Post Doc

**About the role**:
Context: The field of Weakly Supervised Learning (WSL) has recently seen a surge of popularity, with numerous papers addressing different types of ''supervision deficiencies'', namely: poor quality, non-adaptability, and insufficient quantity of labels. Regarding quality, label noise can be of different kinds, including completely-at-random, at-random or even not-at-random). Label noise affect the quality of the models trained. Detecting label noise allow to obtain better model. However, we do not have a clear view of the state of the art of methods able to detect instances which have a wrong label.

Scientific objective - results and obstacles to overcome: The aim of the postdoc is to do a review of the methods of the literature. One of the objectives will be to publish this review and /or a white paper. The second objective could be to develop new methods adapted to biquality learning and to (i) to have a tool for detecting erroneous labels and/or automate data cleansing and to (ii) be able to learn with clean data + data with erroneous labels.

The main issues will be to (i) identify the best state-of-the-art methods (strengths, weaknesses, maturity,...),; (ii) to do the difference between methods which detect the noise ratio and those which detect the noisy samples (the later is harder).

**About you**:

- You have a very good knowledge of AI model development
- You are autonomous and take initiative.
- You enjoy working in a team.
- You have a PhD in Artificial Intelligence or Data Science obtained within the last two years (have completed this PhD in relation to the postdoc topic)
- Mastery of the Python language
- Good knowledge of machine learning and data science algorithms.

**Additional information**:
You will be part of a team of research engineers. The department you will join is at the forefront of innovation and expertise in the field of machine learning. The Post-Doc may lead to industrial (patents) or scientific (publications) valorization activities depending on the results obtained.

**Department**:
Orange Innovation brings together the research and innovation activities and expertise of the Group's entities and countries. We work every day to ensure that Orange is recognized as an innovative operator by its customers, and we create value for the Group and the Brand in each of our projects. With 740 researchers, thousands of marketers, developers, designers and data analysts, it is the expertise of our 6,000 employees that fuels this ambition every day. Orange Innovation anticipates technological breakthroughs and supports the Group's countries and entities in making the best technological choices to meet the needs of our consumer and business customers.

Within Innovation, you will be integrated into a research team at the forefront of innovation and expertise and know-how in knowledge extraction from data. The department provides tools and algorithms based on predictive analysis techniques. In particular, it carries the activities of automating the valorization of data. You will be part of a research ecosystem, working alongside research engineers as well as design engineers (and data scientists) in anticipation (shorter term) allowing the concrete implementation of the concepts studied.

**Contract**:
Post Doc