stp2y

arXiv:2405.17450v1 Announce Type: new Abstract: Next-frame prediction is a useful and powerful method for modelling and understanding the dynamics of video data. Inspired by the empirical success of causal language modelling and next-token prediction in language modelling, we explore the extent to which next-frame prediction serves as a strong foundational learning strategy (analogous to language modelling) for inducing an understanding of the visual world. In order to quantify the specific visual understanding induced by next-frame prediction, we introduce six diagnostic simulation video datasets derived from fundamental physical laws created by varying physical constants such as gravity and mass. We demonstrate…

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs. Source link lol

(iQoncept/Shutterstock) People are less likely to believe news written by generative AI, according to a recent public opinion survey conducted by Reuters, but the news would be timelier and cost less for the publisher to produce. The online survey also found GenAI tools are not being as widely adopted by consumers as some people may think. The Reuters Institute for the Study of Journalism (RISJ) at the University of Oxford contracted with YouGov to answer questions in two main areas: Determine how people are adopting GenAI broadly, and gauge their feelings towards GenAI-generated journalism specifically. The survey was conducted online…