It’s commonly assumed that models pretrained to achieve high performance on ImageNet will perform better on other visual tasks after fine-tuning. But is it always true? A new study reached surprising conclusions.
It’s commonly assumed that models pretrained to achieve high performance on ImageNet will perform better on other visual tasks after fine-tuning. But is it always true? A new study reached surprising conclusions.
The pandemic has pushed hospitals to their limits. A new machine learning system could help doctors make sure the most severe cases get timely, appropriate care.
The transformer architecture has inspired a plethora of variations. Yet researchers have used a patchwork of metrics to evaluate their performance, making them hard to compare. New work aims to level the playing field.
The transformer architecture has inspired a plethora of variations. Yet researchers have used a patchwork of metrics to evaluate their performance, making them hard to compare. New work aims to level the playing field.
The pandemic has pushed hospitals to their limits. A new machine learning system could help doctors make sure the most severe cases get timely, appropriate care.
Models that interpret the interplay of words and images tend to be trained on richer bodies of text than images. Recent research worked toward giving such models a more balanced knowledge of the two domains.
Models that interpret the interplay of words and images tend to be trained on richer bodies of text than images. Recent research worked toward giving such models a more balanced knowledge of the two domains.
Performance in language tasks rises with the size of the model — yet, as a model’s parameter count rises, so does the time it takes to render output. New work pumps up the number of parameters without slowing down the network.
Performance in language tasks rises with the size of the model — yet, as a model’s parameter count rises, so does the time it takes to render output. New work pumps up the number of parameters without slowing down the network.
Train a multilingual language translator to translate between Spanish and English and between English and German, and it may be able to translate directly between Spanish and German as well. New work proposes a simple path to better machine translation between languages.
Train a multilingual language translator to translate between Spanish and English and between English and German, and it may be able to translate directly between Spanish and German as well. New work proposes a simple path to better machine translation between languages.
The same models trained on the same data may show the same performance in the lab, and yet respond very differently to data they haven’t seen before. New work finds this inconsistency to be pervasive.
The same models trained on the same data may show the same performance in the lab, and yet respond very differently to data they haven’t seen before. New work finds this inconsistency to be pervasive.
Doctors who treat breast cancer typically use a quick, inexpensive tumor-tissue stain test to diagnose the illness and a slower, more costly one to determine treatment. A new neural network could help doctors to go straight from diagnosis to treatment.
Stay updated with weekly AI News and Insights delivered to your inbox