XAI - eXplainable Artificial Intelligence
Go beyond the traditional borders with eXplainable Artificial Intelligence
AI algorithms have the power to improve credit scoring and pricing capabilities of trade credit and surety insurers. However, our regulated industry will have to overcome the challenge of explainability to make use of the most accurate ones.
See what Tinubu has produced on XAI.
Here you will find the assets developed by the Innovation LAB Team. This list will be completed progressively with new articles when developed.
- Infographic
- Technical Paper
- Scientific publication
- Technical achievement (or Prototype)
/2022_visuals%20by%20topic/resources_LAB_TP_338x237px-1_WEBP.webp "Tinubu LAB - Technical Paper")
[White Paper excerpt]
Hire a XAI Underwriter
AI implementation has brought to light various issues and problems that must be addressed. There is a need for a regulatory framework and governance to ensure AI is secure, fair, and responsible. To achieve this, there has been research on eXplainable Artificial Intelligence (XAI) to create more transparent and understandable models. XAI can help governments, civil society organizations, and industries define what is required to achieve responsible AI from a technical perspective. This paper will describe XAI's main advantages and disadvantages and show how it can be applied to Tinubu's risk assessment activity.
[White Paper excerpt]
Extracting insights from News Feeds
Underwriters need more and more collecting information to assess the health of the companies in their portfolio between two publications of financial statements. This activity is time-consuming and requires analyzing unstructured data such as news feeds or social media content. This paper illustrates automating this task by leveraging Natural Language Processing to perform sentiment analysis.
[White Paper excerpt]
Simulating Risk Propagation
Value chains are more and more interconnected in the global economy. Therefore, a failure in one part of the globe will have consequences on other parts. Such supply chain disruption may result in insurers' claims, which are difficult to anticipate. This article shows how crisis impacts can be simulated thanks to the input-output matrix from the OECD.
[White Paper excerpt]
Predicting turnover evolution
The evolution of a company’s turnover over time represents what statisticians call a Time series. Statistical approaches and deep learning offer options to predict this indicator's future value based on the time series' past behavior. This paper explores various methodologies and compares their performance. It also explores the value of adding the evolution of Google Trends of the industry in which the company operates to improve predictions.
[White Paper excerpt]
Glossary of Artificial Intelligence & Quantum Computing
Related to the white paper 'The Augmented Underwriter'
For a better understanding of the technical terms used in the white paper, click here for direct access to explanations.
/2022_visuals%20by%20topic/resources_LAB_WP_338x237px_WEBP.webp "LAB white paper")
[White Paper]
The Augmented Underwriter: can statistics, artificial intelligence, and quantum computing empower further Credit Insurers?
This white paper presents a selection of works on Risk Management by the LAB Team. We strongly believe that Innovation is facilitated by collaboration. We are confident that this material will trigger your curiosity and make you want to contact the team to submit research topics or discuss research partnership opportunities.
Please fill in this form to receive all the available publications
Visit the other LAB-dedicated pages
Blockchain
Virtual Assistant
Quantum Computing
[Technical Paper]
Keep control of your AI by making it eXplainable.
While the pandemic has forced Insurers to quickly set up digital processes and abandon the “we’ve always done
it this way...” refrain, Artificial Intelligence (AI) can make their digital transformation more sustainable and secure, provided that it is Explainable.
So, let’s get XAI (eXplainable Artificial Intelligence).
"What prevents us from trying advanced technologies?”
/2022_visuals%20by%20topic/resources_LAB_infographic_338x237px_WEBP.webp "LAB infographic")
[Infographic]
Modern AI: Adopt it if you can explain it!
Step by step, discover in this infographic how new developments require revisiting the way to use and implement AI.
[Technical Paper]
Use Artificial Intelligence to generate code with natural language
How can you build a Simple Web Application with GPT-3 and Dash that Converts Natural Language into Source Code in 10 Minutes?
We are entering a new technological era. Transformers may reshape the way applications deeply will be developed in the coming years. Soon, most of the low-value code a developer produces might be generated by AI and Transformers.
This article mostly focuses on source code generation, with some bonus tracks showing great text generation capabilities. It comes along with the entire Python source code used to build this playground web app.
