On the Temporality for Sketch Representation Learning
This work addresses a gap in sketch representation learning for researchers, but it is incremental as it focuses on optimizing existing sequence-based methods.
The paper tackled the problem of understanding the relevance of temporal aspects in sketch representation learning, finding that absolute coordinates outperform relative ones and non-autoregressive decoders are better than autoregressive ones, with temporality's importance varying by order and task.
Sketches are simple human hand-drawn abstractions of complex scenes and real-world objects. Although the field of sketch representation learning has advanced significantly, there is still a gap in understanding the true relevance of the temporal aspect to the quality of these representations. This work investigates whether it is indeed justifiable to treat sketches as sequences, as well as which internal orders play a more relevant role. The results indicate that, although the use of traditional positional encodings is valid for modeling sketches as sequences, absolute coordinates consistently outperform relative ones. Furthermore, non-autoregressive decoders outperform their autoregressive counterparts. Finally, the importance of temporality was shown to depend on both the order considered and the task evaluated.