Stochastic Approximation

By F. Downton. 1969

How researchers see this paper

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

In Optimization & Training

Citations in the corpus

Taming the Noise in Reinforcement Learning via Soft Updates (2015, background): “In particular, given conditions (45) for α t , the contractive nature of B ∗ , inﬁnite visits to each pair ( s t , a t ) and assuming that | z t | < ∞ , G t is guaranteed to converge to the optimal G ∗ with probability 1 [17, 22].”
DeltaGrad: Rapid retraining of machine learning models (2020, neutral): “See e.g., (Gladyshev, 1965; Amari, 1967; Kul’chitskiy & Mozgovoy, 1992; Bert-sekas & Tsitsiklis, 1996; Moulines & Bach, 2011; Karimi et al., 2016; Bottou et al., 2018; Gorbunov et al., 2019; Gower et al., 2019), etc, and references therein.”
The statistical evaluation of social network dynamics (2001, methodology): “This approach to network evolution is in line with the theoretical sociological principle of methodological individualism, and was referred to by Snijders (1996) as a stochastic actor-oriented model.”
A one-measurement form of simultaneous perturbation stochastic approximation (1997, background): “Ruppert, D. (1983). Kiefer-Wolfowitz procedure. In N. L. Johnson and S. Katz (Eds), Encyclopedia of Statistical Sciences. Vol. 4. DD. 379-381. Wilev. New York. One of the areas for which SPSAl seems most appropriate Ruppert, D. (1988). Efficient estimators from a slowly is in feedback control proble”
An incremental network for on-line unsupervised classification and topology learning (2006, background): “After the network size becomes stable, the network is fine tuned by stochastic approximation (Wasan, 1969).”
Reward Constrained Policy Optimization (2018, methodology): “Thus, ( λ Finally, as seen in Theorem 2 of Chapter 2 of Borkar et al. (2008), θ n → θ ∗ a.s. then λ n → λ ( θ ∗ ) a.s. which completes the proof.”
Pattern-recognizing stochastic learning automata (1985, methodology): “We follow the presentation given in [4] and [26] and consider only the case of two pattern classes.”
The Asymptotic Convergence-rate of Q-learning (1997, methodology): “Firstly, notice that as a consequence of this change the process ^ Q t clearly converges to Q and this convergence may be investigated along each component (x; a) separately using standard stochastic-approximation techniques (see e.g. (Wasan, 1969; Poljak and Tsypkin, 1973)).”
Adaptation, Learning, and Optimization over Networks (2014, background): “The work by [201] dealt primarily with scalar weights w and was extended by [40, 218] to weight vectors — see [259].”
A Correlated Probit Model for Joint Modeling of Clustered Binary and Continuous Responses (2001, background): “Such weighting schemes may also have a substantial effect in small samples (Ruppert 1991).”
Stochastic Subgradient Method Converges on Tame Functions (2018, methodology): “Results of this type under more stringent assumptions, and with similar arguments, have previously appeared for example in Bena¨ım-Hofbauer-Sorin [2, 3] and Borkar [5].”
A companion for the kiefer-wolfowitz-blum stochastic approximation algorithm (2006, methodology): “The well known approach used to obtain optimal convergence rates for stochastic approximation algorithms without a tedious condition on the step size is to use the averaging principle independently introduced by Ruppert [32] and Polyak [28].”