A boundary-forcing masked modeling paradigm for self-supervised vision pretraining yields a 1B model rivaling 7B models on dense spatial perception tasks.
Emerging properties in self-supervised vision transformers
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
years
2026 4representative citing papers
HGPM learns compositional patterns in hypergraphs by subset tokenization and inclusion-aware masked Transformer reconstruction, matching or exceeding SOTA on ten benchmarks and correctly identifying inhibitory drug additions in adverse-event prediction where prior methods fail.
Formalizes Fashion Detail Generation task, releases FDBench benchmark with 40K+ pairs, and proposes CFAD distillation method plus RL consistency reward that outperforms open-source baselines.
Decouples semantic and spatial tokens in NVS transformers to resolve representation ambiguity, yielding consistent gains with near-zero added latency.
citing papers explorer
-
Vision Pretraining for Dense Spatial Perception
A boundary-forcing masked modeling paradigm for self-supervised vision pretraining yields a 1B model rivaling 7B models on dense spatial perception tasks.
-
Hypergraph Pattern Machine: Compositional Tokenization for Higher-Order Interactions
HGPM learns compositional patterns in hypergraphs by subset tokenization and inclusion-aware masked Transformer reconstruction, matching or exceeding SOTA on ten benchmarks and correctly identifying inhibitory drug additions in adverse-event prediction where prior methods fail.
-
DetailAnywhere: Fashion Detail Generation via Cross-Modal Feature Alignment Distillation
Formalizes Fashion Detail Generation task, releases FDBench benchmark with 40K+ pairs, and proposes CFAD distillation method plus RL consistency reward that outperforms open-source baselines.
-
Resolving Representation Ambiguity in Feedforward Novel View Synthesis Transformer via Semantic-Spatial Decoupling
Decouples semantic and spatial tokens in NVS transformers to resolve representation ambiguity, yielding consistent gains with near-zero added latency.