Reverse Engineering Self-Supervised Learning — mentioned by haltist

Fri, 15 Dec 2023 01:24:13 GMT

The data that AlphaFold was trained on included all that information and more. The database they used for training included software simulations (and real world data) that accounted for atomic (quantum) interactions. The 3D structure of proteins includes all the quantum interactions. More generally, AI models (aka very large function graphs) are trained on tuples that represent mappings of inputs to outputs (input -> output). The idea then is that whatever structure exists in those pairs/tuples/mappings is discovered by the training process with the help of gradient descent which tunes the…

Paper and discussion history. A mention is not an endorsement.

Reverse Engineering Self-Supervised Learning — HN submission

Sun, 04 Jun 2023 11:43:46 GMT

Reverse Engineering Self-Supervised Learning

86 points; 16 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Reverse Engineering Self-Supervised Learning · HackerPapers discussions

Reverse Engineering Self-Supervised Learning — mentioned by haltist

Reverse Engineering Self-Supervised Learning — HN submission