Xuyang Chen

Wi-Fi sensing accuracy jumps to 92.83% by decomposing CSI into trajectory-guided tokens that capture complex temporal dynamics.

ReLViC achieves robust video coding by dispersing latent tokens, outperforming established codecs even in the face of severe packet loss.

By framing camera trajectory generation as a language-grounded spatial reasoning problem, CinemaTraj achieves high-quality, collision-free cinematographic outputs that align closely with user prompts.

Selective disclosure in watermarking could revolutionize how we manage privacy and access control in AI-generated text.

Chart design can mislead vision models into learning visual cues instead of actual temporal patterns, fundamentally altering their performance.