Towards more human-like concept learning in machines : compositionality, causality, and learning-to-learn

44 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

What can the brain teach us about building artificial intelligence? (2016, background): “What additional ingredients may be needed to rapidly learn more powerful and more general-purpose representations? A relevant case study is from our own work on the Characters Challenge (sect. 3.1; Lake 2014; Lake et al. 2015a).”
Gradient-based Hyperparameter Optimization through Reversible Learning (2015, methodology): “We illustrate this “architecture learning” with a multitask learning problem, the Omniglot data set (Lake, 2014).”
Deep Neural Networks Predict Category Typicality Ratings for Images (2015, background): “While the convnet was trained on an average of 1200 images per class, people need far less data in order to learn a new category (Lake, 2014).”
Modeling, Inference and Optimization With Composable Differentiable Procedures (2016, methodology): “We illustrate this “architecture learning” with a multitask learning problem using the Omniglot data set [56].”
Sticking the Landing: Simple, Lower-Variance Gradient Estimators for Variational Inference (2017, methodology): “Benchmark Datasets We evaluate our path derivative estimator using two benchmark datasets: MNIST, a dataset of handwritten digits [LeCun et al., 1998], and Omniglot, a dataset of handwritten characters from many different alphabets [Lake, 2014].”