We introduce a Mixture of Raytraced Experts, a stacked Mixture of Experts (MoE) architecture which can dynamically select sequences of experts, producing computational graphs of variable width and depth. Existing MoE architectures generally require a fixed amount of computation for a given sample. Our approach, in contrast, yields predictions with increasing accuracy as the computation cycles through the experts' sequence. We train our model by iteratively sampling from a set of candidate experts, unfolding the sequence akin to how Recurrent Neural Networks are trained. Our method does not require load-balancing mechanisms, and preliminary experiments show a reduction in training epochs of 10\% to 40\% with a comparable/higher accuracy. These results point to new research directions in the field of MoEs, allowing the design of potentially faster and more expressive models. The code is available at https://github.com/nutig/RayTracing

"Computing is the first productive force of the digital economy." On July 29, 2022, at the SenseTime Science and Technology Sub-forum of the First Computing Conference in China, Gao Shanshan, director of Shandong Sino US Digital Media International Cooperation Research Center said that, in the era of AI, computing infrastructure keeps changing such industries as finance, medicine and data center. Therefore, AI computing has become the main increment of digital economy development in various countries, and also the foundation of the era of digital economy. Computing represents a new type of productivity. Who owns the computing of the future digital economy industry development will have the ultimate power to lead the development of the digital economy in digital economy.

Improvement of steam engine by Watt triggered the first industrial revolution, and human society entered industrial civilization from agricultural civilization; electromagnetic induction principle triggered the second industrial revolution, and human society entered the age of electricity; the strong power brought by the third industrial revolution marked by information technology and the fourth industrial revolution featured by artificial intelligence deeply influenced our economy and society. In the age of digitalization, data is new means of production and computing is new productivity. Network reaches everywhere, computing exists everywhere and intelligence is used everywhere. Those who have strong computing have the password to win the future. At the new beginning of the age of computing, China is running extremely fast on the new track of digital economy.

This paper presents the Computing Networks (CNs) framework. CNs are used to generalize neural and swarm architectures. Artificial neural networks, ant colony optimization, particle swarm optimization, and realistic biological models are used as examples of instantiations of CNs. The description of these architectures as CNs allows their comparison. Their differences and similarities allow the identification of properties that enable neural and swarm architectures to perform complex computations and exhibit complex cognitive abilities. In this context, the most relevant characteristics of CNs are the existence multiple dynamical and functional scales. The relationship between multiple dynamical and functional scales with adaptation, cognition (of brains and swarms) and computation is discussed.