Enhancing disease prediction using Deep Learning
Lieu de stage: SIMOB - IBISC (EA 4526) - Univ. Paris-Sacay (UEVE)
Salary an perspectives: According to background and experience (a minimum of 577.50 euros/month). Possibility to pursue with a 3-year-funded PhD contract with international research partners.
Since the last decade, interest for articial intelligence (AI) has grown thanks to recent advances in the eld of machine learning [1, 6]. Deep learning provides a new class of biological inspired methods (neural networks) which spectacularly outperforms most of the previously existing techniques in the field of computer vision. This internship investigates how to adapt an ecient deep learning architecture for biomedical images recognition and disease detection. The motivating application is an automatic system that can detect thrombus from brain scanner images.
The study focuses on convolutional neural networks (CNN) which have proven their efficiency in the field of computer vision . An effort will be made to find the best architecture for the proposed classication task of biomedical for disease prediction.
From a theoretical point of view, we expect to better understand the role of the architecture and of the optimized loss function to provide the best prediction results. Possible approaches to cope with an insufficient amount of data may use variational auto-encoder (VAE) , generative adversarial networks (GAN)  and data augmentation techniques .
An evaluation protocol with its corresponding testbench will be proposed to objectively compare an arbitrary number of techniques for a given annotated dataset. A comparison with a given baseline method is expected in order to validate the effectiveness of the new proposed enhancements measured in terms of recognition rate.
 Min Chen, Yixue Hao, Kai Hwang, Lu Wang, and Lin Wang. Disease prediction by machine learning over big data from healthcare communities. IEEE Access, vol 5, pages: 8869-8879, 2017.
 Carl Doersch. Tutorial on variational autoencoders. arXiv preprint arXiv :1606.05908, 2016.
 Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. In Advances in neural information processing systems, pages: 2672-2680, 2014.
 Bharath Raj. Data augmentation | how to use deep lear-
ning when you have limited data. https://medium.com/nanonets/
how-to-use-deep-learning-when-you-have-limited-data-part-2-data-augmentation-c26971dc8ced. Accessed : 18 septembre 2018.
 George Seif. Deep learning for image recognition : why it's challen-
ging, where we've been, and what's next. https://towardsdatascience.com/
deep-learning-for-image-classication-why-its-challenging-where-we-ve-been-and-what-s-next-93b56948fcef. Accessed : 18 septembre 2018.
 Ahelam Tikotikar and Mallikarjun Kodabagi. A survey on technique for prediction of disease in medical data. In Proc. IEEE International Conference on Smart Technologies For Smart Nation (SmartTech-Con), pages: 550-555, 2017.
(c) GdR 720 ISIS - CNRS - 2011-2018.