Google DeepMind launches AlphaGenome 2, trained on 100M+ cell profiles across 68 tissues to predict genetic mutation effects, for non-commercial research (Robert Hart/The Verge)
DeepMind's new DNA model better predicts how mutations affect different cells, potentially opening paths to treatment.
Researchers at Google DeepMind employed agentic AI to mathematically verify several key cryptographic guarantees of the Signal protocol, marking a significant milestone in the world of cryptography and artificial intelligence. This project, completed in just four weeks, holds dual importance. Firstly, it provides independent validation of the cryptography safeguarding vast amounts of global communications in platforms such as Signal, Google Messages, and WhatsApp.
Secondly, it showcases Google DeepMind's capabilities in formal verification through a novel AI framework named AlphaProof Code. Beyond this, the use of AI in formal verification is experiencing a surge, though its practical application and effectiveness in securing large codebases remain uncertain. Formal verification, often hailed as a holy grail in software security, employs mathematical checks to ensure software operates as intended, effectively eliminating critical vulnerabilities.
Written by urgent.news from Ars Technica's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
- Google DeepMind just mathematically verified key parts of Signal arstechnica.com
