Jan Hula

TL;DR: A new version of Y OLO with better performance and extended with instance segmentation called Poly-YOLO, which has the same precision as YOLOv3, but it is three times smaller and twice as fast, thus suitable for embedded devices.

TL;DR: The authors conducted a large-scale systematic study of candidate pretraining tasks, comparing 19 different tasks both as alternatives and complements to language modeling and found that language modeling can support the use of language modeling, especially when combined with pretraining on additional labeled-data tasks.

TL;DR: This paper conducted a large-scale systematic study of candidate pretraining tasks, comparing 19 different tasks both as alternatives and complements to language modeling and found that language modeling can support the use of language modeling, especially when combined with pretraining on additional labeled-data tasks.

TL;DR: The primary results support the use of language modeling as a pretraining task and set a new state of the art among comparable models using multitask learning with language models and suggest that the widely-used paradigm of pretraining and freezing sentence encoders may not be an ideal platform for further work.

TL;DR: This paper proposes a preprocessing technique that enriches the original image data using local intensity information and introduces a new image structure named image represented by a fuzzy function to encode this information into an image.