It supported backprop from day one, so I'm not sure where you got that Idea. From the ICML paper you mention above:
@jeffdean
-
CPU GPU TPU support evolution in machine learning framework
By
–
(Initial version supported CPUs. Later versions added support for GPUs and TPUs).
-
DistBelief trained 10000+ models across architectures
By
–
Distbelief was used internally for training 10,000+ distinct models of all kinds of architectures (convnets, RNNs, LSTMs, feed forward networks, sparse MoEs, etc), with all kinds of different training objectives (supervised, unsupervised, RL, etc).
-
DistBelief: Distributed Training and Knowledge Distillation Framework
By
–
Or https://
arxiv.org/abs/1503.02531, which describes how distbelief was used to train the baseline model (a quite large model for the circa 2014/2015 time frame), and later describes using the framework to train specialists and then distill them into a single model. -
Supervised ConvNets Scale Better With Smaller Parameter Count
By
–
Supervised convnets worked just fine with disbelief (in fact they scale better because of the smaller parameter count). For example: https://
papers.nips.cc/paper_files/pa
per/2013/hash/7cce53cf90577442771720a370c3c723-Abstract.html
… -
Professor Cordelia Schmid Wins Monte-Carlo Woman Year Award
By
–
Congratulations on the well-deserved recognition to my colleague, @CordeliaSchmid
! "The Monte-Carlo International Woman of the Year Award was presented by Princess Stephanie of Monaco to Professor Cordelia Schmid, a world leading specialist in computer assisted recognition, -
1 Trillion Edges: Breakthrough in AI Model Architecture
By
–
An edgy post… 1 trillion edges, in fact! https://t.co/ymbl16TL2X
— Jeff Dean (@JeffDean) 1 mai 2024An edgy post… 1 trillion edges, in fact!
-
Med-Gemini Models Advance Healthcare AI Capabilities
By
–
The Med-Gemini models are research models for helping us evaluate the capabilities of these approaches in the medical domain. The best of these will make their way into Google Cloud's MedLM product offerings (and MedLM models with strong capabilities are already available there
-
AI Models Transforming Clinical Care and Patient Understanding
By
–
I'm very excited about the possibilities of these models to help clinicians deliver better care, as well as to help patients better understand their medical conditions. AI for healthcare is going to be one of the most impactful application domains for AI, in my opinion.