Deep Multi-Species Embedding

Chen, Di; Xue, Yexiang; Chen, Shuo; Fink, Daniel; Gomes, Carla P.

doi:10.48550/arxiv.1609.09353

Cited by 5 publications

(6 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Interest in joint species distribution modeling with neural networks has only grown as deep learning has come to maturity [89]. Convolutional neural networks in particular have created a new opportunity: the ability to extract features from spatial arrays of environmental features [43,51] instead of using hand-selected environmental feature vectors.…”

Section: Machine Learning Methodsmentioning

confidence: 99%

“…There is a considerably amount of domain knowledge and ecological theory which would ideally be incorporated into SDMs [85]. This might include knowledge about species dispersal [25,52,72,126], spatial patterns of community composition [44,49,103], and constraints on species ranges (e.g. cliffs, water) [47,65,69,126].…”

Section: Incorporating Ecological Theory and Expert Knowledgementioning

confidence: 99%

See 1 more Smart Citation

Species Distribution Modeling for Machine Learning Practitioners: A Review

Beery

Cole

Parker

et al. 2021

ACM SIGCAS Conference on Computing and Sustainable Societies (COMPASS)

View full text Add to dashboard Cite

Figure 1. Species distribution models describe the relationship between environmental conditions and (actual or potential) species presence. However, the link between the environment and species distribution data can be complex, particularly since distributional data comes in many different forms. Above are four different sources of distribution data for the Von Der Decken's Hornbill [11]: (from left to right) raw point observations, regional checklists, gridded ecological surveys, and data-driven expert range maps. All images are from Map of Life [101].

show abstract

Section: Machine Learning Methodsmentioning

confidence: 99%

Section: Incorporating Ecological Theory and Expert Knowledgementioning

confidence: 99%

Species Distribution Modeling for Machine Learning Practitioners: A Review

Beery

Cole

Parker

et al. 2021

ACM SIGCAS Conference on Computing and Sustainable Societies (COMPASS)

View full text Add to dashboard Cite

show abstract

“…Species coexistence in nature follows complex, unknown patterns that Machine Learning should be able to capture thanks to its high degree of expressivity (i.e. the capacity of a model to express complex relations) (Balamurugan et al, 2019;Chen et al, 2017;Harris, 2015;Raghu et al, 2017;Tang et al, 2018). Here, we ex- G maximizes the likelihood that the discriminative model makes a mistake (Goodfellow et al, 2014).…”

Section: Introductionmentioning

confidence: 99%

A deep Generative Artificial Intelligence system to predict species coexistence patterns

et al. 2022

View full text Add to dashboard Cite

Predicting coexistence patterns is a current challenge to understand diversity maintenance, especially in rich communities where these patterns' complexity is magnified through indirect interactions that prevent their approximation with classical experimental approaches. We explore cutting‐edge Machine Learning techniques called Generative Artificial Intelligence (GenAI) to predict species coexistence patterns in vegetation patches, training generative adversarial networks (GAN) and variational AutoEncoders (VAE) that are then used to unravel some of the mechanisms behind community assemblage. The GAN accurately reproduces real patches' species composition and plant species' affinity to different soil types, and the VAE also reaches a high level of accuracy, above 99%. Using the artificially generated patches, we found that high‐order interactions tend to suppress the positive effects of low‐order interactions. Finally, by reconstructing successional trajectories, we could identify the pioneer species with larger potential to generate a high diversity of distinct patches in terms of species composition. Understanding the complexity of species coexistence patterns in diverse ecological communities requires new approaches beyond heuristic rules. Generative Artificial Intelligence can be a powerful tool to this end as it allows to overcome the inherent dimensionality of this challenge.

show abstract

“…Multi-label learning (MLL) learns from examples each associated with multiple labels simultaneously and aims to derive a predictive model which can assign a set of relevant labels for the unseen instance [30,44]. During the past decade, multi-label learning has been widely employed to learn from data with rich semantics, such as multimedia content annotation [38,33], text categorization [29,27], music emotion analysis [21,35], and bioinformatics analysis [3], etc. However, in practice, obtaining ground-truth labels for training datasets is costly due to the expensive and time-consuming manual annotations.…”

Section: Introductionmentioning

confidence: 99%

One Positive Label is Sufficient: Single-Positive Multi-Label Learning with Label Enhancement

Xu¹,

Qiao²,

Lv³

et al. 2022

Preprint

View full text Add to dashboard Cite

Multi-label learning (MLL) learns from the examples each associated with multiple labels simultaneously, where the high cost of annotating all relevant labels for each training example is challenging for real-world applications. To cope with the challenge, we investigate single-positive multi-label learning (SPMLL) where each example is annotated with only one relevant label and show that one can successfully learn a theoretically grounded multi-label classifier for the problem. In this paper, a novel SPMLL method named SMILE, i.e., Single-positive MultIlabel learning with Label Enhancement, is proposed. Specifically, an unbiased risk estimator is derived, which could be guaranteed to approximately converge to the optimal risk minimizer of fully supervised learning and shows that one positive label of each instance is sufficient to train the predictive model. Then, the corresponding empirical risk estimator is established via recovering the latent soft label as a label enhancement process, where the posterior density of the latent soft labels is approximate to the variational Beta density parameterized by an inference model. Experiments on benchmark datasets validate the effectiveness of the proposed method.Preprint. Under review.

show abstract

Deep Multi-Species Embedding

Cited by 5 publications

References 15 publications

Species Distribution Modeling for Machine Learning Practitioners: A Review

Species Distribution Modeling for Machine Learning Practitioners: A Review

A deep Generative Artificial Intelligence system to predict species coexistence patterns

One Positive Label is Sufficient: Single-Positive Multi-Label Learning with Label Enhancement

Contact Info

Product

Resources

About