Attention-based Deep Multiple Instance Learning
The official implementation of a method for multiple instance learning using deep neural networks and attention mechanism.
Keywords: multiple instance learning, deep learning, attention, PyTorch, weak classification
The reference implementation for the attri2vec algorithm.
Keywords: Attri2vec, representation learning, graph machine learning, node classification, link prediction
AugLy is a data augmentations library that currently supports four modalities (audio, image, text & video) and over 100 augmentations. Each modality’s augmentations are contained …
Keywords: data augmentation, machine learning, Python, PyTorch
AutoGL is an AutoML framework & toolkit for machine learning on graphs. AutoGL is developed for researchers and developers to conduct AutoML on graph datasets …
Keywords: AutoML, graph neural architecture search, machine learning, graph neural networks, NAS, GraphNAS, Python, PyTorch
Backprop is a Python library for fine-tuning and deploying machine learning models. It includes pre-trained models for natural language processing, e.g., question answering, text summarisation, …
Keywords: neural networks, pre-trained models, fine-tuning, library, natural language processing, computer vision, Python, PyTorch
Implementation of multiple instance learning (MIL) with the interactions between bags modelled using a Bayesian graph neural network (GNN).
Keywords: multiple instance learning, graph neural networks, Bayesian, MIL, GNN, machine learning, Python, Keras, TensorFlow
Implements scalable variational inference methods for Bayesian structure learning. Bayesian Causal Discovery Nets (BCD Nets) is a variational inference framework for estimating a distribution over …
Keywords: causal machine learning, causal structure learning, Bayesian methods, variational inference, Python, Jax
A framework for benchmarking the performance of graph neural network algorithms including datasets.
Keywords: graph neural network, machine learning, datasets, benchmark, graph convolutions, PyTorch
BERTScore leverages the pre-trained contextual embeddings from BERT and matches words in candidate and reference sentences by cosine similarity. It has been shown to correlate …
Keywords: evaluation metric, text generation, text similarity, language model, NLP, natural language processing, Python, PyTorch
As missing values are frequently present in genomic data, practical methods to handle missing data are necessary for downstream analyses that require complete datasets. In …
Keywords: variational auto-encoder, VAE, beta VAE, data imputation, genomics, Python, PyTorch