Deep Learning

DeltaCNN: Efficient processing of CNN inference for continuous mobile vision
Fast calculation of convolutional layers in continuous video streams using the fact that the video scene does not change significantly
PDF
DeltaCNN: Efficient processing of CNN inference for continuous mobile vision
Rocket Recycling with Reinforcement Learning
Hovering & landing task with soft actor critic & curiosity-driven experience replay
PDF
Rocket Recycling with Reinforcement Learning
SPADE-based Line Art Colorization
SPADE-based model that colorizes a given line art image using a hint image
PDF Code Slides
SPADE-based Line Art Colorization
Validating Labeling Functions in Domain Shift
Domain shift detection using programmatic weak supervision
PDF Slides
Validating Labeling Functions in Domain Shift
Will Overly Polite Sentences Harm Model Performance? Adversarial Pragmatic Perturbation for NLP
Adversarial attack on LLM model with tone perturbation and its defense
PDF
Will Overly Polite Sentences Harm Model Performance? Adversarial Pragmatic Perturbation for NLP