How Uber Uses Neural Networks

The article discusses the application of relational graph learning, specifically relational graph convolutional networks (RGCN), to detect collusion in fraudulent activities within the Uber platfor...

This article discusses the use of Graph Neural Networks for predicting vehicle trajectories in autonomous driving by modeling pairwise interactions among agents.

The article discusses Uber's participation in the NeurIPS 2019 conference, highlighting their commitment to advancing machine learning through research and practical applications.

The article discusses how Uber Eats utilizes graph learning techniques to enhance its food recommendation system, improving user experience by providing personalized dish and restaurant suggestions.

The article introduces Graph Recurrent Attention Networks (GRANs), a new family of deep generative models designed for efficient graph generation.

The article explores the Lottery Ticket Hypothesis, which posits that within large neural networks, smaller subnetworks can be identified that perform comparably to the full network when trained in...

The article discusses the role of women in data science at Uber, highlighting their contributions to various projects and the importance of diversity in technical fields.

The article discusses an innovative approach to enhance the performance of Convolutional Neural Networks (CNNs) by utilizing JPEG's internal representations.

The article discusses the evolution of Recurrent Neural Networks (RNNs), specifically focusing on the advancements made through evolutionary and reinforcement learning techniques.

The article discusses the limitations of Convolutional Neural Networks (CNNs) in performing coordinate transformations and introduces the CoordConv layer as a solution.

The article discusses Uber's innovative hybrid Exponential Smoothing-Recurrent Neural Network (ES-RNN) model that won the M4 Forecasting Competition.

The article discusses the importance of accurately segmenting building footprints using a novel framework called Deep Structured Active Contours (DSAC).

The article presents a novel approach for semi-automatic annotation of object instances in images, shifting from traditional pixel-labeling to polygon prediction.

The article discusses Graph Partition Neural Networks (GPNN), an advanced extension of Graph Neural Networks (GNNs) designed to efficiently handle large graphs.

This article explores the theoretical properties of deep neural networks, specifically their relationship with Gaussian processes.

The article discusses the use of Graph Neural Networks (GNNs) for inference in probabilistic graphical models, highlighting their ability to outperform traditional message-passing algorithms like b...

The article discusses NerveNet, a novel approach to learning structured policies for continuous control using Graph Neural Networks.

The article discusses SBNet, an open-source algorithm developed by Uber ATG that leverages activation block sparsity to enhance the speed of Convolutional Neural Networks (CNNs).

The article summarizes key highlights from the Uber Engineering Blog in 2017, showcasing advancements in technology that have enhanced user experiences across Uber's services.

The article discusses the Reversible Residual Network (RevNet), a variant of Residual Networks (ResNets) that allows for backpropagation without storing intermediate activations, thereby reducing m...

The article discusses the effectiveness of genetic algorithms (GAs) as a competitive alternative to traditional gradient-based methods for training deep neural networks (DNNs) in reinforcement lear...

The article discusses the challenges of applying neuroevolution to large, deep neural networks and introduces safe mutation operators that allow for effective exploration without disrupting existin...

The article discusses the development of a 3D Graph Neural Network (3DGNN) for RGBD semantic segmentation, emphasizing the integration of 2D appearance and 3D geometric information.

This article discusses the implementation of uncertainty estimation in neural networks for time series prediction at Uber, focusing on the use of Bayesian neural networks (BNNs) and the Monte Carlo...

The article discusses the use of Graph Neural Networks (GNNs) for recognizing situations in images by predicting salient verbs and their associated semantic roles.

The article discusses Uber's approach to extreme event forecasting using Recurrent Neural Networks (RNNs), specifically Long Short Term Memory (LSTM) architecture.