دانلود کتاب معماری یادگیری ماشین در عصر هوش مصنوعی

A clear danger of writing a book about the disciplinary implications of artificial intelligence is a short shelf life: as the galleys arrive at the printing house, the thinking behind the text may already be obsolete. And surely there has been enormous energy invested in the architecture | engineering | construction | operations (AECO) technology space since the first edition of this book was finished in late 2021, and particularly in the market for AI-based applications.1 A central thesis of the first edition was that AECO AI solutions faced considerable challenges of representation, training, precision, knowhow, and risk and reward of their use. Four years later, as this second edition goes to press, many of these barriers remain. This is not to say that there has not been considerable progress in the underlying technology, particularly in widely used language and image generation models. Consider first this 2025 response to the text provocation ‘when computers design buildings, the public will lose out’, first given to ChatGPT-3 (see the Introduction to the first edition) and now to ChatGPT-4o.