Graph related to imple Contrastive Learning (SimCLR)
Google

Self-Supervised Simplicity

A simple linear classifier paired with a self-supervised feature extractor outperformed a supervised deep learning model on ImageNet, according to new research.
FixMatch example
Google

Less Labels, More Learning

In small data settings where labels are scarce, semi-supervised learning can train models by using a small number of labeled examples and a larger set of unlabeled examples. A new method outperforms earlier techniques.
Simplified depiction of LSH Attention
Google

Transformers Transformed

Transformer networks have revolutionized natural language processing, but they hog processor cycles and memory. New research demonstrates a more frugal variation.
Association for the Advancement of Artificial Intelligence conference in New York
Google

Meeting of the Minds

Geoffrey Hinton, Yoshua Bengio, and Yann LeCun presented their latest thinking about deep learning’s limitations and how to overcome them.
Capture of a chatbot telling jokes developed by Google Brain
Google

Bot Comic

Androids may not dream of electric sheep, but some crack jokes about horses and cows. Meena, a 2.6-billion parameter chatbot developed by Google Brain, showed impressive conversational ability, discussing a variety of topics.
Graph related to Mixture of Softmaxes (MoS)
Google

Upgrading Softmax

Softmax commonly computes probabilities in a classifier’s output layer. But softmax isn’t always accurate in complex tasks — say, in a natural-language task, when the length of word vectors is much smaller than the number of words in the vocabulary.
Breast cancer screening
Google

Cancer in the Crosshairs

Computer vision has potential to spot cancer earlier and more accurately than human experts. A new system surpassed human accuracy in trials, but critics aren’t convinced.
EfficientDet explained
Google

Easy on the Eyes

Researchers aiming to increase accuracy in object detection generally enlarge the network, but that approach also boosts computational cost. A novel architecture sets a new state of the art in accuracy while cutting the compute cycles required.
Illustration of a fireplace with "Happy holidays" cards in English, Spanish and French
Google

Natural Language Processing Models Get Literate: Top NLP Advances in 2019

Earlier language models powered by Word2Vec and GloVe embeddings yielded confused chatbots, grammar tools with middle-school reading comprehension, and not-half-bad translations. The latest generation is so good, some people consider it dangerous.
Sesame Street characters together
Google

Inside AI’s Muppet Empire: Why Are So Many NLP Models Named After Muppets?

As language models show increasing power, a parallel trend has received less notice: The vogue for naming models after characters in the children’s TV show Sesame Street.
Excerpt from Ring commercial
Google

Neighborhood Watchers

Smart doorbell maker Ring has built its business by turning neighborhoods into surveillance networks. Now the company is drawing fire for using private data without informing customers and sharing data with police.
Observational dropout
Google

Seeing the World Blindfolded

In reinforcement learning, if researchers want an agent to have an internal representation of its environment, they’ll build and train a world model that it can refer to. New research shows that world models can emerge from standard training, rather than needing to be built separately.
Information related to Explainable AI (xAI)
Google

Google's AI Explains Itself

Google's AI platform offers a view into the mind of its machines. Explainable AI (xAI) tools show which features exerted the most influence on a model’s decision, so users can evaluate model performance and potentially mitigate biased results.
Information related to a model that predicts a chemical's smell
Google

Nose Job

Predicting a molecule’s aroma is hard because slight changes in structure lead to huge shifts in perception. Good thing deep learning is developing a sense of smell.
Illustration of health related icons connected to a cloud
Google

When Private Data is Not Private

Google spent the past year training an AI-powered health care program using personal information from one of the largest hospital systems in the U.S. Patients had no idea — until last week.

Subscribe to The Batch

Stay updated with weekly AI News and Insights delivered to your inbox