Amazon Science||1 min read
Graph-centric agentic intelligence
Augmenting a network graph with agentic AI produces a “digital twin” that can help isolate network failures.
Source archive
Amazon Science AI updates
Research updates and technical insights from Amazon Science. This source page gives readers and crawlers a stable route for the latest Amazon Science coverage.
Indexed briefings
6
Latest source-linked updates, ordered newest first.
Latest
Recent briefings
Amazon Science||1 min read
A kernel-centric path to real-time video generation on Trainium
Using the Neuron Kernel Interface, a Reactor–AWS collaboration tackled the dynamic shapes, memory access patterns, and cache management that make real-time autoregressive...
Amazon Science||1 min read
Why don’t machine learning research agents overfit?
New research indicates that AI agents learn compressible models of data, which don’t have enough space to enable memorization.
Amazon Science||1 min read
SOP-Bench: A new benchmark for evaluating AI agents on real business procedures
Extendable framework enables testing agents on the full set of capabilities required to successfully complete a procedure, not isolated proxy tasks.
Amazon Science||1 min read
A new benchmark for evaluating patient-facing health AI agents
PatientAgentBench generates a synthetic patient health record, a realistic clinical vignette, and a patient agent that converses with the AI system under evaluation, to c...
Amazon Science||1 min read
Capturing token IDs during agentic interactions for better reinforcement learning
A new Rust proxy called Turnstile sits between the model backend and the agent harness to capture information lost in mere text transcripts.