Papers
14
Total Citations
675
H-Index
11
About
Kate Saenko is a prominent computer vision and robotics researcher whose work sits at the intersection of domain adaptation, robotic perception, and deep learning. She is perhaps best known for her contributions to bridging the simulation-to-reality gap — a fundamental challenge in both computer vision and robotics where models trained on synthetic or controlled data must generalize to messy real-world environments. Her benchmark dataset paper, "VisDA: A Synthetic-to-Real Benchmark for Visual Domain Adaptation" (181 citations), has become a foundational resource for researchers tackling this problem, establishing rigorous standards for evaluating domain adaptation methods. Saenko has also made significant strides in robotic grasping, developing visuomotor controllers that leverage simulated depth images to enable reliable real-world manipulation (100 citations), and advancing grasp pose detection in dense, cluttered point clouds (57 citations). Her work on hierarchical reinforcement learning and adapting deep visuomotor representations further demonstrates her broad impact across autonomous systems. With contributions spanning object detection, indoor localization, and interactive learning, Saenko's research has collectively shaped how modern AI systems transfer knowledge across domains — a challenge central to deploying robust, real-world intelligent machines.
Research Focus
Key Achievements
Top Papers
- 1VisDA: A Synthetic-to-Real Benchmark for Visual Domain Adaptation181 citations · 2018
- 2
- 3Adapting Deep Visuomotor Representations with Weak Pairwise Constraints81 citations · 2020
- 4Learning Multi-Level Hierarchies with Hindsight76 citations · 2017
- 5
- 6High precision grasp pose detection in dense clutter57 citations · 2016
- 7Grasp Pose Detection in Point Clouds40 citations · 2017
- 8
- 9Interactive adaptation of real-time object detectors15 citations · 2014
- 10