With information from trade, academia, and main researchers
On the coronary heart of the “Generative AI revolution” is the Transformer mannequin — launched by Google in 2017.
However each tech revolution creates confusion. In a speedy development setting, it’s tough to evaluate improvements unbiasedly — not to mention estimate their influence.
Transformers, which kickstarted this AI breakthrough, have grow to be a “controversial mannequin”. There are 2 excessive viewpoints:
- Zealous adopters: They use Transformers in all places, together with exterior of NLP. They use them even when they will’t or don’t need to — they’re compelled by their employers, managers, traders, and so on.
- Skeptics and luddites: They criticize AI fashions, together with Transformers. They will’t perceive/settle for that scaling a mannequin with extra information and layers can usually outperform elegant mathematical fashions based mostly on rigorous proofs.
Now that the mud is settling, it’s time for neutral analysis.
This text focuses on Transformers for forecasting. I’ll talk about the newest developments from academia, trade, and main researchers.
I may also clarify how and underneath what circumstances Transformer-based forecasting…
