Deep Set Prediction Networks

This is the official implementation of our NeurIPS 2019 paper Deep Set Prediction Networks. We propose a new way of predicting sets with a neural network that doesn't suffer from discontinuity issues. This is done by backpropagating through a set encoder to act as a set decoder. You can take a look at the poster for NeurIPS 2019 or the poster for the NeurIPS 2019 workshop on Sets & Partitions.

To use the decoder, you only need dspn.py. You can see how it is used in model.py with build_net and the Net class. For details on the exact steps to reproduce the experiments, check out the README in the dspn directory. You can download pre-trained models and the predictions thereof from the Resources page.

BibTeX entry

@inproceedings{zhang2019dspn,
    author        = {Yan Zhang and Jonathon Hare and Adam Pr\"ugel-Bennett},
    title         = {{Deep Set Prediction Networks}},
    booktitle     = {Advances in Neural Information Processing Systems 32},
    year          = {2019},
    eprint        = {1906.06565},
    url           = {https://arxiv.org/abs/1906.06565},
}