Deep learning

Learning quantile function without quantile crossing for distribution-free time series forecasting

Youngsuk Park, Danielle Maddix Robinson, Yuyang (Bernie) Wang, Jan Gasthaus

ICML 2021 Workshop on Distribution-Free Uncertainty Quantification

2021

Quantile regression is an effective technique to quantify uncertainty, and fit challenging underlying distributions. Generating full probabilistic predictions requires multiple quantile regressions over multiple quantile levels. As a result, quantile crossing is a common drawback to these approaches since it violates the desirable monotone property of the conditional quantile function. In this work, we

Machine learning

Deep AutoAugment

Yu Zheng, Zhi Zhang, Shen Yan, Mi Zhang

ICML 2021 Workshop on Machine Learning for Data: Automated Creation, Privacy, Bias, ICLR 2022

2021

While recent automatic data augmentation works lead to state-of-the-art results, their design spaces and the derived data augmentation strategies still incorporate strong human priors. In this work, instead of selecting a set of hand-picked default augmentations alongside the searched data augmentations, we propose a fully automated approach for data augmentation search called Deep AutoAugment (DeepAA).

Computer vision

Deep hierarchical product classification based on pre-trained multilingual knowledge

Wen Zhang, Yanbin Lu, Bella Dubrov, Zhi Xu, Shang Shang, Emilio Maldonado

IEEE - The Bulletin of the Technical Committee on Data Engineering

2021

The customer experience of online shopping is largely contingent on the accuracy of product classification. Considering the amount of products and all the possible categories, it is desirable to construct a framework to auto-assign products into correct categories at scale. Machine learning based systems often suffer from poor data quality, such as incomplete item descriptions, adversarial noise in the

Conversational AI

LoWino: Towards efficient low-precision Winograd convolutions on modern CPUs

Guangli Li, Zhen Jia, Xiaobing Feng, Yida Wang

ICPP 2021

2021

Low-precision computation, which has been widely supported in contemporary hardware, is considered as one of the most effective methods to accelerate convolutional neural networks. However, low-precision computation is not widely used to speed up Winograd, an algorithm for fast convolution computation, due to the numerical error introduced by combining Winograd transformation and quantization. In this paper

Cloud and systems

Scene generation from backgrounds to objects and anything in between: A deep learning robotics survey

Bouchra Bouqata, Krishna Aswani, David Bailey

ArabWIC 2021

2021

The recent rapid progress of deep learning algorithms in generating realistic images, especially in Generative Adversarial Networks (GAN) and Variational Auto-Encoders (VAE), has helped advance new applications. Examples of such applications range from generating and manipulating new synthetic data for self-driving cars, to building/urban architectures, to interior design, and gaming. Furthermore, several

Robotics

Learning self-consistency for deepfake detection

Eric Zhao, Xiang Xu, Mingze Xu, Hui Ding, Yuanjun Xiong, Wei Xia

ICCV 2021

2021

We propose a new method to detect deepfake images using the cue of the source feature inconsistency within the forged images. It is based on the hypothesis that images’ distinct source features can be preserved and extracted after going through state-of-the-art deepfake generation processes. We introduce a novel representation learning approach, called pair-wise self-consistency learning (PCL),for training

Computer vision

GAN-control: Explicitly controllable GANs

Alon Shoshan, Nadav Bhonker, Igor Kviatkovsky, Gérard Medioni

ICCV 2021

2021

We present a framework for training GANs with explicit control over generated facial images. We are able to control the generated image by settings exact attributes such as age, pose, expression, etc. Most approaches for manipulating GAN-generated images achieve partial control by leveraging the latent space disentanglement properties, obtained implicitly after standard GAN training. Such methods are able

Computer vision

Learning hierarchical graph neural networks for image clustering

Yifan Xing, Tong He, Tianjun Xiao, Richard Wang, Yuanjun Xiong, Wei Xia, David Wipf, Zheng Zhang, Stefano Soatto

ICCV 2021

2021

We propose a hierarchical graph neural network (GNN) model that learns how to cluster a set of images into an unknown number of identities using a training set of images annotated with labels belonging to a disjoint set of identities. Our hierarchical GNN uses a novel approach to merge connected components predicted at each level of the hierarchy to form a new graph at the next level. Unlike fully unsupervised

Computer vision

Video contrastive learning with global context

Haofei Kuang, Yi Zhu, Zhi Zhang, Xinyu Li, Joe Tighe, Sören Schwertfeger, Cyrill Stachniss, Mu Li

ICCV 2021 Workshop on AI for Creative Video Editing and Understanding

2021

Contrastive learning has revolutionized the self-supervised image representation learning field and recently been adapted to the video domain. One of the greatest advantages of contrastive learning is that it allows us to flexibly define powerful loss objectives as long as we can find a reasonable way to formulate positive and negative samples to contrast. However, existing approaches rely heavily on the

Computer vision

Tabular data concept type detection using star-transformers

Yiwei Zhou, Siffi Singh, Christos Christodoulopoulos

CIKM 2021

2021

Tabular data is an invaluable information resource for search, information extraction and question answering about the world. It is critical to understand the semantic concept types for table columns in order to fully exploit the information in tabular data. In this paper, we focus on learning-based approaches for column concept type detection without relying on any metadata or queries to existing knowledge

Conversational AI

Deep learning

Work with us