The literature largely focuses on lossy compression of model updates in deterministic FL. In contrast, stochastic (Bayesian) FL considers distributions over parameters, enabling uncertainty quantification, better generalization, and, crucially, inherent communication-regularized training through a mirror-descent structure. In this paper, we consider both uplink and downlink communication in stochastic FL, and propose a communication framework based on remote source generation. Employing Minimal Random Coding (MRC) for remote generation, we allow the server and the clients to sample from local and global posteriors (sources), respectively, rather than transmitting locally sampled updates. The framework encompasses communication-regularized local optimization and principled compression of model updates, leveraging gradually updated prior distributions as side information. Through extensive simulations, we show that our method achieves 5 32 reduction in total communication cost while preserving accuracy. We further analyze the communication cost, refining existing MRC bounds and enabling precise quantification of uplink and downlink trade-offs. We also extend our method to conventional FL via stochastic quantization and prove a contraction property for the biased MRC compressor to facilitate convergence analysis.

The first algorithm based on multi-armed bandit (MAB) choosesk < K arms (neighbors) repeatedly.

A few months ago, I published a blog that highlighted Qualcomm's plans to enter the data center market with the Cloud AI100 chip sometime next year. While preparing the blog, our founder and principal analyst, Patrick Moorhead, called to point out that Qualcomm, not NVIDIA, probably has the largest market share in AI chip volume thanks to its leadership in devices for smartphones. Turns out, we were both right; it just depends on what you are counting. In the mobile and embedded space, Qualcomm powers hundreds of consumer and embedded devices running AI; it has shipped well over one billion Snapdragons and counting, all which support some level of AI today. In the data center, however, NVIDIA likely has well over 90% share of the market for training.