Spectral Inference Networks: Unifying Deep and Spectral Learning

Authors: David Pfau, Stig Petersen, Ashish Agarwal, David G. T. Barrett, Kimberly L. Stachenfeld

ICLR 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We show results of training Spectral Inference Networks on problems in quantum mechanics and feature learning for videos on synthetic datasets. Our results demonstrate that Spectral Inference Networks accurately recover eigenfunctions of linear operators and can discover interpretable representations from video in a fully unsupervised manner.
Researcher Affiliation Industry David Pfau1, Stig Petersen1, Ashish Agarwal2, David G. T. Barrett1 & Kimberly L. Stachenfeld1 1Deep Mind 2Google Brain London, UK Mountain View, CA, USA {pfau, svp, agarwal, barrettdavid, stachenfeld}@google.com
Pseudocode Yes The full algorithm for training Spectral Inference Networks is given in Alg. 1, with Tensor Flow pseudocode in the supplementary material in Sec. B.
Open Source Code Yes Code is available at https://github.com/deepmind/spectral_inference_networks
Open Datasets Yes We provide a qualitative comparison of the performance of Sp IN with the SFA objective against the successor feature approach for learning eigenpurposes Machado et al. (2018) on the Arcade Learning Environment (Bellemare et al., 2013).
Dataset Splits No The paper does not explicitly describe training, validation, and test splits with specific percentages or counts. It mentions using 'held-out frames' for testing, but no distinct validation set.
Hardware Specification No The paper does not provide specific details about the hardware used, such as GPU or CPU models.
Software Dependencies No The paper mentions 'Tensor Flow' in the supplementary material but does not specify a version number.
Experiment Setup Yes Details of the training network and experimental setup are given in the supplementary material in Sec. C.1.