{"slug": "why-every-weight-in-a-neural-network-is-born-divided-by-the-square-root-of-n", "title": "Why Every Weight in a Neural Network Is Born Divided by the Square Root of n.", "summary": "Neural network weights are initialized by dividing by the square root of the number of inputs to prevent the network from being dead before training begins. This technique, known as Xavier initialization, ensures stable signal propagation through deep networks.", "body_md": "Before a network learns anything — before it sees a single example — it can already be dead.\nContinue reading on Towards AI »", "url": "https://wpnews.pro/news/why-every-weight-in-a-neural-network-is-born-divided-by-the-square-root-of-n", "canonical_source": "https://pub.towardsai.net/why-every-weight-in-a-neural-network-is-born-divided-by-the-square-root-of-n-d09dde79eb4c?source=rss----98111c9905da---4", "published_at": "2026-06-26 04:04:30+00:00", "updated_at": "2026-06-26 04:11:32.784652+00:00", "lang": "en", "topics": ["neural-networks", "machine-learning", "artificial-intelligence"], "entities": ["Xavier initialization"], "alternates": {"html": "https://wpnews.pro/news/why-every-weight-in-a-neural-network-is-born-divided-by-the-square-root-of-n", "markdown": "https://wpnews.pro/news/why-every-weight-in-a-neural-network-is-born-divided-by-the-square-root-of-n.md", "text": "https://wpnews.pro/news/why-every-weight-in-a-neural-network-is-born-divided-by-the-square-root-of-n.txt", "jsonld": "https://wpnews.pro/news/why-every-weight-in-a-neural-network-is-born-divided-by-the-square-root-of-n.jsonld"}}