Hack Session: Content Based Recommender System using Transfer Learning

Nov 15, 2019


Auditorium 3

50 minutes

Computer Vision

Identifying products a specific customer likes most can significantly increase the earnings of a company. For that purpose, a neural network classifier is used as a data-driven, visually-aware feature extractor. The latter then serves as input for similarity-based recommendations using a ranking algorithm. Combined with more traditional content-based recommendation systems, image-based recommendations can help to increase robustness and performance, for example, by better matching a particular customer style. In this hack session, learn how to build content based recommender systems using image data.


Key Takeaways:

Understanding of Recommender Systems

  • Collaborative Systems.
  • Content Based Recommender Systems.

Deep Learning Algorithms for Unsupervised Computer Vision

  • Convolutional Neural Networks (Convolution, MaxPooling, BatchNorm)
  • Transfer Learning for CNN Architectures
    • Inception Models
    • RESNET Models
    • VGG Models

Understanding Similarity Measures

  • Euclidean Distance measures
  • Cosine Similarity measures

Building an End to End Content Based Recommender System

  • Image and Data Preprocessing techniques
  • Creating a Deep Learning model
  • Apply Hyperparameter tuning to improve model accuracy
  • Create a Flask based Web App for showcasing Shirts
  • Running the end to end demo.


Check out the below video to know more about the session.

  • Sitaram Tadepalli

    Data Scientist


Copyright 2019 Analytics Vidhya. All rights reserved