Machine learning to resolve poor data ‘quality’
The use of machine learning comes up against ‘data scarcity’, i.e. poor data ‘quality’, which can only be exploited if it has been ‘labelled’ before and according to the analysis you want to carry out.
Explicability of ‘black box’ ai* models
Artificial intelligence models are used in a number of fields (e.g. granting credit, fraud detection) and they deliver very good results. However, they lack transparency and explicability, which hinders their adoption by business teams and management, as well as their use in the eyes of regulators.
Causal inference: the uplift
‘Uplift modelling’ is the name given to a set of scoring methods used to assess individual sensitivity to sales prospecting.
Incorporating climate risks into the value of property assets
Real estate is vulnerable to a wide range of climate risks (physical, regulatory, and economic) that can significantly reduce its value. This depreciation may go so far as to generate ‘stranded assets’.
AI* to help assess the impact of climate change on the activities in the financial sector
AI makes it easier to assess and anticipate financial climate risks (e.g. risks for insurers, credit risks, and/or market risks).
AI* and NLP* in the service of CSR
The proliferation of data (news, social networking, company reports, and press releases) has become a major challenge for CSR experts and decision makers in terms of extracting and interpreting key data.