General conditions for predictivity in learning theory

By Tomaso A. Poggio, Ryan M. Rifkin, Sayan Mukherjee, Partha Niyogi. 2004

304 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computational Theory and Mathematics

How researchers see this paper

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

In Artificial Intelligence

In Computational Theory and Mathematics

Citations in the corpus

Generalization in anti-causal learning (2018, background): “Formally, for ERM this corresponds to the notion of consistency, (Vapnik, 1998; Poggio et al., 2004).”
Robustness and Regularization of Support Vector Machines (2008, background): “It is now well known that a stable algorithm such as SVM has desirable generalization properties, and is statistically consistent under mild technical conditions; see for example Bousquet and Elisseeff (2002); Kutin and Niyogi (2002); Poggio et al. (2004); Mukherjee et al. (2006) for details.”
Making AI Forget You: Data Deletion in Machine Learning (2019, neutral): “In statisticalandcomputationallearningtheory, thisideaisknowntoas stability [60,47,50,29,77,68].”
The Dynamics of Learning: A Random Matrix Approach (2018, background): “To answer such a question, statistical learning theory has proposed interpretations from the viewpoint of system complexity (Vapnik, 2013; Bartlett & Mendelson, 2002; Poggio et al., 2004).”
Controlling Bias in Adaptive Data Analysis Using Information Theory (2015, methodology): “e selection rule Tcould represent the choice made by a data-analyst, like Bob, after performing several rounds of exploratory analysis. A powerful line of work in computer science and learning theory [12, 13, 14] has explored the role of algorithmic stability in preventing overﬁtting. Related to sta”
Learning theory: stability is sufficient for generalization and necessary and sufficient for consistency of empirical risk minimization (2006, background): “There are other “simple” stability conditions equivalent to LOO stability [17, 19].”
Vector Field Learning via Spectral Filtering (2010, background): “These algorithms are motivated by the results connecting learning theory and regularization of ill-posed problems [3] and the relations between stability and generalization [4,5].”
Spectral Methods for Regularization in Learning Theory (2006, background): “Anyway this point of view is not new to learning theory since the connection between stability and generalization was considered in Bousquet and Elisseeff (2002), Mukherjee et al. (2004), Poggio et al. (2004).”
Understanding deep learning requires rethinking generalization (2016, background): “This is commonly done with some notion of stability, such as uniform stability (Mukherjee et al., 2002; Bousquet & Elisseeff, 2002; Poggio et al., 2004).”
Stability of Stochastic Gradient Descent on Nonsmooth Convex Losses (2020, unknown): “nd Vondrák, 2018, 2019; Bousquet et al., 2019). A long line of work focuses on the relationship between various notions of stability and learnability in supervised setting (see (Kearns and Ron, 1999; Poggio et al., 2004; Shalev-Shwartz et al., 2010) for an overview). These works employ relatively we”
Tikhonov, Ivanov and Morozov regularization for support vector machine learning (2016, background): “We can link, instead, both the Ivanov and Tikhonov formulations to the Algorithmic Stability framework (Bousquet and Elisseeff 2002; Poggio et al. 2004; Tomasi 2004; Oneto et al. 2014; Elisseeff et al. 2005; Evgeniou et al. 2004).”
How Much Does Your Data Exploration Overfit? Controlling Bias via Information Usage (2015, background): “Apowerful line of work in computer science and learning theory [6, 26, 27] has explored the role of algorithmic stability in preventing overﬁtting.”
Online Learning via the Differential Privacy Lens (2017, unknown): “Stability of output in presence of small changes to input is a desirable feature of methods in statistics and machine learning [11, 19, 31, 42].”
Adaptive Learning with Robust Generalization Guarantees (2016, result): “The second follows from an important line of work [BE02, PRMN04, SSSSS10] that establishes connections between the stability of a learning algorithm and its ability to generalize.”