Amortized Predictability-aware Training Framework for Time Series Forecasting and Classification

2026-02-18 · Xu Zhang, Peng Wang, Yichen Li, Wei Wang arxiv

Time series data are prone to noise in various domains, and training samples may contain low-predictability patterns that deviate from the normal data distribution, leading to training instability or convergence to poor local minima. Therefore, mitigating the adverse effects of low-predictability samples is crucial for time series analysis tasks such as time series forecasting (TSF) and time series classification (TSC). While many deep learning models have achieved promising performance, few consider how to identify and penalize low-predictability samples to improve model performance from the training perspective. To fill this gap, we propose a general Amortized Predictability-aware Training Framework (APTF) for both TSF and TSC. APTF introduces two key designs that enable the model to focus on high-predictability samples while still learning appropriately from low-predictability ones: (i) a Hierarchical Predictability-aware Loss (HPL) that dynamically identifies low-predictability samples and progressively expands their loss penalty as training evolves, and (ii) an amortization model that mitigates predictability estimation errors caused by model bias, further enhancing HPL's effectiveness. The code is available at https://github.com/Meteor-Stars/APTF.

📄 PDF Abstract BibTeX arXiv:2602.16224

Code (0)

등록된 구현이 없습니다.

Tasks

Time Series Classification Time Series Forecasting Time Series Analysis

Uncertainty-Aware Surrogate-based Amortized Bayesian Inference for Computationally Expensive Models

2025-05-13 · Stefania Scheurer, Philipp Reiser, Tim Brünnette, Wolfgang Nowak 외

Bayesian inference typically relies on a large number of model evaluations to estimate posterior distributions. Established methods like Markov Chain Monte Carlo (MCMC) and Amortized Bayesian Inference (ABI) can become c…

Bayesian Inference

Amortized Safe Active Learning for Real-Time Data Acquisition: Pretrained Neural Policies from Simulated Nonparametric Functions

2025-01-26 · Cen-You Li, Marc Toussaint, Barbara Rakitsch, Christoph Zimmer

Safe active learning (AL) is a sequential scheme for learning unknown systems while respecting safety constraints during data acquisition. Existing methods often rely on Gaussian processes (GPs) to model the task and saf…

Active Learning Decision Making Experimental Design Gaussian Processes

Beyond Model Ranking: Predictability-Aligned Evaluation for Time Series Forecasting

2025-09-27 · Wanjin Feng, Yuan Yuan, Jingtao Ding, Yong Li arxiv

In the era of increasingly complex AI models for time series forecasting, progress is often measured by marginal improvements on benchmark leaderboards. However, this approach suffers from a fundamental flaw: standard ev…

Time Series Forecasting

AToM: Amortized Text-to-Mesh using 2D Diffusion

2024-02-01 · Guocheng Qian, Junli Cao, Aliaksandr Siarohin, Yash Kant 외

We introduce Amortized Text-to-Mesh (AToM), a feed-forward text-to-mesh framework optimized across multiple text prompts simultaneously. In contrast to existing text-to-3D methods that often entail time-consuming per-pro…

Text to 3D

U-DAVI: Uncertainty-Aware Diffusion-Prior-Based Amortized Variational Inference for Image Reconstruction

2026-02-12 · Ayush Varshney, Katherine L. Bouman, Berthy T. Feng arxiv

Ill-posed imaging inverse problems remain challenging due to the ambiguity in mapping degraded observations to clean images. Diffusion-based generative priors have recently shown promise, but typically rely on computatio…

Image Reconstruction