Search papers, labs, and topics across Lattice.
Universit茅 Paris-Saclay
2
0
5
Discrete proxy-tokens in Composer enhance visual grounding accuracy by 9.0 points, bridging the semantic-spatial gap that plagues traditional methods.
Transformer-based object detectors are surprisingly resistant to adversarial attacks crafted for CNNs, but can be hardened further by training against a diverse "cocktail" of strong attacks.
