Deep Learning
On sensitivity of meta-learning to support data
Meta-learning algorithms are widely used for few-shot learning.
On sensitivity of meta-learning to support data
Meta-learning algorithms are widely used for few-shot learning.
Excess Capacity and Backdoor Poisoning
From a computational standpoint, we show that under certain assumptions, adversarial training can detect the presence of backdoors in a training set.
HA WQ-V2: Hessian Aware trace-Weighted Quantization of Neural Networks
Furthermore, we present results for object detection on Microsoft COCO, where we achieve 2.6 higher mAP than direct uniform quantization and 1.6 higher mAP than the recently proposed method of
HA WQ-V2: Hessian Aware trace-Weighted Quantization of Neural Networks
Furthermore, we present results for object detection on Microsoft COCO, where we achieve 2.6 higher mAP than direct uniform quantization and 1.6 higher mAP than the recently proposed method of
Human-Adversarial Visual Question Answering Sasha Sheng Amanpreet Singh V edanuj Goswami Jose Alberto Lopez Magana Tristan Thrush Wojciech Galuba Devi Parikh Douwe Kiela
Performance on the most commonly used Visual Question Answering dataset (VQA v2) is starting to approach human accuracy.