Accurately recognizing a revisited place is crucial for embodied agents to localize and navigate. This requires visual representations to be distinct, despite strong variations in camera viewpoint and scene appearance. Existing visual place recognition pipelines encode the "whole" image and search for matches. This poses a fundamental challenge in matching two images of the same place captured from different camera viewpoints: "the similarity of what overlaps can be dominated by the dissimilarity of what does not overlap". We address this by encoding and searching for "image segments" instead of the whole images. We propose to use open-set image segmentation to decompose an image into `meaningful' entities (i.e., things and stuff). This enables us to create a novel image representation as a collection of multiple overlapping subgraphs connecting a segment with its neighboring segments, dubbed SuperSegment. Furthermore, to efficiently encode these SuperSegments into compact vector representations, we propose a novel factorized representation of feature aggregation. We show that retrieving these partial representations leads to significantly higher recognition recall than the typical whole image based retrieval. Our segments-based approach, dubbed SegVLAD, sets a new state-of-the-art in place recognition on a diverse selection of benchmark datasets, while being applicable to both generic and task-specialized image encoders. Finally, we demonstrate the potential of our method to ``revisit anything'' by evaluating our method on an object instance retrieval task, which bridges the two disparate areas of research: visual place recognition and object-goal navigation, through their common aim of recognizing goal objects specific to a place. Source code: https://github.com/AnyLoc/Revisit-Anything.

Traffic4cast is an annual competition to predict spatio temporal traffic based on real world data. We propose an approach using Graph Neural Networks that directly works on the road graph topology which was extracted from OpenStreetMap data. Our architecture can incorporate a hierarchical graph representation to improve the information flow between key intersections of the graph and the shortest paths connecting them. Furthermore, we investigate how the road graph can be compacted to ease the flow of information and make use of a multi-task approach to predict congestion classes and ETA simultaneously. Our code and models are released on Github.

Short-term traffic forecasting is an extensively studied topic in the field of intelligent transportation system. However, most existing forecasting systems are limited by the requirement of real-time probe vehicle data because of their formulation as a time series forecasting problem. Towards this issue, the NeurIPS 2022 Traffic4cast challenge is dedicated to predicting the citywide traffic states with publicly available sparse loop count data. This technical report introduces our second-place winning solution to the extended challenge of ETA prediction. We present a similarity-based feature extraction method using multiple nearest neighbor (NN) filters. Similarity-based features, static features, node flow features and combined features of segments are extracted for training the gradient boosting decision tree model. Experimental results on three cities (including London, Madrid and Melbourne) demonstrate the strong predictive performance of our approach, which outperforms a number of graph-neural-network-based solutions in the task of travel time estimation. The source code is available at \url{https://github.com/c-lyu/Traffic4Cast2022-TSE}.

Accurate traffic forecasting is of the utmost importance for optimal travel planning and for efficient city mobility. IARAI (The Institute of Advanced Research in Artificial Intelligence) organizes Traffic4cast, a yearly traffic prediction competition based on real-life data [https://www.iarai.ac.at/traffic4cast/], aiming to leverage artificial intelligence advances for producing accurate traffic estimates. We present our solution to the IARAI Traffic4cast 2022 competition, in which the goal is to develop algorithms for predicting road graph edge congestion classes and supersegment-level travel times. In contrast to the previous years, this year's competition focuses on modelling graph edge level behaviour, rather than more coarse aggregated grid-based traffic movies. Due to this, we leverage a method familiar from tabular data modelling -- gradient-boosted decision tree ensembles. We reduce the dimensionality of the input data representing traffic counters with the help of the classic PCA method and feed it as input to a LightGBM model. This simple, fast, and scalable technique allowed us to win second place in the core competition. The source code and references to trained model files and submissions are available at https://github.com/skandium/t4c22 .

Travel-time prediction constitutes a task of high importance in transportation networks, with web mapping services like Google Maps regularly serving vast quantities of travel time queries from users and enterprises alike. Further, such a task requires accounting for complex spatiotemporal interactions (modelling both the topological properties of the road network and anticipating events -- such as rush hours -- that may occur in the future). Hence, it is an ideal target for graph representation learning at scale. Here we present a graph neural network estimator for estimated time of arrival (ETA) which we have deployed in production at Google Maps. While our main architecture consists of standard GNN building blocks, we further detail the usage of training schedule methods such as MetaGradients in order to make our model robust and production-ready. We also provide prescriptive studies: ablating on various architectural decisions and training regimes, and qualitative analyses on real-world situations where our model provides a competitive edge. Our GNN proved powerful when deployed, significantly reducing negative ETA outcomes in several regions compared to the previous production baseline (40+% in cities like Sydney).

Google users contribute more than 20 million pieces of information on Maps every day – that's more than 200 contributions every second. The uncertainty of traffic can crash the algorithms predicting the best ETA. There is also a chance of new roads and buildings being built all the time. Though Google Maps gets its ETA right most of the time, there is still room for improvement. Researchers at Alphabet-owned DeepMind have partnered with the Google Maps team to improve the accuracy of the real-time ETAs by up to 50% in places like Berlin, Jakarta, São Paulo, Sydney, Tokyo, and Washington D.C.

Google Maps is one of the company's most widely-used products, and its ability to predict upcoming traffic jams makes it indispensable for many drivers. Each day, says Google, more than 1 billion kilometers of road are driven with the app's help. But, as the search giant explains in a blog post today, its features have got more accurate thanks to machine learning tools from DeepMind, the London-based AI lab owned by Google's parent company Alphabet. In the blog post, Google and DeepMind researchers explain how they take data from various sources and feed it into machine learning models to predict traffic flows. This data includes live traffic information collected anonymously from Android devices, historical traffic data, information like speed limits and construction sites from local governments, and also factors like the quality, size, and direction of any given road.

Labor Day weekend typically spurs highway and neighborhood traffic. The AAA didn't offer travel estimates this year based on the COVID-19 pandemic, but last year 43 million Americans traveled for Memorial Day Weekend, the second-highest travel volume on record since the company began tracking holiday travel volumes in 2000. Google estimates that people use Google Maps to drive more than 700 billion miles daily in more than 220 countries and territories worldwide. The app shows the driver which the direction to travel, whether traffic along route is heavy or light, an estimated travel time, and an estimated time of arrival. It may appear simple, but the technology behind the app is quite complex because conditions such as an accident or rockslide in a canyon can change the directions in a matter of seconds.