Thales Student Innovation Championship
The Thales Competition is intended to give students exposure to problems that businesses face every day and to encourage you to come up with innovative ways of solving them. It intends to encourage healthy competition, professionalism and interaction between themselves and with the industry professionals they would one day like to work with.
Throughout human history, any topic that arose usually triggered a number of diverging opinions. Those opinions can be argued and defended with the support of evidences brought forward by their champions. In order to reach a consensus, every opinion has to be taken into account and every piece of evidence has to be collected and analyzed carefully so as to validate or dismiss it. A process which can be easily biased as the number of participants grows.
Nowadays, this complexity is amplified with the internet spreading an unprecedented volume of information at a speed and easiness ever greater reaching an increasing number of people.
The challenge, for a given topic provided by the end-user, is twofold:
- Collecting all available opinions and their supporting evidences
- Validating every piece of evidence, weighting its contribution to the topic and explaining this contribution
Although it will never be possible to crawl all data ever expressed around the world on any topic, it is possible to design a crawling engine ensuring the completeness of the pool of opinions and supporting evidences, pertaining to a specific topic, available on the internet.
More importantly, with the recent progresses made in the field of Artificial Intelligence, it should be possible to design an AI capable of automatically finding the opinion clusters and analyzing a piece of evidence in order to:
- Classify it with respect to the expressed opinions
- Explain why this piece of evidence is aligned with the opinion(s) it has been assigned to
- Show which pieces of evidence are in contradiction with it
- Explain why those pieces of evidence are in contradiction with it
The winning submission will:
- Provide a clear explanation of the architecture that could be employed to determine all different pieces of evidence (e.g. articles, images, videos, text, etc.) related to a topic.
- Describe innovative algorithms or applications of algorithms that may be capable of grouping those pieces of evidence into different opinions.
- Describe innovative algorithms or applications of algorithms that may allow the system to explain its decision-making process.
- Describe implementation options that could include preconfigured datasets, generated datasets and/or information gathering on the web.
- Describe a method to be able to visualize the decision-making process.