arXiv subjects

Stephen Law

Publications and source records attributed to Stephen Law.

23 records · Page 2Linked to original sources

arXiv2021

Explaining Classifiers using Adversarial Perturbations on the Perceptual Ball

We present a simple regularization of adversarial perturbations based upon the perceptual loss. While the resulting perturbations remain imperceptible to the human eye, they differ from existing adversarial perturbations in that they are semi-sparse alterations that highlight objects and regions of interest while leaving the background unaltered. As a semantically meaningful adverse perturbations, it forms a bridge between counterfactual explanations and adversarial perturbations in the space of images. We evaluate our approach on several standard explainability benchmarks, namely, weak localization, insertion deletion, and the pointing game demonstrating that perceptually regularized counterfactuals are an effective explanation for image-based classifiers.

cs.CV↗

arXiv2021

Jane Jacobs in the Sky: Predicting Urban Vitality with Open Satellite Data

The presence of people in an urban area throughout the day -- often called 'urban vitality' -- is one of the qualities world-class cities aspire to the most, yet it is one of the hardest to achieve. Back in the 1970s, Jane Jacobs theorized urban vitality and found that there are four conditions required for the promotion of life in cities: diversity of land use, small block sizes, the mix of economic activities, and concentration of people. To build proxies for those four conditions and ultimately test Jane Jacobs's theory at scale, researchers have had to collect both private and public data from a variety of sources, and that took decades. Here we propose the use of one single source of data, which happens to be publicly available: Sentinel-2 satellite imagery. In particular, since the first two conditions (diversity of land use and small block sizes) are visible to the naked eye from satellite imagery, we tested whether we could automatically extract them with a state-of-the-art deep-learning framework and whether, in the end, the extracted features could predict vitality. In six Italian cities for which we had call data records, we found that our framework is able to explain on average 55% of the variance in urban vitality extracted from those records.

cs.CV↗

arXiv2019

Take a Look Around: Using Street View and Satellite Images to Estimate House Prices

When an individual purchases a home, they simultaneously purchase its structural features, its accessibility to work, and the neighborhood amenities. Some amenities, such as air quality, are measurable while others, such as the prestige or the visual impression of a neighborhood, are difficult to quantify. Despite the well-known impacts intangible housing features have on house prices, limited attention has been given to systematically quantifying these difficult to measure amenities. Two issues have led to this neglect. Not only do few quantitative methods exist that can measure the urban environment, but that the collection of such data is both costly and subjective. We show that street image and satellite image data can capture these urban qualities and improve the estimation of house prices. We propose a pipeline that uses a deep neural network model to automatically extract visual features from images to estimate house prices in London, UK. We make use of traditional housing features such as age, size, and accessibility as well as visual features from Google Street View images and Bing aerial images in estimating the house price model. We find encouraging results where learning to characterize the urban quality of a neighborhood improves house price prediction, even when generalizing to previously unseen London boroughs. We explore the use of non-linear vs. linear methods to fuse these cues with conventional models of house pricing, and show how the interpretability of linear models allows us to directly extract proxy variables for visual desirability of neighborhoods that are both of interest in their own right, and could be used as inputs to other econometric methods. This is particularly valuable as once the network has been trained with the training data, it can be applied elsewhere, allowing us to generate vivid dense maps of the visual appeal of London streets.

econ.EM↗

arXiv2019

An unsupervised approach to Geographical Knowledge Discovery using street level and street network images

Recent researches have shown the increasing use of machine learn-ing methods in geography and urban analytics, primarily to extract features and patterns from spatial and temporal data using a supervised approach. Researches integrating geographical processes in machine learning models and the use of unsupervised approacheson geographical data for knowledge discovery had been sparse. This research contributes to the ladder, where we show how latent variables learned from unsupervised learning methods on urbanimages can be used for geographic knowledge discovery. In particular, we propose a simple approach called Convolutional-PCA(ConvPCA) which are applied on both street level and street network images to find a set of uncorrelated and ordered visual latentcomponents. The approach allows for meaningful explanations using a combination of geographical and generative visualisations to explore the latent space, and to show how the learned representation can be used to predict urban characteristics such as streetquality and street network attributes. The research also finds that the visual components from the ConvPCA model achieves similaraccuracy when compared to less interpretable dimension reduction techniques.

cs.CV↗

arXiv2018

Urban Analytics: Multiplexed and Dynamic Community Networks

In the past decade, cities have experienced rapid growth, expansion, and changes in their community structure. Many aspects of critical urban infrastructure are closely coupled with the human communities that they serve. Urban communities are composed of a multiplex of overlapping factors which can be distinguished into cultural, religious, social-economic, political, and geographical layers. In this paper, we review how increasingly available heterogeneous mobile big data sets can be leveraged to detect the community interaction structure using natural language processing and machine learning techniques. A number of community layer and interaction detection algorithms are then reviewed, with a particular focus on robustness, stability, and causality of evolving communities. The better understanding of the structural dynamics and multiplexed relationships can provide useful information to inform both urban planning policies and shape the design of socially coupled urban infrastructure systems.

cs.SI↗