Researchers have developed the Sparse Transformer, a deep neural network that can predict what comes next in a sequence of text, images, or sound. This model uses an improved attention mechanism to analyze long sequences, 30 times longer than previously possible. The development is based on work by OpenAI. --- Why it matters: This breakthrough matters for researchers and engineers working with generative models because it enables the analysis of much longer sequences, which can be essential for tasks like natural language processing, image recognition, or music generation. Source: https://openai.com/index/sparse-transformer