- The 9th FinVolution Global Data Science Competition focuses on deepfake audio detection, tackling the challenge of distinguishing cloned voices from real voices in the LLM era.
- This competition is part of the IJCAI 2024 Competitions and Challenges track, which fosters global collaboration and innovation among AI enthusiasts.
- By integrating fake voices generated by LLM into the test dataset, the competition increases complexity and fosters innovation.
Shanghai, May 9, 2024 /PRNewswire/ — FinVolution, a leading fintech service provider, launches nine servicesth The FinVolution Global Data Science Competition, which takes place today, focuses on “Deepfake Audio Detection.” This competition is part of the competition and challenge track of his IJCAI (International Joint Conference on Artificial Intelligence), the top international conference in AI.
As speech synthesis technology continues to evolve, the line between cloned and authentic voices becomes increasingly blurred in the era of large language models (LLMs), creating significant challenges for data security and asset protection. Masu.
The competition aims to inspire AI enthusiasts and experts around the world to innovate in the fight against voice cloning and deepfake fraud. Contestants will utilize deep learning technology to develop models and algorithms based on FinVolution's test dataset. The competition includes LLM-generated fake voices to increase complexity and encourage innovation.
The total prize money is 310,000 RMB, the competition consists of preliminaries, semi-finals, and finals, and aims to authenticate the real voice and the voice behind it. The highest ranked contestants will participate in IJCAI 2024 to be held in Jeju Island. South Korea, receive awards, and interact with experts from academia and industry. FinVolution is proudly sponsoring IJCAI 2024.
Lee Cheol-jung“Since its inception nine years ago, the FinVolution Global Data Science Competition has evolved into a widely recognized event in the field of data technology, fostering technology exchange around the world,” said FinVolution CEO. We are partnering with IJCAI at the International AI Conference to demonstrate our commitment to advancing deep speech recognition technology.”
Challenge to deep fakes
During qualifying (May 10th to June 12th), participants will design algorithms based on white-box datasets provided by FinVOlution and submit their scoring results to qualify for the semi-finals. The dataset primarily consists of audio recordings totaling 20 to 40 hours.
At the semi-final stage (From June 13th to June 28th), contestants are expected to improve their algorithms based on black-box datasets provided by the contest organizers and compete for a spot in the finals. This dataset mainly consists of personal data and contains 5 to 10 hours of recordings.
Participants can register from the official website. May 9th – June 3rddownload and view the dataset.
Complying with AI ethics
Voice cloning is emerging as a major form of communication scamas fraudsters exploit AI technology to make it increasingly difficult to distinguish between real and fake voices.
This contest focuses on protecting user privacy and combating fraud by accurately identifying cloned voices.
Lei Chen, vice president and head of big data and AI at FinVolution, said: “The application of large-scale language models has far exceeded the corresponding detection technologies and poses great challenges to information security. We look forward to seeing the continued use of AI deepfake voice detection technology.” Keep pace with LLM development and protect public data security. With this concept in mind, the FinVoluting Global Data Science Competition is not only a platform for technology competition, but also an opportunity to explore how AI can better adhere to ethical principles. and serve the public. ”
To date, the FinVolution Global Data Science Competition has attracted a total of nearly 10,000 participants worldwide, making it a well-known event in the field of digital financial technology.
Held annually since 2016, the contest themes span a diverse range of disciplines, all rooted in real-world fintech business scenarios. These topics range from risk management algorithms, financial data applications, and product development to semantic similarity recognition, asset portfolio cash flow forecasting, and credit facilities for small and medium-sized businesses.
About FinVoltion Group
FinVolution Group is a leading fintech company that connects millions of consumers and small and medium-sized businesses with financial institutions.
Founded in 2007 and listed on the New York Stock Exchange in 2017, we are at the forefront of the credit technology industry across Asia and have pioneered innovative technologies in credit risk assessment. scam Discovery, big data and artificial intelligence. With a track record of strong growth across pan-Asian countries, we have established a world-leading fintech platform. China, Indonesiaand Philippines.
Source FinVolution Group
