Stability and Generalization in Structured Prediction

By Ben London, Bert Huang, Lise Getoor. 2016

61 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 Statistical Learning Theory

Citations in the corpus

Wireless Network Intelligence at the Edge (2018, background): “The term Γ is the training dataset’s dependency matrix whose definition is elaborated in [139].”
Stable and Fair Classification (2019, methodology): “, the quality of the estimation |R(AS)− E(AS)|, depends on the number of samples N and algorithm A, and has been well studied in the literature (Adomavicius & Zhang, 2011; Bousquet & Elisseeff, 2002; Wainwright, 2006; London et al., 2016).”
Hinge-Loss Markov Random Fields and Probabilistic Soft Logic (2015, background): “In addition to convenience, recent work in learning theory says that repeated dependencies with tied parameters can be the key to generalizing from a few—or even one—large, structured training example(s) (London et al., 2016).”
Fast and Accurate Sequence Labeling with Iterated Dilated Convolutions (2017, background): “may also have better sample complexity, as it imposes more prior knowledge about the structure of the interactions among the tags (London et al., 2016).”
Parametric Adversarial Divergences are Good Task Losses for Generative Modeling (2017, background): “…of learning in structured prediction, some recent papers are devoted to theoretical understanding of structured prediction such as Cortes et al. (2016) and London et al. (2016) which propose generalization error bounds in the same vein as Osokin et al. (2017) but with data dependencies.”
Deterministic PAC-Bayesian generalization bounds for deep networks via generalizing noise-resilience (2019, background): “Our requirement is in fact philosophically similar to London et al. (2016); McAllester (2003), although technically different.”