Skip to content

Autoencoders and PCA #129

@AmanGoyalNayan

Description

@AmanGoyalNayan
  • Abstract (2-3 lines)
    Autoencoders are one of the generative models. They work on unsupervised algorithm. Autoencoders are mainly used for dimensionality reduction , image denoising , image generation and also recommendation systems.

  • Brief Description and Contents to be covered
    I would be covering the following topics :
    -> Dimensionality Reduction Problem
    -> PCA
    -> Autoencoders
    -> Latent Space
    -> PCA vs Autoencoders
    -> Types of Autoencoders
    -> Applications of Autoencoders
    -> Implementation : Image Denoising of Facial Images Dataset using Autoencoders

  • Pre-requisites for the talk

  • Basic concepts of Neural Networks should be clear

  • Basic knowledge of python and Keras is also required.

  • Time required for the talk
    60 - 70 mins

  • Link to slides
    https://bit.ly/DLQAutoencoders

  • Will you be doing hands-on demo as well?
    Yes

  • Link to ipython notebook (if any)
    https://bit.ly/Autoencoders

  • About yourself
    I am currently pursuing B.Tech in CS , 3rd year at NIT Kurukshetra. I am an AI enthusiast.
    Also last year , I had done a deep learning internship after which I was selected as Google AI Explore ML Facilitator due to which I had taken 10 sessions in my college which was attended by over 250 students where I had taught them various ML and DL concepts.
    Also recently during Quarantine times , I have been taking Deep Learning and Computer Vision webinars.
    Currently I am doing a 6 month internship as a DLCV intern at Nayan Technologies which is one of the leading startups and has collabration with Dubai Police.
    Also I was recently selected by Omdena as a Junior ML Engineer where I would be working on analysing the domestic violence and online harrasment trends during COVID-19.
    I have also written a research paper on Multimodal Emotion Recognition which I would be submitting for publication in one of the conferences.

  • Are you comfortable if the talk is recorded and uploaded to PyData Delhi's YouTube channel ?
    Yes Sure.

  • Any query ?
    No

Metadata

Metadata

Assignees

No one assigned

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions