. . . . . " You can not run a closed model yourself, right? well...\nThey prevent information leakage through quantum light properties\nWhat?!\nSimply:you can compute the network outputs but not recreate the weights and the API cannot access your data!\n@RLEatMIT\n@KfirSulimany @Dirk_Englund https://twitter.com/LChoshen/status/1823374595334009145/photo/1\n\n The basic Idea is that for every multiplication in the network you multiply two symmetric things an input and an output matrix.\nSo, if each side can send a matrix but only read the output each can keep their secret (one the input one the model weights)\n\nhttps://www.alphaxiv.org/abs/2408.05629\n\n And apparently you can multiply (super fast) by encoding the matrix as a light wave, then the multiplication is the two waves meeting and the output is more or less the only thing you can do.\n\n I was involved in the \"more or less\", e.g. to avoid reconstruction through a canonical basis as inputs you can use invariances such as sending a permuted network each time\nhttps://www.alphaxiv.org/abs/2408.05629\n\n" . "machinelearning" . "privacy" . "quantum-securemultipartydeeplearning" . "quantumcomputing" . "securemultipartycomputation" . . "webpage" . . . . . . . . . . . . "RSA" . "MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEArHtI92jm8pAYVsvJabxLGfOT+7G0JyJGh2gwjB5x2pFPga6wWTd+rNBWWUZViIFnaJrBEsJpgdnoupLU9ppwn+khMiGRfxqGsDDzwHcj3Jc75CRys7d3etwXdBdoXfBgjsJiZBazwm13idr6tljRrC1TaEJBnRQAqzBw9cLDeGY77cSznzXT39feUGT168dpCSE9O6u/48DvvWVqciHGsH9cQ+LroJJVsMrorwtsdZnAK+q48wtIP6pIpw5shSJ5LnA0qeN/f4TvTFDV6ItYIXjiWWpTECc/Bxmfnyat3B5xWCu9nvz8fEs7Ns0TuzQwT3/K55iSKDEIi/E0nO97xwIDAQAB" . "iZMDkFBvlGddiWxopE98TVDnJaqe+eg8yCmgBEh2aeNiZH79KlCHUDehcDY6BAyotwsdISjlYlpGKpcmqepcfHTq25+7IrFymLy4eSykfNgpAfspMDHe9+SLIHnTsMWldSw7tCy5XPSN9wzGwa4Jv9jtiYnKRD3KeBO18Mg8RhgqKwRrIkDzeSzbTn9Tp+u/2NBfu/00Z/nTljLurkdSLwiNdgCgrCdimptDuXsXW+rJJhM/1cCqWC22ZtZGbRHNCW14IihurNUNuUu7kZVp90x+lKqVPD83aM5I6X6Q/n7xhU4Jy08LJj2MXIVx68qdoCUP3d4I0KAdoNfArAHO5w==" . . . . "2024-09-03T21:16:02.887Z"^^ . . . . . "CoSMO Semantic Post" . . "0xf6ECcfD463afB464dcC85b051DF2E93E2646E6D2" . . "Leshem Choshen 🤖🤗 @ICML wanna talk?" .