Benlei Cui

Evolving video agents can boost long-form video understanding accuracy by over 13 percentage points while dramatically reducing processing costs.

Backdoor attacks on self-supervised models can be effectively countered without any reliance on labels or training data, achieving substantial performance gains across multiple attack types.

Stop guessing why your GUI agent failed: GUIDE breaks down complex trajectories into interpretable subtasks, pinpointing errors and suggesting fixes with unprecedented accuracy.

Diffusion models get a 2.88x speed boost in low-step regimes thanks to a new Padé approximation technique that predicts feature evolution more accurately than Taylor-based methods.