Deep Archetypal Analysis

Keller, Sebastian Mathias; Samarin, Maxim; Wieser, Mario; Röth, Volker

doi:10.1007/978-3-030-33676-9_12

Cited by 15 publications

(20 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To provide a fair comparison, we used 128 classical AA patterns as was used in DAA assessment. We used the implementation of Chen et al for AA [21] and used the implementations in [5], [7] for DAA analysis.…”

Section: Resultsmentioning

confidence: 99%

“…Drawback: AA effectively models the convex hull of data, however, it is limited in modeling either the average or local characteristics of data. To overcome these limitations, deep archetypal analysis has been proposed by our team and others, recently [5]- [7].…”

Section: A Archetypal Analysismentioning

confidence: 99%

“…This characteristic of DAA atoms can be attributed to the fact that dictionaries obtained at deeper layers (L > 1) are different convex combinations of archetypes obtained at the first layer. Since the combination of convex representations is also a convex representation [5], at any L'th layer (L > 1), a new convex representation matrix is obtained by factorizing A L1 using AA. Hence, atoms of these deeper dictionaries can lie anywhere on the data-spread including the boundary.…”

Section: B Deep Archetypal Analysismentioning

confidence: 99%

“…However, several variations of AA has been proposed in the literature [2]- [4]. Deep AA (DAA), as a nonlinear extension of the initial AA was recently proposed to address several limitations of the AA [5]- [7]. Deep AA does not rely on expert knowledge to combine relevant dimensions, learns appropriate transformations when combining features of different types and is able to incorporate additional information into the learning process.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Convex Representations Using Deep Archetypal Analysis for Predicting Glaucoma

Thakur

Goldbaum

Yousefi

2020

IEEE J. Transl. Eng. Health Med.

View full text Add to dashboard Cite

The purpose of this study was to identify clinically relevant patterns of glaucomatous vision loss through convex representation to predict glaucoma several years prior to disease onset. Methods: We developed a deep archetypal analysis to identify patterns of glaucomatous vision loss, and then projected visual fields over the identified patterns. Projections provided a representation that was more accurate in detecting glaucomatous vision loss, thus, more appropriate for recognizing preclinical signs of glaucoma prior to disease development. To overcome the class imbalance in prediction, we implemented a class-balanced bagging with neural networks. Results: Using original visual field as features of the class-balanced bagging classification provided an area under the receiver-operating characteristic curve (AUC) of 0.55 for predicting glaucoma approximately four years prior to disease development. Using convex representation of the visual fields as input features provided an AUC of 0.61 while using deep convex representation as input features improved the AUC to 0.71. Relevance vector machine (RVM) achieved an AUC of 0.64. Conclusion: Deep archetypal analysis representation of visual functional features with balanced bagging classification could serve as an automated tool for predicting glaucoma. Significance: Glaucoma is the second leading cause of worldwide blindness. Most people with glaucoma have no early symptoms or pain, delaying diagnosis in many patients until they reach late irreversible vision loss stages. In fact, about 50% of people with glaucoma are unaware they have the disease. Deep archetypal analysis models may impact clinical practice in effectively identifying at-risk glaucoma patients well prior to disease development.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: A Archetypal Analysismentioning

confidence: 99%

Section: B Deep Archetypal Analysismentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Convex Representations Using Deep Archetypal Analysis for Predicting Glaucoma

Thakur

Goldbaum

Yousefi

2020

IEEE J. Transl. Eng. Health Med.

View full text Add to dashboard Cite

show abstract

“…Further, no side information is incorporated which can-and in our opinion should-be used to constrain potentially over-flexible neural networks and guide the optimisation process towards learning a meaningful representation. The work presented here is an extension of a conference paper by Keller et al (2019). The extension highlights the wide scope of application of the proposed method by including extensive experiments on "Archetypes in Image-based Sentiment Analysis".…”

Section: Related Workmentioning

confidence: 93%

Learning Extremal Representations with Deep Archetypal Analysis

et al. 2020

Self Cite

View full text Add to dashboard Cite

Archetypes represent extreme manifestations of a population with respect to specific characteristic traits or features. In linear feature space, archetypes approximate the data convex hull allowing all data points to be expressed as convex mixtures of archetypes. As mixing of archetypes is performed directly on the input data, linear Archetypal Analysis requires additivity of the input, which is a strong assumption unlikely to hold e.g. in case of image data. To address this problem, we propose learning an appropriate latent feature space while simultaneously identifying suitable archetypes. We thus introduce a generative formulation of the linear archetype model, parameterized by neural networks. By introducing the distance-dependent archetype loss, the linear archetype model can be integrated into the latent space of a deep variational information bottleneck and an optimal representation, together with the archetypes, can be learned end-to-end. Moreover, the information bottleneck framework allows for a natural incorporation of arbitrarily complex side information during training. As a consequence, learned archetypes become easily interpretable as they derive their meaning directly from the included side information. Applicability of the proposed method is demonstrated by exploring archetypes of female facial expressions while using multi-rater based emotion scores of these expressions as side information. A second application illustrates the exploration of the chemical space of small organic molecules. By using different kinds of side information we demonstrate how identified archetypes, along with their interpretation, largely depend on the side information provided.

show abstract