Tech Meridian ← ENTITY INDEX
PROMY MERIDIAN RU

COMPANY · ENTITY #11135

Pushmeet Kohli

Related event timeline, sources and context from the news index.

EVENT TIMELINE

3

RESEARCH · 3 SOURCES · WIRED AI · Google DeepMind · The Decoder

DeepMind's AlphaProof Code formally verifies parts of the Signal protocol

Google DeepMind used a new agentic AI framework called AlphaProof Code to formally verify several key cryptographic properties of the Signal protocol in a four-week proof-of-concept project. The analysis found no new vulnerabilities in the examined portions, confirmed that older known-vulnerable versions fail verification, and released the specification and proof, while researchers cautioned that scaling AI-driven formal verification remains an open challenge.

9.0

MODELS · 1 SOURCE · Google DeepMind

AlphaGenome 2: conditional DNA sequence model predicts variant effects in specific cell types

AlphaGenome 2 is a next-generation DNA-sequence model that predicts how sequence variants affect molecular readouts (gene expression, chromatin accessibility, histone marks, DNA methylation) in specific human cell types and states. Trained on over 107 million individual cell profiles across 68 tissues, it uses a conditional architecture that takes a biological condition plus a DNA sequence, improves detection of cell-type–specific regulatory variants versus the prior model, supports cross-modality imputation, and is available for non-commercial research via the AlphaGenome API (commercial preview private).

8.0

RESEARCH · 1 SOURCE · Google DeepMind

SynthID Bio: watermarking AI-designed proteins while fine‑tuning AlphaFold 3

The authors introduce SynthID Bio, a family of watermarking methods that embeds imperceptible signatures into protein sequences and predicted 3D coordinates so the watermark is verifiable on synthesized proteins; in wet‑lab tests on three targets (VEGF‑A, SARS‑CoV‑2 RBD, PD‑L1) watermarked protein binders retained hit rates, binding affinities, and sequence diversity comparable to unwatermarked designs. For folding, the approach fine‑tunes part of AlphaFold 3’s diffusion network to make predicted coordinates carry a detectable signature while preserving prediction accuracy and structural feature distributions.

8.0