arXiv subjects

Pål Sundsøy

Publications and source records attributed to Pål Sundsøy.

5 recordsLinked to original sources

arXiv2017

Modeling the Temporal Nature of Human Behavior for Demographics Prediction

Mobile phone metadata is increasingly used for humanitarian purposes in developing countries as traditional data is scarce. Basic demographic information is however often absent from mobile phone datasets, limiting the operational impact of the datasets. For these reasons, there has been a growing interest in predicting demographic information from mobile phone metadata. Previous work focused on creating increasingly advanced features to be modeled with standard machine learning algorithms. We here instead model the raw mobile phone metadata directly using deep learning, exploiting the temporal nature of the patterns in the data. From high-level assumptions we design a data representation and convolutional network architecture for modeling patterns within a week. We then examine three strategies for aggregating patterns across weeks and show that our method reaches state-of-the-art accuracy on both age and gender prediction using only the temporal modality in mobile metadata. We finally validate our method on low activity users and evaluate the modeling assumptions.

cs.LG↗

arXiv2017

Big Data for Social Sciences: Measuring patterns of human behavior through large-scale mobile phone data

Through seven publications this dissertation shows how anonymized mobile phone data can contribute to the social good and provide insights into human behaviour on a large scale. The size of the datasets analysed ranges from 500 million to 300 billion phone records, covering millions of people. The key contributions are two-fold: 1. Big Data for Social Good: Through prediction algorithms the results show how mobile phone data can be useful to predict important socio-economic indicators, such as income, illiteracy and poverty in developing countries. Such knowledge can be used to identify where vulnerable groups in society are, reduce economic shocks and is a critical component for monitoring poverty rates over time. Further, the dissertation demonstrates how mobile phone data can be used to better understand human behaviour during large shocks in society, exemplified by an analysis of data from the terror attack in Norway and a natural disaster on the south-coast in Bangladesh. This work leads to an increased understanding of how information spreads, and how millions of people move around. The intention is to identify displaced people faster, cheaper and more accurately than existing survey-based methods. 2. Big Data for efficient marketing: Finally, the dissertation offers an insight into how anonymised mobile phone data can be used to map out large social networks, covering millions of people, to understand how products spread inside these networks. Results show that by including social patterns and machine learning techniques in a large-scale marketing experiment in Asia, the adoption rate is increased by 13 times compared to the approach used by experienced marketers. A data-driven and scientific approach to marketing, through more tailored campaigns, contributes to less irrelevant offers for the customers, and better cost efficiency for the companies.

cs.CY↗

arXiv2016

Estimating individual employment status using mobile phone network data

This study provides the first confirmation that individual employment status can be predicted from standard mobile phone network logs externally validated with household survey data. Individual welfare and households vulnerability to shocks are intimately connected to employment status and professions of household breadwinners. At a societal level unemployment is an important indicator of the performance of an economy. By deriving a broad set of novel mobile phone network indicators reflecting users financial, social and mobility patterns we show how machine learning models can be used to predict 18 categories of profession in a South-Asian developing country. The model predicts individual unemployment status with 70.4 percent accuracy. We further show how unemployment can be aggregated from individual level and mapped geographically at cell tower resolution, providing a promising approach to map labor market economic indicators, and the distribution of economic productivity and vulnerability between censuses, especially in heterogeneous urban areas. The method also provides a promising approach to support data collection on vulnerable populations, which are frequently under-represented in official surveys.

cs.SI↗

arXiv2016

Quantum Computing: Linear Optics Implementations

One of the main problems that optical quantum computing has to overcome is the efficient construction of two-photon gates. Theoretically these gates can be realized using Kerr-nonlinearities, but the techniques involved are experimentally very difficult. We therefore employ linear optics with projective measurements to generate these non-linearities. The downside is that the measurement-induced nonlinearities achieved with linear optics are less versatile and the success rate can be quite low. This project is mainly the result of a literature study but also a theoretical work on the physics behind quantum optical multiports which is essential for realizing two-photon gates. By applying different postcorrection techniques we increase the probability of success in a modifed non-linear sign shift gate which is foundational for the two photon controlled-NOT gate. We prove that it's not possible to correct the states by only using a single beam splitter. We show that it might be possible to increase the probability of success using a more complex setup with at least two error-correcting beam splitters.

quant-ph↗

arXiv2016

Can mobile usage predict illiteracy in a developing country?

The present study provides the first evidence that illiteracy can be reliably predicted from standard mobile phone logs. By deriving a broad set of mobile phone indicators reflecting users financial, social and mobility patterns we show how supervised machine learning can be used to predict individual illiteracy in an Asian developing country, externally validated against a large-scale survey. On average the model performs 10 times better than random guessing with a 70% accuracy. Further we show how individual illiteracy can be aggregated and mapped geographically at cell tower resolution. Geographical mapping of illiteracy is crucial to know where the illiterate people are, and where to put in resources. In underdeveloped countries such mappings are often based on out-dated household surveys with low spatial and temporal resolution. One in five people worldwide struggle with illiteracy, and it is estimated that illiteracy costs the global economy more than 1 trillion dollars each year. These results potentially enable costeffective, questionnaire-free investigation of illiteracy-related questions on an unprecedented scale

cs.AI↗