{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreifl3vxp7bfdpppc7phi65o7ojqqo6wo6bhbtxvgk6vincheyasiea",
"uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3mivuan5npq42"
},
"path": "/t/snn-credit-assignment-problem-is-not-unsolved-anymore/175042#post_1",
"publishedAt": "2026-04-07T09:19:59.000Z",
"site": "https://discuss.huggingface.co",
"textContent": "The credit assignment problem in Spiking Neural Networks (SNNs) has been treated as unsolved for years due to reliance on BPTT and unstable training.\n\nI’ve been working on a **data-driven, event-based approach** that enables effective credit assignment without full BPTT.\n\nEarly results show:\n\n * Stable training in deeper SNNs\n\n * Better temporal credit propagation\n\n * Lower compute overhead\n\n\n\n\nThis is backed by **real experimental results** , and I’m preparing a research paper.\n\nI believe this problem is no longer “unsolved” we’re closer to practical SNN learning than people think.\n\n**Looking for collaborators and feedback** (SNN, neuromorphic, biologically plausible learning).",
"title": "SNN Credit Assignment Problem is NOT Unsolved Anymore"
}