{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreic44avijfvoru5aeoysj4mz6vpohvjbjddlq6zjbzctwrmc5iqdoe",
"uri": "at://did:plc:4rgrdigiftglskeax4wvmsev/app.bsky.feed.post/3mflbfahusmz2"
},
"coverImage": {
"$type": "blob",
"ref": {
"$link": "bafkreiflo6xt7is6b2iafwghkjahlgggocme5jwjsbeuqqwcywuvjhmszm"
},
"mimeType": "image/png",
"size": 24783
},
"path": "/abs/2602.19552v1",
"publishedAt": "2026-02-24T01:00:00.000Z",
"site": "https://arxiv.org",
"tags": [
"Kasper Green Larsen",
"Markus Engelund Mathiasen",
"Chirag Pabbaraju",
"Clement Svendsen"
],
"textContent": "**Authors:** Kasper Green Larsen, Markus Engelund Mathiasen, Chirag Pabbaraju, Clement Svendsen\n\nIn this paper, we consider the problem of replicable realizable PAC learning. We construct a particularly hard learning problem and show a sample complexity lower bound with a close to $(\\log|H|)^{3/2}$ dependence on the size of the hypothesis class $H$. Our proof uses several novel techniques and works by defining a particular Cayley graph associated with $H$ and analyzing a suitable random walk on this graph by examining the spectral properties of its adjacency matrix. Furthermore, we show an almost matching upper bound for the lower bound instance, meaning if a stronger lower bound exists, one would have to consider a different instance of the problem.",
"title": "The Sample Complexity of Replicable Realizable PAC Learning"
}