analiza aiqbench
Sakana AI's "Diffusing Blame" paper presents a significant architectural shift by training dual-stream networks that obey Dale's principle without relying on backpropagation or weight transport. By utilizing modulo error routing to scale this local learning rule from MNIST to CIFAR-10, the research offers a biologically plausible alternative for deep learning models where standard gradient descent is physically unfeasible.
Sakana AI’s Error Diffusion Trains Dale-Compliant Dual-Stream Networks, Reaching 96.7% MNIST and 61.7% CIFAR-10 Without Backpropagation
Backpropagation relies on weight transport, which biological circuits likely cannot implement. Sakana AI's Error Diffusion sidesteps that constraint, training dual-stream excitatory/inhibitory networks that obey Dale's principle. This piece breaks down how modulo error routing scales the rule from MNIST to CIFAR-10 and reinforcement learning, and what its task-dependent ablations reveal. The post Sakana AI’s Error Diffusion Trains Dale-Compliant Dual-Stream Networks, Reaching 96.7% MNIST and 61.7% CIFAR-10 Without Backpropagation appeared first on MarkTechPost.
Backpropagation dominates deep learning, yet it uses a mechanism the brain likely cannot. Specifically, the backward pass needs exact transposes of forward weight matrices. This is the weight transport problem. Sakana AI's new paper, Diffusing Blame, confronts this constraint directly. The research team trains networks that obey Dale's principle while avoiding weight transport entirely. What is Error Diffusion? Error Diffusion (ED) is a local learning rule, first proposed by Kaneko (2000). Each weight update depends on three signals only: presynaptic activity, a postsynaptic activation derivative, and a single global error sign. Consequently, ED never transports…
Extras afișat conform termenilor RSS ai publicației · 95 cuvinte · textul integral și figurile sunt pe site-ul publicației
Subiecte conexe
Build a Physical AI model factory with NVIDIA Cosmos 3 on SageMaker HyperPod
AWS · 4 sept. 2026
Physics-Grounded Fluid Video Generation with a Simulation Dataset and Dual-Stream Optical-Flow Supervision
arXiv · 29 iul. 2026
TrioPose: Native Triple-Stream Diffusion Transformers for Pose-Guided Text-to-Image Generation
arXiv · 8 iun. 2026
One Diffusion Model, Two Roles: Guided Trajectory Planning and Safety-Critical Scenario Generation in Closed-Loop Simulation
arXiv · 7 sept. 2026