International Conference on Learning Representations (ICLR)

By M. Abbas, Jyri J. Kivinen, Tapani Raiko. 2016

662 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Evaluation & Benchmarks

Citations in the corpus

Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks (2017, methodology): “This approach has been applied to learning to optimize deep networks (Hochreiter et al., 2001; Andrychowicz et al., 2016; Li & Malik, 2017), as well as for learning dynamically changing recurrent networks (Ha et al., 2017).”
Key-Value Memory Networks for Directly Reading Documents (2016, methodology): “The Key-Value Memory Network model is based on the Memory Network (MemNNs) model (Weston et al., 2015; Sukhbaatar et al., 2015) which has proven useful for a variety of document reading and question answering tasks: for reading children’s books and answering questions about them (Hill et al.,…”
Order Matters: Sequence to sequence for sets (2015, neutral): “We also show the importance of ordering in a variety of tasks in which seq2seq has successfully been applied, and include experimental results to support our claims and extensions to the existing models.”
Continuous Deep Q-Learning with Model-based Acceleration (2016, methodology): “This makes it possible to train policies for complex tasks with minimal feature and policy engineering, using the raw state representation directly as input to the neural network.”
Stronger data poisoning attacks break data sanitization defenses (2018, background): “Designing models that are robust to such attacks, as well as coming up with more effective attacks, is an active area of research (Papernot et al., 2016c; Madry et al., 2017; Tramèr et al., 2017; Wong & Kolter, 2018; Raghunathan et al., 2018; Sinha et al., 2018; Athalye et al., 2018; Papernot & McDa”
Conditional variance penalties and domain shift robustness (2017, background): “the problems of cause-effect inference and generative adversarial networks (Chalupka et al. 2014; Lopez-Paz et al. 2017; Lopez-Paz and Oquab 2017; Goudet et al. 2017; Bahadori et al. 2017; Besserve et al. 2018; Kocaoglu et al. 2018).”
Group Equivariant Convolutional Networks (2016, background): “Their “sameness” (as judged by the network) is preserved under symmetry transformations.”
Faster ILOD: Incremental Learning for Object Detectors based on Faster RCNN (2020, background): “often fail on the previously trained tasks — a problem called catastrophic forgetting (Goodfellow et al., 2014; McCloskey and Cohen, 1989).”
Continual Learning with Adaptive Weights (CLAW) (2019, background): “Other algorithms based on replay mechanisms include (Aljundi et al., 2019a;b).”
Stick-Breaking Variational Autoencoders (2016, methodology): “To the best of our knowledge, neither SGVB nor any of the other recently proposed amortized VI methods (Kingma & Welling, 2014b; Rezende & Mohamed, 2015; Rezende et al., 2014; Tran et al., 2016) have been used in conjunction with BNP priors.”
An Empirical Comparison of Instance Attribution Methods for NLP (2021, methodology): “…predictions (which may reveal potentially problematic training examples), they have not yet been widely adopted, at least in part because even approximating inﬂuence functions (Koh and Liang, 2017)—arguably the most principled at-tribution method—can be prohibitively expensive in terms of compute.”