TriHead-GAN proposes a Transformer-based Generative Adversarial Network with a triple-head discriminator to address the scarcity of city-level carbon emission data. This framework improves time series generation by better preserving cross-variable correlations and realistic step-wise variability compared to existing models.
This paper studies asynchronous alignment in multimodal learning, where a dense primary stream must be fused with sporadic external context, requiring models to reason explicitly about freshness and trust. It proposes CGCMA (Conditionally-Gated Cross-Modal Attention), a model that separates text-conditioned grounding from lag-aware trust control, tested on cryptocurrency markets.
This paper introduces COM (Continuity and Ordinality Matter), a strategy that integrates geometric constraints into both the initialization and training stages of token-based time series large language models (TS-LLMs). The research demonstrates that preserving continuity and ordinality in time series token embeddings significantly improves model performance and generalizability.
Liquid Neural Networks (LNNs) model hidden state evolution as a continuous differential equation, addressing the limitations of discrete-time RNNs and LSTMs in capturing fluid temporal dynamics. This paper benchmarks LNNs against LSTMs across four sequential modalities, revealing LNNs' superior parameter efficiency and robustness, especially in native temporal domains and clinical environments.
This paper proposes a novel method called forecast-necessity testing for interpretable causal discovery in nonlinear time-series models. It aims to move beyond traditional coefficients to better understand complex causal relationships.
This content explains how to find seasonality patterns in time series using the Fourier Transform.
This article presents five practical Python scripts designed for time series analysis, a common task in various fields such as finance, operations, engineering, and research. These scripts are intended to cover frequently encountered analytical challenges with time series data.
This paper tackles the Variable Gapped Longest Common Subsequence (VGLCS) problem, a generalization of LCS with flexible gap constraints, relevant to molecular sequence comparison and time-series analysis. It proposes a root-based state graph search framework combined with an iterative beam search strategy to manage combinatorial explosion and find high-quality solutions.