Hi, I'm Vincent!

I'm a senior Software Engineer at Google, where I use Machine Learning to help hundreds of millions of users in their daily lives. Before joining Google, I was a Machine Learning Scientist at Microsoft where I leveraged user interactions with Office products to build intelligent features. I successfully designed and shipped models which helped hundreds of millions of users to achieve more, using Reinforcement Learning to improve user experiences quickly and efficiently.

Before that, I was an applied researcher at Microsoft Research, working on Word Flow for iOS. In 2015, I earned my PhD from EPFL in Lausanne, Switzerland, under the supervision of Prof. Matthias Grossglauser and Prof. Patrick Thiran.

I am interested in Machine Learning and Data Mining, more specifically in combining models and datasets to solve real-world problems. I built Sidekick and Predikon to make some of my results available to the public.

During my PhD, I was an intern at the Machine Learning Group of Microsoft Research in Redmond, WA. I also did two internships at the Machine Learning Department of NEC Labs in Princeton, NJ, working on NLP.

Publications

• Vincent Etter, Mohammad Emtiyaz Khan, Matthias Grossglauser, Patrick Thiran. Online Collaborative Prediction of Regional Vote Results. Proceedings of the 3rd IEEE International Conference on Data Science and Advanced Analytics (DSAA'16), 2016.
  PDF BibTeX Abstract Slides Code & data

  @inproceedings{etter2016dsaa,
  	author = "Etter, Vincent and Khan, Mohammad Emtiyaz and Grossglauser, Matthias and Thiran, Patrick",
  	title = "Online Collaborative Prediction of Regional Vote Results",
  	booktitle = "Proceedings of the 3rd IEEE International Conference on Data Science and Advanced Analytics",
  	series = "DSAA '16",
  	year = "2016",
  }

  We consider online predictions of vote results, where regions across a country vote on an issue under discussion. Such online predictions before and during the day of the vote are useful to media agencies, polling institutes, and political parties, e.g., to identify regions that are crucial in determining the national outcome of a vote. We analyze a unique dataset from Switzerland. The dataset contains 281 votes from 2352 regions over a period of 34 years. We make several contributions towards improving online predictions. First, we show that these votes exhibit a bi-clustering of the vote results, i.e., regions that are spatially close tend to vote similarly, and issues that discuss similar topics show similar global voting patterns. Second, we develop models that can exploit this bi-clustering, as well as the features associated with the votes and regions. Third, we show that, when combining vote results and features together, Bayesian methods are essential to obtaining good performance. Our results show that Bayesian methods give better estimates of the hyperparameters than non-Bayesian methods such as cross-validation. The resulting models generalize well to many different tasks, produce robust predictions, and are easily interpretable.

• Vincent Etter. Combine and Conquer: Mining Social Systems for Prediction. EPFL, 2015.
  PDF BibTeX Abstract Publisher's website

  @phdthesis{etter2015phd,
  	author = "Etter, Vincent",
  	title = "Combine and Conquer: Mining Social Systems for Prediction",
  	school = "EPFL",
  	year = "2015",
  }

  In this thesis, we explore the application of data mining and machine learning techniques to several practical problems. These problems have roots in various fields such as social science, economics, and political science. We show that computer science techniques enable us to bring significant contributions to solving them. Moreover, we show that combining several models or datasets related to the problem we are trying to solve is key to the quality of the solution we find.

  The first application we consider is human mobility prediction. We describe our winning contribution to the Nokia Mobile Data Challenge, in which we predict the next location a user will visit based on his history and the current context. We first highlight some data characteristics that contribute to the difficulty of the task, such as sparsity and non-stationarity. Then, we present three families of models and observe that, even though their average accuracies are similar, their performances vary significantly across users. To take advantage of this diversity, we introduce several strategies to combine models, and show that the combinations outperform any individual predictor.

  The second application we examine is predicting the success of crowdfunding campaigns. We collected data on Kickstarter (one of the most popular crowdfunding platforms) in order to predict whether a campaign will reach its funding goal or not. We show that we obtain good performances by simply using information about money, but that combining this information with social features extracted from Kickstarter's social graph and Twitter improves early predictions. In particular, predictions made a few hours after the beginning of a campaign are improved by 4%, to reach an accuracy of 76%.

  Then, we move to the realms of politics, and first investigate the ideologies of politicians. Using their opinion on several aspects of politics, gathered on a voting advice application (VAA), we show that the themes that divide politicians the most are the ones that we usually associate with left-wing/right-wing and liberal/conservative, thus validating the simplified two-dimensional view of the political system that many people use. We bring attention to the potentially malicious uses of VAAs by creating a fake candidate profile that is able to gather twice as many voting recommendations as any other. To counter this, we demonstrate that we are able to monitor politicians after they were elected, and potentially detect changes of opinion, by combining the data extracted from the VAA with the votes that they cast at the Parliament.

