TSLANet

TSLANet stands for Time Series Lightweight Adaptive Network. It combines convolutional operations and spectral analysis, using an Adaptive Spectral Block to capture long- and short-term temporal relationships while reducing noise through adaptive thresholding.

It is not a general chat model; it is a model line for continuous time data in sensors, finance, industry, healthcare, and related domains.

Time-series data is important but less visible in AI applications. Many enterprise and public-sector signals are not text or images, but continuous measurements: load, prices, electricity usage, patient vitals, and equipment state.

TSLANet represents the extension of the "foundation model" idea into non-text data.

SUTD’s strength sits at the intersection of engineering, design, and systems. Projects such as TSLANet fit Singapore’s real industrial scenarios: urban infrastructure, industrial systems, health monitoring, and financial risk.

This page can later add concrete benchmarks, datasets, and whether local industry applications emerge.