Abstract
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News – Researchers detail new data in artific ial intelligence. According to news originating from Bridgeport, Connecticut, by NewsRx correspondents, research stated, “In recent years, transformer-based mod els have gained prominence in multivariate long-term time series forecasting (LT SF), demonstrating significant advancements despite facing challenges such as hi gh computational demands, difficulty in capturing temporal dynamics, and managin g long-term dependencies.” Our news correspondents obtained a quote from the research from University of Br idgeport: “The emergence of LTSF-Linear, with its straightforward linear archite cture, has notably outperformed transformerbased counterparts, prompting a reev aluation of the transformer’s utility in time series forecasting. In response, t his paper presents an adaptation of a recent architecture, termed extended LSTM (xLSTM), for LTSF. xLSTM incorporates exponential gating and a revised memory st ructure with higher capacity that has good potential for LTSF. Our adopted archi tecture for LTSF, termed xLSTMTime, surpasses current approaches. We compare xLS TMTime’s performance against various state-of-the-art models across multiple rea l-world datasets, demonstrating superior forecasting capabilities.”
NETL
NSTL