There is a term from the field of #AI and #machinelearning that I think would be useful for those of us trying to make sense of how people are behaving about #Covid which unfortunately is still very much a thing. It’s called “catastrophic forgetting.”
When a neural network is trained on new examples, the earliest ones submitted to it tend to lose their impact, disproportionately. It’s basically the silicon equivalent of recency bias. Numerous strategies are being employed to overcome it. 🧵
Ed Suominen /
npub185…w4t5m
2024-06-26 00:52:20
Author Public Key
npub185qansar7yhangapt68z9rz23mnlrs7gjt3cmdfnjgnj09d2yhrq6w4t5mPublished at
2024-06-26 00:52:20Event JSON
{
"id": "1cd0faac48fc4170c2e4f33023690210aba6ec43bb513d58aedcc55fcc22302a",
"pubkey": "3d01d9c3a3f12fd9a3a15e8e228c4a8ee7f1c3c892e38db53392272795aa25c6",
"created_at": 1719363140,
"kind": 1,
"tags": [
[
"t",
"ai"
],
[
"t",
"machinelearning"
],
[
"t",
"covid"
],
[
"proxy",
"https://hachyderm.io/users/edsuom/statuses/112680182788376387",
"activitypub"
]
],
"content": "There is a term from the field of #AI and #machinelearning that I think would be useful for those of us trying to make sense of how people are behaving about #Covid which unfortunately is still very much a thing. It’s called “catastrophic forgetting.”\n\nWhen a neural network is trained on new examples, the earliest ones submitted to it tend to lose their impact, disproportionately. It’s basically the silicon equivalent of recency bias. Numerous strategies are being employed to overcome it. 🧵",
"sig": "26abbf3d5b8e63f3665dcf6c1aa6c2d38b547c228a768747c4cfb31d47c6f392c5b5a7ddb75bbc76f79eb97075c7cfcbca971e200b02a9efeb9587a549e740e9"
}