts process variable documents into frozen semantic vectors that scale the numerical input window before a time-series backbone, yielding average MAE reductions of 2.9–3.6% across industrial forecasting tasks.
citing papers explorer
-
Scale-Equivariant Generative Forecasting: Weight-Tied Dilated Convolutions, Wavelet Scattering Inputs, and Spectral-Consistency Training for Self-Similar Time Series
Presents SE-WaveNet with weight-tied dilated convolutions plus wavelet and spectral components that reproduces empirical scaling collapse on financial returns while using L times fewer convolutional parameters.
-
CastFlow: Learning Role-Specialized Agentic Workflows for Time Series Forecasting
CastFlow introduces a role-specialized agentic workflow with memory retrieval and multi-view toolkit for iterative ensemble time series forecasting, using two-stage SFT+RLVR training on a domain-specific LLM to outperform static baselines.
-
Open-Ended Scenario Reasoning for Specialist Model Adaptation
ROAM freezes specialist models and uses LLM priors plus online evidence in a 5-D semantic latent space to cut major-shift MAE by over 20% with under 0.02 ms overhead.
-
NoRIN: Backbone-Adaptive Reversible Normalization for Time-Series Forecasting
NoRIN replaces affine reversible normalization with a Johnson S_U non-linear transform whose shape parameters are initialized by quantile fitting and refined by Bayesian optimization on validation data, yielding backbone-dependent optima that differ from the linear limit.
-
FTimeXer: Frequency-aware Time-series Transformer with Exogenous variables for Robust Carbon Footprint Forecasting
FTimeXer improves power-grid carbon intensity forecasts by combining an FFT frequency branch with gated fusion and stochastic exogenous masking plus consistency regularization, showing gains on three real datasets.
-
LLM-Guided Task-Semantic Field Factorization for Industrial Process Forecasting
TSF converts process variable documents into frozen semantic vectors that scale the numerical input window before a time-series backbone, yielding average MAE reductions of 2.9–3.6% across industrial forecasting tasks.