
Let Aiden help you "Keep Pace" with Patents
How did AI arrive at that decision? Easy question - hard to answer.
-
Companies are in the early phases of AI adoption
-
Failures can be just as important as success
-
Don't be surprised by companies that do announce their successes
-
Make sure Competitive Intelligence is part of your AI Strategy
Spent $150 Million on an AI Strategy
Leveraging AI for fraud detection
Hired a top Google Executive to head their Artificial Intelligence efforts

Aiden Suggests
How to learn more about Explainable AI
Suggestion #1: Go to a Conference or a local Meetup
Patents
Yes. Google is well known for filing multiple patents for ML algorithms and methods:
Dropout: Patent WO2014105866A1 - System and method for addressing overfitting in a neural network
Linear regression: Patent US8250003 - Computationally efficient probabilistic linear regression
Random Forest classifier: Patent US6009199 - Classification technique using random decision forests
Parallel neural network training: Patent WO2014105865A1 - System and method for parallelizing convolutional neural networks
Reinforcement learning: Patent US20150100530 - Methods and apparatus for reinforcement learning
Classification (they patented classification, holy crap): Classifying Data Objects - GOOGLE INC.
Word embeddings: Patent US20120310627 - Document classification with weighted supervised n-gram embedding
While that may seem scary at first, Google doesn’t restrict the usage of these algorithms by anyone. This is done mostly to prevent patent trolling.