Deep Learning
Reviewer 1 [ The ] methodology combines multiple different ideas in causal inference (multi-headed deep learning
The baselines in their evaluations are not completely clear . In addition [...] We have clarified this. It seems weird that Equation 2.2 has no hyperparameter ... We have clarified this. Indeed, there is a hyperparameter. We used an arbitrary fixed value (1.0) to avoid unfairly advantaging our method via hyperparam search.
Heterogeneous Graph Learning for Visual Commonsense Reasoning
Weijiang Yu, Jingwen Zhou, Weihao Yu, Xiaodan Liang, Nong Xiao
It poses great challenges over the proper semantic alignment between vision and linguistic domains and knowledge reasoning to generate persuasive reasoning paths.
NeurVPS: Neural Vanishing Point Scanning via Conic Convolution
Yichao Zhou, Haozhi Qi, Jingwei Huang, Yi Ma
Neural Information Processing Systems http://nips.cc/
Saccader: Improving Accuracy of Hard Attention Models for Vision
Gamaleldin Elsayed, Simon Kornblith, Quoc V. Le
Neural Information Processing Systems http://nips.cc/
Supplement material for " Walsh-Hadamard Variational Inference for Bayesian Deep Learning " Simone Rossi
(the cube in Figure 1). For each dimension, the orange dots represent 20 repetitions. The median distance is displayed in black. Few outliers (with distance greater than 3.0) appeared, possibly due to imperfect numerical optimization. Results are reported in Table 2. Comparison of test error w.r.t. the number model parameters ( top: mean field, bottom: full covariance).
Walsh-Hadamard Variational Inference for Bayesian Deep Learning
Bayesian inference is desirable but extremely challenging.
Beyond temperature scaling: Obtaining well-calibrated multi-class probabilities with Dirichlet calibration
Meelis Kull, Miquel Perello Nieto, Markus Kängsepp, Telmo Silva Filho, Hao Song, Peter Flach
Neural Information Processing Systems http://nips.cc/