Near-optimal sensor placements in Gaussian processes

By Andreas Krause, Amarjeet Singh, Carlos Guestrin. 2005

1,916 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Signal Processing

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 Signal Processing

Citations in the corpus

Super-Samples from Kernel Herding (2010, positive): “This is a significant reduction whenever evaluatingf is expensive, e.g. for user interaction data.”
Parallelizing Exploration-Exploitation Tradeoffs with Gaussian Process Bandit Optimization (2012, methodology): “A similar idea was also employed by Krause et al. (2008) in the Gaussian process setting for experimental design.”
Bayesian Active Learning for Classification and Preference Learning (2011, neutral): “[Guestrin et al., 2005, Krause et al., 2006] also use mutual information.”
Non-monotone submodular maximization under matroid and knapsack constraints (2009, background): “We note that the entropy function is not even approximately monotone (see [30]).”
Safe Exploration for Active Learning with Gaussian Processes (2015, neutral): “This entropy criterion has been frequently used in the active learning literature, e.g. in Seo et al. (2000) or Guestrin et al. (2005).”
Gaussian Processes for Object Categorization (2010, neutral): “A similar observation has also been made by Krause et al. (2008) .”
Online Sparse Multi-Output Gaussian Process Regression and Learning (2019, positive): “GP regression has found diverse applications in indoor localization [3]–[5], image processing and visual tracking [6], [7], change point detection [8] and sensor (robotic) network management [9], [10].”
Gaussian process dynamic programming (2009, background): “…active learning, the expected utility is considered by averaging over possible outcomes.12 Information-based criteria as proposed by MacKay (1992), Krause et al. (2008), and Pfingsten (2006), for example, or their combination with expected outcomes as discussed by Verdinelli and Kadane (1992) and…”
Learning via Hilbert Space Embedding of Distributions (2007, neutral): “Besides the above successes of this embedding approach, we can actually replace information theoretic quantities in many other scenarios as well (for instance, in sensor placement problem [63]).”
Learning Graphs From Data: A Signal Representation Perspective (2018, background): “For example, in large-scale social, biomedical or environmental networks, sampling or active learning may have to be applied to select a limited number of sensors for observations [87], [88].”
Examples are not enough, learn to criticize! Criticism for Interpretability (2016, methodology): “We found the best qualitative performance using the log-determinant regularizer (Krause et al., 2008).”
Near-optimal Nonmyopic Value of Information in Graphical Models (2005, positive): “The analogous subset selection problem for the information gain is to compute For example, we can model our temperature measurement task as indicated in by the graphical model in Fig.”
Bounded approximate decentralised coordination via the max-sum algorithm (2009, neutral): “20This measure was chosen because it has been often used in related work to ascertain the accuracy of sensor predictions [37, 38].”
Data-Driven Sparse Sensor Placement for Reconstruction: Demonstrating the Benefits of Exploiting Known Patterns (2017, neutral): “[4] A.”
Beyond Distributive Fairness in Algorithmic Decision Making: Feature Selection for Procedurally Fair Learning (2018, background): “…maximum likelihood 4Similar problems have previously been studied in the context of feature selection (Iyer and Bilmes 2012), sensor placement (Krause, Singh, and Guestrin 2008), diversification (Ahmed, Dickerson, and Fuge 2017; Ashkan et al. 2015), document summarization (Lin and Bilmes…”
Online Submodular Maximization with Preemption (2015, methodology): “In machine learning, maximization of submodular functions has been used for document summarization [46, 47], sensor placement [41, 42, 44], and information gathering [43].”
Submodular Function Maximization (2014, methodology): “Krause et al. (2008c) develop Saturate , an algorithm that is guaranteed to eﬃciently obtain a set S such that f min ( S ) ≥ OPT k , and | S | ≤ 1 + max v ln i f i ( { v } ) k .”
Sparse Dictionary Learning for Edit Propagation of High-Resolution Images (2014, result): “This demonstrates that the use of KL-divergence tends to capture more representative samples than other alternatives of information entropy, such as MMI.”