{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreidnvfbsm7r7h2xllqyag3fnpgiha56msp3niyqcwqqvrxomtisokq",
    "uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3mi2ofpnemca2"
  },
  "path": "/t/searching-for-endorsement-on-arxiv-cs-cr-cs-lg/174701#post_1",
  "publishedAt": "2026-03-27T14:48:35.000Z",
  "site": "https://discuss.huggingface.co",
  "tags": [
    "doi.org/10.13140/RG.2.2.31561.51040",
    "https://arxiv.org/auth/endorse?x=CNPVKN"
  ],
  "textContent": "I am a 2nd-year Software Engineering undergraduate from Malaysia, and I am seeking an arXiv endorsement to publish my recent work in cs.CR (Cryptography and Security) and cs.LG (Machine Learning).\n\nMy research introduces **AEGIS** , an adversarial network anomaly detection system that completely discards payload inspection. Instead, it utilizes 6-dimensional continuous-time flow physics to mathematically expose advanced cryptographic mimicry (such as VLESS Reality and GhostBear) and automated C2 tunnels.\n\nThis research presents a novel architectural approach by fusing Hyperbolic Poincaré embeddings with Liquid Time-Constants, processed through a linear-time Mamba-3 core to evaluate traffic at sub-millisecond line-rates.\n\n**Key Empirical Results** (evaluated on a 400GB zero-trust adversarial corpus):\n\n  * F1-Score: 0.9952 (AUC: 0.9998)\n\n  * True Positive Rate: 99.50% against zero-day evasion\n\n  * Inference Latency: 262μs (bare-metal RTX 4090)\n\n\n\n\nThe paper is published under CC BY 4.0 and the full 10GB flow-physics dataset is publicly available. Manuscript available on request.\n\nPreprint (ResearchGate): doi.org/10.13140/RG.2.2.31561.51040\n\nDataset : Vix0007 @ HuggingFace\n\nIf anyone with endorsement privileges in cs. would be willing to review my methodology, I would be incredibly grateful.\n\ncs.CR: https://arxiv.org/auth/endorse?x=CNPVKN\n\ncs.LG: IMRASY\n\nThank you for your time and support of independent research.\n\n**Vickson Ferrel** @ Vixero Technology Enterprise | UNIMAS, Malaysia | ORCID: 0009-0009-0155-0913",
  "title": "Searching for endorsement on arxiv (cs.CR/cs.LG)"
}