  Finally, we study the outcome of issue votes. We first show that simply considering vote results at a fine geographical level is sufficient to highlight characteristic geographical voting patterns across a country, and their evolution over time. It also enables us to find representative regions that are crucial in determining the national outcome of a vote. We then demonstrate that predicting the actual result of a vote in all regions (in opposition to the binary national outcome) is a much harder task that requires combining data about regions and votes themselves to obtain good performances. We compare the use of Bayesian and non-Bayesian models that combine matrix-factorization and regression. We show that, here too, combining appropriate models and datasets improves the quality of the predictions, and that Bayesian methods give better estimates of the model's hyperparameters.

• Vincent Etter, Julien Herzen, Matthias Grossglauser, Patrick Thiran. Mining Democracy. Proceedings of the second ACM Conference on Online Social Networks (COSN'14), 2014. Best paper award
  PDF BibTeX Abstract Publisher's website Poster Slides Predikon

  @inproceedings{etter2014cosn,
  	author = "Etter, Vincent and Herzen, Julien and Grossglauser, Matthias and Thiran, Patrick",
  	title = "Mining Democracy",
  	booktitle = "Proceedings of the second ACM Conference on Online Social Networks",
  	series = "COSN '14",
  	year = "2014",
  }

  Switzerland has a long tradition of direct democracy, which makes it an ideal laboratory for research on real-world politics. Similar to recent open government initiatives launched worldwide, the Swiss government regularly releases datasets related to state affairs and politics. In this paper, we propose an exploratory, data-driven study of the political landscape of Switzerland, in which we use opinions expressed by candidates and citizens on a web platform during the recent Swiss parliamentary elections, together with fine-grained vote results and parliament votes.

  Following this purely data-driven approach, we show that it is possible to uncover interesting patterns that would otherwise require both tedious manual analysis and domain knowledge. In particular, we show that traditional cultural and/or ideological idiosyncrasies can be highlighted and quantified by looking at vote results and pre-election opinions. We propose a technique for comparing the candidates' opinions expressed before the elections with their actual votes cast in the parliament after the elections. This technique spots politicians that do not vote consistently with the opinions that they expressed during the campaign. We also observe that it is possible to predict surprisingly precisely the outcome of nationwide votes, by looking at the outcome in a single, carefully selected municipality. Our work applies to any country where similar data is available; it points to some of the avenues created by user-generated data emerging from open government initiatives, which enable new data-mining approaches to political and social sciences.

• Vincent Etter, Matthias Grossglauser, Patrick Thiran. Launch Hard or Go Home! Predicting the Success of Kickstarter Campaigns. Proceedings of the first ACM Conference on Online Social Networks (COSN'13), 2013.
  PDF BibTeX Abstract Publisher's website Poster Slides Sidekick Data

  @inproceedings{etter2013cosn,
  	author = "Etter, Vincent and Grossglauser, Matthias and Thiran, Patrick",
  	title = "Launch Hard or Go Home! Predicting the Success of Kickstarter Campaigns",
  	booktitle = "Proceedings of the first ACM Conference on Online Social Networks",
  	series = "COSN '13",
  	year = "2013",
  }

  Crowdfunding websites such as Kickstarter are becoming increasingly popular, allowing project creators to raise hundreds of millions of dollars every year. However, only one out of two Kickstarter campaigns reaches its funding goal and is successful. It is therefore of prime importance, both for project creators and backers, to be able to know which campaigns are likely to succeed.

  We propose a method for predicting the success of Kickstarter campaigns by using both direct information and social features. We introduce a first set of predictors that uses the time series of money pledges to classify campaigns as probable success or failure and a second set that uses information gathered from tweets and Kickstarter's projects/backers graph.

  We show that even though the predictors that are based solely on the amount of money pledged reach a high accuracy, combining them with predictors using social features enables us to improve the performance significantly. In particular, only 4 hours after the launch of a campaign, the combined predictor reaches an accuracy of more than 76% (a relative improvement of 4%).

• Vincent Etter, Mohamed Kafsi, Ehsan Kazemi, Matthias Grossglauser, Patrick Thiran. Where to Go from Here? Mobility Prediction from Instantaneous Information. Pervasive and Mobile Computing, Volume 9, Issue 9, Pages 784-797, 2013.
  PDF BibTeX Abstract Publisher's website

  @article{etter2013pervasive,
  	author = "Etter, Vincent and Kafsi, Mohamed and Kazemi, Ehsan and Grossglauser, Matthias and Thiran, Patrick",
  	title = "Where to Go from Here? Mobility Prediction from Instantaneous Information",
  	journal = "Pervasive and Mobile Computing",
  	volume = "9",
  	number = "6",
  	pages = "784--797",
  	year = "2013",
  }

  We present the work that allowed us to win the Next-Place Prediction task of the Nokia Mobile Data Challenge. Using data collected from the smartphones of 80 users, we explore the characteristics of their mobility traces. We then develop three families of predictors, including tailored models and generic algorithms, to predict, based on instantaneous information only, the next place a user will visit. These predictors are enhanced with aging techniques that allow them to adapt quickly to the users' changes of habit. Finally, we devise various strategies to blend predictors together and take advantage of their diversity, leading to relative improvements of up to 4%.

