Episode 38 - AI trained with Historial Data: getting the same Crap but with AI
Listen now
Description
AI is making its entrance into the Construction Industry. But it requires data to be trained on. Construction data. We don't have a lot of that. And the data we have, might not be of the best standardized quality... training an AI on this old data might just give us more of the same - which we don't want. So, how can we solve this? That is today's topic. What you'll learn from today's episode of Inventing Construction: - Why you have to train your own AI - The historical bias of AI - Why Amazon had to stop using an AI - How we tried to train our own AI - What could happen if we train AI badly - Why we in Construction have to develop together We would love for you to continue the conversation with us on LinkedIn: Ross: https://www.linkedin.com/in/ross-griffin-kosmos/ Elia: https://www.linkedin.com/in/eliaarquitectatecnica/ KOSMOS: https://www.linkedin.com/company/kosmos-dk Thanks for listening!
More Episodes
Coding has entered the Construction Industry. And so has the sustainability agenda. In Denmark you have to calculate a Life Cycle Assessment of your project before you can receive a Building Permit. This new regulation has resulted in the creation of several new programs that can help Developers...
Published 11/15/24
Published 11/15/24
How do we reach our goals for 2050? Where do we start? These are the thoughts that started ZERO Construct. A non-profit organization and community of like-minded people, working in construction, all trying to make the Industry more sustainable one project at a time.  Our guest, Matthew Jackson,...
Published 10/15/24