arXiv cs.AI· Melanie Schaller, Nick Janssen, Bodo Rosenhahn·· 3 天前AI 评分22
Naga：受吠陀数学启发的深度状态空间模型编码方法
Naga: Vedic Encoding for Deep State Space Models
AI 导读
研究者提出 Naga，一种受吠陀数学结构概念启发的深度状态空间模型（SSM）编码方法，通过双向处理正向与时间反转序列并经 Hadamard 逐元素交互，增强对远距离时间步依赖的捕捉能力。在 ETTh1、ETTh2、ETTm1、ETTm2、Weather、Traffic、ILI 等长期时间序列预测基准上，Naga 优于 28 个当前 SOTA 模型，且效率较现有深度 SSM 方法更优。
正文 · 原文
Abstract:This paper presents Naga, a deep State Space Model (SSM) encoding approach inspired by structural concepts from Vedic mathematics. The proposed method introduces a bidirectional representation for time series by jointly processing forward and time-reversed input sequences. These representations are then combined through an element-wise (Hadamard) interaction, resulting in a Vedic-inspired encoding that enhances the model's ability to capture temporal dependencies across distant time steps. We evaluate Naga on multiple long-term time series forecasting (LTSF) benchmarks, including ETTh1, ETTh2, ETTm1, ETTm2, Weather, Traffic, and ILI. The experimental results show that Naga outperforms 28 current state of the art models and demonstrates improved efficiency compared to existing deep SSM-based approaches. The findings suggest that incorporating structured, Vedic-inspired decomposition can provide an interpretable and computationally efficient alternative for long-range sequence modeling.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Systems and Control (eess.SY) |
| Cite as: | arXiv:2511.13510 [cs.LG] |
| (or arXiv:2511.13510v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2511.13510 arXiv-issued DOI via DataCite |
Submission history
From: Melanie Schaller Dr. [view email]
[v1]
Mon, 17 Nov 2025 15:43:49 UTC (436 KB)
[v2]
Mon, 12 Oct 2026 15:12:47 UTC (429 KB)
来源：arXiv cs.AI · arxiv.org