Grégoire Locqueville on Nostr: A nice, concise definition of the trace, which makes it look a little less ...
A nice, concise definition of the trace, which makes it look a little less arbitrary-looking (if a little more technical) than the standard definition "the sum of elements on the diagonal", and also makes it more evident that it doesn't depend on the choice of coordinates:
The trace is the only linear mapping¹ such that:
trace(|u><v|) = <u|v>
for any vectors |u>, |v>.
(The <|> things are bracket notation: |u> stands for a regular vector (think "column vector"), <u|, for its corresponding dual vector (think "row vector"), |u><v| for the tensor product of |u> and |v>, and <u|v> to their dot product)
This still has the drawback of looking like it depends on a choice of inner product, when it actually doesn't.
¹ From linear mappings from V to V, to the scale field K.
Published at
2025-12-19 20:47:01 UTCEvent JSON
{
"id": "b118fe6d4711a4c7476e09530a6c042792798096281e8edde9c68f15c043af11",
"pubkey": "da2ffb9942d327b65ee66f152d69a427f114a4a42f69a4d9b88ab9cab96339b9",
"created_at": 1766177221,
"kind": 1,
"tags": [
[
"proxy",
"https://mathstodon.xyz/users/glocq/statuses/115748190415998297",
"activitypub"
],
[
"client",
"Mostr",
"31990:6be38f8c63df7dbf84db7ec4a6e6fbbd8d19dca3b980efad18585c46f04b26f9:mostr",
"wss://relay.ditto.pub/"
]
],
"content": "A nice, concise definition of the trace, which makes it look a little less arbitrary-looking (if a little more technical) than the standard definition \"the sum of elements on the diagonal\", and also makes it more evident that it doesn't depend on the choice of coordinates:\n\nThe trace is the only linear mapping¹ such that:\ntrace(|u\u003e\u003cv|) = \u003cu|v\u003e\nfor any vectors |u\u003e, |v\u003e.\n\n(The \u003c|\u003e things are bracket notation: |u\u003e stands for a regular vector (think \"column vector\"), \u003cu|, for its corresponding dual vector (think \"row vector\"), |u\u003e\u003cv| for the tensor product of |u\u003e and |v\u003e, and \u003cu|v\u003e to their dot product) \n\nThis still has the drawback of looking like it depends on a choice of inner product, when it actually doesn't.\n\n¹ From linear mappings from V to V, to the scale field K.",
"sig": "74b9741953fdfd1094e080ad7a70042a6169a8a91da5fe1051095e8e9211c777dc7803b0bba9ba5cddccb29ddca52da298b32ac79f7de44dbebccf1c4de9299f"
}