New approaches foster hope that computers can comprehend paragraphs, classify email as spam, or generate a satisfying end to a short story. [Read More]
Summary… * Machine-Made
In 2012, artificial intelligence researchers revealed a big improvement in computers’ ability to recognize images by feeding a neural network millions of labeled images from a database called ImageNet.
But recent research from fast.ai, OpenAI, and the Allen Institute for #AI suggests a potential breakthrough, with more robust language models that can help researchers tackle a range of unsolved problems.
The most widely tested model, so far, is called Embeddings from Language Models, or ELMo.
For languages other than English, researchers often don’t have enough labeled data to accomplish even basic tasks.
Feed a neural network a billion words, as Peters’ team did, and this approach turns out to be quite effective.
Opinion… * Man-Made