k-DPPs: Fixed-Size Determinantal Point Processes

By Alex Kulesza, B. Taskar. 2011

317 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.

Citations in the corpus

Determinantal Point Processes for Machine Learning (2012, methodology): “3 Experiments: Image Search We demonstrate the use of k-DPPs on an image search task [84].”
Diversity Networks (2015, neutral): “The key ingredient that remains to be specified is the DPP kernel, which we now describe.”
INSIDE: LLMs' Internal States Retain the Power of Hallucination Detection (2024, methodology): “The eigenvalue or determinant of covariance matrix captures the variability of the data and has been widely explored as divergence measure in a wide range of machine learning tasks (Wold et al., 1987; Kulesza & Taskar, 2011; Xu et al., 2021; Zhouyin & Liu, 2021; Cai et al., 2015).”
Learning Everything about Anything: Webly-Supervised Visual Concept Learning (2014, background): “This formulation searches for a subset of ngrams that are visually manageable (have reliable ngram classifiers) and cover the space of variance within a concept (similar to [3, 22]).”
Optimal Transport Kernels for Sequential and Parallel Neural Architecture Search (2020, methodology): “A k-determinantal point process (k-DPP) [27] is a distribution over all subsets of a ground set Pt of cardinality k.”
Fair and Diverse DPP-based Data Summarization (2018, methodology): “2 Following [17], each image was processed with the vlfeat toolbox to obtain sets of 128dimensional SIFT descriptors [21, 26].”
Fast Greedy MAP Inference for Determinantal Point Process to Improve Recommendation Diversity (2017, methodology): “Besides its early applications in quantum physics and random matrices [35], it has also been recently applied to various machine learning tasks such as multiple-person pose estimation [27], image search [28], document summarization [29], video summarization [19], product recommendation [18], and twe”
Diverse Sequential Subset Selection for Supervised Video Summarization (2014, methodology): “Our focus on the determinantal point process as the building block is largely inspired by its appealing property in modeling diversity in subset selection, as well as its success in search and ranking [17], document summarization [14], news headline displaying [28], and pose estimation [29].”
Balanced Mini-batch Sampling for SGD Using Determinantal Point Processes (2017, methodology): “It assigns probabilities to subsets of size k , Apart from conditioning on the size of the subset of points, the k -DPP has the same diversiﬁcation effect as the DPP [18].”
Near-Optimal MAP Inference for Determinantal Point Processes (2012, methodology): “More recently, they have been applied as probabilistic models for machine learning problems [10, 13, 5, 11].”
Stochastic Learning on Imbalanced Data: Determinantal Point Processes for Mini-batch Diversification (2017, methodology): “Therefore, a slightly modified version of the DPP is needed, which is called k-DPP [18].”
Deep Batch Active Learning by Diverse, Uncertain Gradient Lower Bounds (2019, background): “As mentioned earlier, our sampling criterion has resemblance to sampling from k-determinantal point processes (k-DPPs) [5].”
Predicting Visual Exemplars of Unseen Classes for Zero-Shot Learning (2016, methodology): “Also, for each of these numbers, we explore the following subset selection strategies (more details are in the supplementary material): (i) Uniform random: Randomly selected B unseen classes from the uniform distribution; (ii) Heavy-toward-seen random Randomly selectedB classes that are semantically”
DLow: Diversifying Latent Flows for Diverse Human Motion Prediction (2020, background): “ons [7] and multiple choice learning [28,44]. Alternatively, submodular function maximization has been applied to select a diverse subset of garments from fashion images [31]. Another type of methods [40,20,19,32,5,74,70] seeks diversity using determinantal point processes (DPPs; [50,41]) which are ”
Effective Diversity in Population-Based Reinforcement Learning (2020, background): “ states but using different mechanisms. Random corresponds to uniform sampling, Zero to a zero function trained on all seen states, where we select the maximum variance states. DPP stands for a k-DPP [20]. As we see in Fig. 6, both the number of states chosen and the mechanism for choosing them appe”
Constrained Dominant Sets for Multimodal Document Question Answering (2026, unknown): “-DPP 45 on } k -DPP [ ] selects a size-B subset by greedy MAP L ij = q i q j K ij , where q i = max(cos( z q , z i ) , 0) and K ij = cos( z i , z j ) , thereby promoting diversity through determinantal volume rather than query-anchored repulsion.”