Our Research Featured ReMix: Training Generalized Person Re-identification on a Mixture of Data Our WACV 2025 paper demonstrates that intelligently leveraging a mixture of heterogeneous data can significantly enhance the quality of generalized person Re-ID.
News Nvidia announced Blackwell B200 GPU: key points Nvidia has created chips for the new "Generative AI Era".
News Apple wants to integrate Google's Gemini into the iPhone Apple breaks into the artificial intelligence race, but risks being left behind.
Paper Review Paper Review: Matryoshka Representation Learning We can train the model with embeddings of different sizes without retraining.
Paper Review Featured Paper Review: TETRIS: Towards Exploring the Robustness of Interactive Segmentation It turns out that modern benchmarks incorrectly assess the robustness of interactive segmentation methods.
Paper Review How OpenAI's Sora works: key insights Work that will accelerate the development of text-to-video generative AI and create many problems with deepfakes.
Guides How to automatically summarize research papers using ChatGPT? ChatGPT can already help researchers speed up the process of choosing good scientific papers and reading them.
Guides How to run open source LLM locally without programming? With Jan, anyone can run a variety of open source LLMs on their local computer without any programming skills.
News OpenAI wants to license articles from a dozen publishers OpenAI has chosen the path to legalize data for training neural networks, similar to Apple.
Paper Review Paper Review: Weak-to-Strong Generalization by OpenAI OpenAI has made a significant step towards superintelligence, but there is still a long way to go.
Opinion What is new in AI to expect in 2024? Apparently, 2024 will be a breakthrough in the field of artificial intelligence. In this post, we describe the main directions of AI development in the new year.
News The Times sues OpenAI and Microsoft over copyright infringement by AI This is probably the first serious lawsuit concerning AI and copyright.
News Apple wants to buy publishers' articles to train generative AI Apple's approach differs significantly from that of other companies, which buy rights to content after training generative models.
Opinion Featured Why does Apple call neural networks machine learning? Interpretability and validity are key factors for Apple when choosing names for its products.
Paper Review Paper Review: Real-World Humanoid Locomotion with Reinforcement Learning Robots have learned to decide for themselves what and how to do it.
News Gemini is Google's answer to GPT-4: the revolution failed Google's new neural networks have impressive performance, but are inferior to OpenAI's GPT-4.
News Meta and IBM launch the AI Alliance: pros and cons The new initiative by large companies has several positive and negative aspects for the artificial intelligence industry and startups.
Simple Terms Featured Why does AI draw the wrong number of fingers on a hand? We explain why a task that is simple for humans is very difficult for modern artificial intelligence.
Simple Terms Featured How are smartphone cameras similar to human vision? The cameras of modern smartphones are very similar to our vision and brain. In this article, we will explain this similarity.