• Vincent Etter, Mohamed Kafsi and Ehsan Kazemi. Been There, Done That: What Your Mobility Traces Reveal about Your Behavior. Proceedings of Mobile Data Challenge by Nokia Workshop at the Tenth International Conference on Pervasive Computing, 2012. Winner of the Next Place Prediction challenge
  PDF BibTeX Abstract Conference website

  @inproceedings{etter2012mdc,
  	title="Been There, Done That: What Your Mobility Traces Reveal about Your Behavior",
  	author="Etter, Vincent and Kafsi, Mohamed and Kazemi, Ehsan",
  	booktitle="Proceedings of Mobile Data Challenge by Nokia Workshop at the Tenth International Conference on Pervasive Computing",
  	year="2012"
  }

  Mobility is a central aspect of our life; the locations we visit reflect our tastes and lifestyle and shape our social relationships. The ability to foresee the places a user will visit is therefore beneficial to numerous applications, ranging from forecasting the dynamics of crowds to improving the relevance of location-based recommendations. To solve the Next Place Prediction task of the Nokia Mobile Data Challenge, we developed several mobility predictors, based on graphical models, neural networks, and decision trees, and explain some of the challenges that we faced. Then, we combine these predictors using different blending strategies, which improve the prediction accuracy over any individual predictor.

• Vincent Etter, Ivana Jovanovic and Martin Vetterli. Use of Learned Dictionaries in Tomographic Reconstruction. SPIE Optical Engineering + Applications, International Society for Optics and Photonics, 2011.
  PDF BibTeX Abstract Publisher's website

  @inproceedings{etter2011tomography,
  	title="Use of Learned Dictionaries in Tomographic Reconstruction",
  	author="Etter, Vincent and Jovanovic, Ivana and Vetterli, Martin",
  	booktitle="SPIE Optical Engineering+ Applications",
  	pages="81381C--81381C",
  	year="2011",
  	organization="International Society for Optics and Photonics",
  }

  We study the use and impact of a dictionary in a tomographic reconstruction setup. First, we build two different dictionaries: one using a set of bases functions (Discrete Cosine Transform), and the other that is learned using patches extracted from training images, similar to the image that we would like to reconstruct. We use K-SVD as the learning algorithm. These dictionaries being local, we convert them to global dictionaries, ready to be applied on whole images, by generating all possible shifts of each atom across the image. During the reconstruction, we minimize the reconstruction error by performing a gradient descent on the image representation in the dictionary space. Our experiments show promising results, allowing to eliminate standard artifacts in the tomographic reconstruction, and to reduce the number of measurements required for the inversion. However, the quality of the results depends on the convergence of the learning process, and on the parameters of the dictionaries (number of atoms, convergence criterion, atom size, etc.). The exact influence of each of these remains to be studied.

• Vincent Etter. Semantic Vector Machines. Master Thesis, EPFL/NEC Labs, 2009.
  PDF BibTeX Abstract

  @mastersthesis{etter2009master,
  	title="Semantic Vector Machines",
  	author="Etter, Vincent",
  	school="EPFL",
  	month="4"
  	year="2009",
  }

  We first present our work in machine translation, during which we used aligned sentences to train a neural network to embed n-grams of different languages into an d-dimensional space, such that n-grams that are the translation of each other are close with respect to some metric. Good n-grams to n-grams translation results were achieved, but full sentences translation is still problematic. We realized that learning semantics of sentences and documents was the key for solving a lot of natural language processing problems, and thus moved to the second part of our work: sentence compression. We introduce a flexible neural network architecture for learning embeddings of words and sentences that extract their semantics, propose an efficient implementation in the Torch framework and present embedding results comparable to the ones obtained with classical neural language models, while being more powerful.

Press

I was lucky enough to have some press coverage of my work. I gathered below a few articles talking about my different projects.

• Intelligent keyboard
  • 
  • 
  • 
  • 
  • 
  • 
• Political data analysis
  • 
  • 
  • 
  • 
  • 
  • 
  • 
  • 
  • 
  • 
  • 
• Kickstarter predictions
  • 
  • 
  • 
  • 
  • 
  • 
  • 
  • 
  • 
  • 
• Mobility prediction
  • 
  • 
  • 
  • 
  • 
  • 

More information & Contact

• For more details about my different jobs and internships, have a look at my LinkedIn profile.
• Consult my Google Scholar profile.
• I am the co-founder of Dootix, a software-engineering company based in Switzerland.
• Do not hesitate to contact me by email (vincent@etter.io) or Skype (W00ka1)