PhD student at Mila (UdeM) researching AI models with inductive biases from consciousness and high-level human cognition.
deep_bilateral_network. PyTorch implementation of https://github.com/google/hdrnet for black-box image transformation
65complexity_compositionality. Jupyter Notebook
9encoder_dimensionality. Jupyter Notebook
8distribution_clustering. Implements (high-dimenstional) clustering algorithm described in https://arxiv.org/pdf/1804.02624.pdf
6CPPN. A CPPN neural net to generate high resolution pattern images
6syntax_tree_network. Tree-structured neural network that uses explicit syntax as an inductive bias to learn compositional representations of language.
2MusicGenerator. A deep belief net coupled with a recurrent LSTM network that composes original piano pieces. Implemented in TensorFlow.
2neural_scene_representation. Train a GQN neural network and compare its representational structure to that of scene regions from human fMRI data.
2lighthouse_deep_learning_tutorial. Jupyter Notebook
1neural-production-systems. Implementation (unofficial) of work described in https://arxiv.org/abs/2103.01937
1encoder_warping. Jupyter Notebook
1pyro_gmmvae. A VAE implemented in Pyro with a Gaussian Mixture Model specifying the prior or latent variables. The Gaussian Mixture prior encourages natural clustering of the latent space according to the classes of the data. It also permits the generation of a desired class after training, since we can manually select which Gaussian component to sample from in order to generate a sample from p(x|z).
1StochasticNeuralNetwork. Experimentations training a stochastic neural network on MNIST data
1prequential_coding. Python
1Circuit-Solver. An iOS application to solve circuits
1