Variational AutoEncoder For Regression: Application to Brain Aging Analysis

Zhao, Qingyu; Adeli, Ehsan; Honnorat, Nicolas; Leng, Tuo; Pohl, Kilian M.

doi:10.48550/arxiv.1904.05948

Cited by 3 publications

(4 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…AND ADVERSARIAL EXAMPLES DETECTION The detection algorithm proposed in this paper extends the work in [7] by using a VAE-based regression model [12]. The method is based on an LEC architecture which integrates the regression model into the VAE and uses inductive conformal anomaly detection [10].…”

Section: Variational Autoencoder For Regressionmentioning

confidence: 99%

“…The objective is to model the relationship between the observation x and the low-dimensional latent variable z. The architecture presented in [12] integrates a regression model into the generative model to disentangle the regression target variable from the latent space. The architecture of the VAE-based regression model is shown in Fig.…”

Section: A Variational Autoencoder For Regressionmentioning

confidence: 99%

See 1 more Smart Citation

Detecting Adversarial Examples in Learning-Enabled Cyber-Physical Systems using Variational Autoencoder for Regression

Cai

Koutsoukos

2020

Preprint

View full text Add to dashboard Cite

Learning-enabled components (LECs) are widely used in cyber-physical systems (CPS) since they can handle the uncertainty and variability of the environment and increase the level of autonomy. However, it has been shown that LECs such as deep neural networks (DNN) are not robust and adversarial examples can cause the model to make a false prediction. The paper considers the problem of efficiently detecting adversarial examples in LECs used for regression in CPS. The proposed approach is based on inductive conformal prediction and uses a regression model based on variational autoencoder. The architecture allows to take into consideration both the input and the neural network prediction for detecting adversarial, and more generally, out-of-distribution examples. We demonstrate the method using an advanced emergency braking system implemented in an open source simulator for self-driving cars where a DNN is used to estimate the distance to an obstacle. The simulation results show that the method can effectively detect adversarial examples with a short detection delay.

show abstract

Section: Variational Autoencoder For Regressionmentioning

confidence: 99%

Section: A Variational Autoencoder For Regressionmentioning

confidence: 99%

Detecting Adversarial Examples in Learning-Enabled Cyber-Physical Systems using Variational Autoencoder for Regression

Cai

Koutsoukos

2020

Preprint

View full text Add to dashboard Cite

show abstract

“…In [30], a Variational Autoencoder (VAE) was used to synthesise aged brain images, but the target age is not controlled, and the quality of the synthesised image appears poor (blurry). Similarly, [31] from temporal progression, then used the first few layers of the VAE as feature extractor to improve the age prediction task. In summary, most previous methods either built atlases [1],…”

Section: Related Workmentioning

confidence: 99%

“…[2], [5], [26], or required longitudinal data [3], [15], [27] to simulate the brain ageing process. Other methods that did not need longitudinal data [4], [30], [31], on the other hand, produced blurry images and lost subject identity.…”

Section: Related Workmentioning

confidence: 99%

Learning to synthesise the ageing brain without longitudinal data

Xia,

Chartsias,

Wang

et al. 2019

Preprint

View full text Add to dashboard Cite

Brain ageing is a continuous process that is affected by many factors including neurodegenerative diseases. Understanding this process is of great value for both neuroscience research and clinical applications. However, revealing underlying mechanisms is challenging due to the lack of longitudinal data. In this paper, we propose a deep learning-based method that learns to simulate subject-specific brain ageing trajectories without relying on longitudinal data. Our method synthesises aged images using a network conditioned on two clinical variables: age as a continuous variable, and health state, i.e. status of Alzheimer's Disease (AD) for this work, as an ordinal variable. We adopt an adversarial loss to learn the joint distribution of brain appearance and clinical variables and define reconstruction losses that help preserve subject identity. To demonstrate our model, we compare with several approaches using two widely used datasets: Cam-CAN and ADNI. We use ground-truth longitudinal data from ADNI to evaluate the quality of synthesised images. A pre-trained age predictor, which estimates the apparent age of a brain image, is used to assess age accuracy. In addition, we show that we can train the model on Cam-CAN data and evaluate on the longitudinal data from ADNI, indicating the generalisation power of our approach. Both qualitative and quantitative results show that our method can progressively simulate the ageing process by synthesising realistic brain images. The code will be made publicly available at https://github.com/xiat0616/BrainAgeing.

show abstract

Automatic evaluation-feedback system for automated social skills training

Saga

Tanaka

Matsuda

et al. 2023

Sci Rep

View full text Add to dashboard Cite

Social skills training (SST), which is a rehabilitation program for improving daily interpersonal communication, has been used for more than 40 years. Although such training’s demand is increasing, its accessibility is limited due to the lack of experienced trainers. To tackle this issue, automated SST systems have been studied for years. An evaluation-feedback pipeline of social skills is a crucial component of an SST system. Unfortunately, research that considers both the evaluation and feedback parts of automation remains insufficient. In this paper, we collected and analyzed the characteristics of a human–human SST dataset that consisted of 19 healthy controls, 15 schizophreniacs, 16 autism spectrum disorder (ASD) participants, and 276 sessions with score labels of six clinical measures. From our analysis of this dataset, we developed an automated SST evaluation-feedback system under the supervision of professional, experienced SST trainers. We identified their preferred or most acceptable feedback methods by running a user-study on the following conditions: with/without recorded video of the role-plays of users and different amounts of positive and corrective feedback. We confirmed a reasonable performance of our social-skill-score estimation models as our system’s evaluation part with a maximum Spearman’s correlation coefficient of 0.68. For the feedback part, our user-study concluded that people understood more about what aspects they need to improve by watching recorded videos of their own performance. In terms of the amount of feedback, participants most preferred a 2-positive/1-corrective format. Since the average amount of feedback preferred by the participants nearly equaled that from experienced trainers in human–human SSTs, our result suggests the practical future possibilities of an automated evaluation-feedback system that complements SSTs done by professional trainers.

show abstract

Variational AutoEncoder For Regression: Application to Brain Aging Analysis

Cited by 3 publications

References 6 publications

Detecting Adversarial Examples in Learning-Enabled Cyber-Physical Systems using Variational Autoencoder for Regression

Detecting Adversarial Examples in Learning-Enabled Cyber-Physical Systems using Variational Autoencoder for Regression

Learning to synthesise the ageing brain without longitudinal data

Automatic evaluation-feedback system for automated social skills training

Contact Info

Product

Resources

About