Are you enthusiastic about AI and data engineering and want to learn the concepts of machine learning models that go beyond human labeling in areas like CV and NLP? At this event, you will learn more about these interesting topics. The event will be divided into two parts. Part 1: Can your model beat a human? Anatomy of Toloka’s WSDM’23 challenge (1 hour) Think you can prove that our WSDM challenge doesn't have to be hard for machines? Let's share ideas and refreshments and get ready to win! Humans are natural at CV and NLP tasks, but equivalent ML solutions require huge datasets, well-designed architectures, and plenty of time for training and fine-tuning. Are researchers close to finding ML models outperforming human labeling in fields like CV and NLP? Evgeniya will share how these questions were answered by the research team at Toloka and tell you about the challenges we took on when designing Toloka’s WSDM 2023 Cup challenge. Break for networking Part 2: Let’s get real with crowdsourcing: hands-on tutorial! (1 hour) A tutorial based on a real use case realized on Toloka. You will be able to choose the use case to run! Together with a Toloka Crowd Solution Architect you will brainstorm the pipeline and see which approaches work better for various projects. Then you will run a real project on Toloka and see how the crowdsourcing methods work in real life. This is a hybrid event. You can attend in person or follow the live stream.
Friday, November 25, 2022
5:00 PM – 7:15 PM UTC
Toloka.ai
Data Evangelist
Toloka.ai
Crowd Solution Architect
TUM
M.Sc. Computer Science
TUM
M.Sc. Computer Science
GDSC Lead
GDSC Lead
Core Team Member
Core Team Member
Core Team Member
Core Team Member
Core Team Member
Core Team Member
Core Team Member
TUM
Core Team Member
Core Team Member
Core Team Member