Master Thesis - Machine Learning

Leveraging transfer learning for multi-label attribute prediction in fashion images, using deep learning

 

About Babyshop Group

E-commerce is at the heart of Babyshop Group where premium childrens’ fashion is retailed on a multitude of websites (eg. lekmer.se, babyshop.se, alexandalexa.se). Babyshop Group has had continuous growth in the past years, currently selling and shipping internationally with a revenue of more than 1.5 billion SEK. In order to sustain such growth, Babyshop Group will continue to work in a data-driven environment, which you hopefully can be a part of! We at Babyshop work with machine learning (ML) on a daily basis, whether it being exploratory data analysis and modelling in a Jupyter Notebook or building ML-pipelines integrating cloud solutions. If this is something that you want to develop within and you find the thesis proposal interesting, apply! By the way, at our five story office in Karlaplan, you will find lots of tasty coffee, tea, and inspiring individuals. 

Thesis proposal

At Babyshop Group, there exist tens of thousands of images of children’s apparel. Currently, new items are labeled by hand, where each item has multiple attributes, eg. color, category, pattern, neckline, and style, etc. Since the workload of this manual task is immense, a need for a support system is high. In literature, a multitude of previous work has been done in regards to multi-label attribute prediction on fashion clothing, most notably DeepFashion and iMaterialist Fashion. Here, deep learning models have been trained on hundreds of thousands of labeled images to predict corresponding attributes, which has proven highly successful. However, almost all of these images are of adult fashion, with very little emphasis on childrens’ fashion.  

Recently, transfer learning has proved very successful in multiple areas, most prominently in regards to object-detection in images, using YoloV3, as well as natural language processing using BERT and GPT. Because of this, ML-practitioners with access to limited sets of labeled data can successfully leverage these transfer learning frameworks that are trained on immense datasets.

The question then arises, is it possible to combine the pre-trained models from previously mentioned multi-label attribution prediction models that are trained on almost exclusively adults’ fashion, together with the learnings from transfer-learning frameworks - in order to successfully perform multi-label attribution prediction on childrens’ fashion?

Qualifications

For a good fit we believe it to be preferable if you have the following qualifications:

  • Students of Computer Science, Machine Learning, Computer Engineering, Mathematics, Physics, or related field (e.g. applied mathematics/statistics)
  • A good understanding of the use of some of the current state of the art machine learning frameworks such as Keras, TensorFlow, PyTorch, Scikit-Learn, Spark etc. in developing ML applications
  • Proficiency in Python and Git
  • Knowledge about data wrangling and data munging, using SQL, Pandas, and Numpy
  • Good communication skills in written and spoken English.

What we offer

  • Supervision from experienced data scientists and ML practitioners
  • The compute-resources needed for such a modelling-heavy task
  • 1000 SEK per higher education credit as compensation per student
  • Inspiring culture and fun after-works!

 

Starting, early January till mid-June

How to apply: You can apply both individually and as pairs, but if you apply as a pair, please attach Resumes from both candidates, as well as a jointly written Cover Letter in the application form below.

Last application date is 30/11, but please note that continuous selection is carried out

Contact: For questions please contact Marcus Svensson, Data Scientist at Babyshop Group marcus.svensson@babyshop.se

Apply

The link to your LinkedIn profile can be found here

Apologies, your application wasn't recieved correctly due to what seems like a malformed file. Make sure the file is not corrupt and is of common format, preferably PDF.

privacy policy