Category Archives: Trip

Conferences, Trip

European Actuarial Journal Conference (in Lisbon)

Leave a comment

Next week, I will be at the European Actuarial Journal Conference, at the Lisbon School of Economics and Business, EAJ’24.

I will give a talk on calibration of actuarial models, based on our recent paper with Agathe Fernandes Machado, Emmanuel Flachaire, Ewen Gallic and François Hu, mainly “Probabilistic Scores of Classifiers, Calibration is not Enough” (as well as recent work on recalibration). Slides are available.

In binary classification tasks, accurate representation of probabilistic predictions is essential for various real-world applications such as predicting payment defaults or assessing medical risks. The model must then be well-calibrated to ensure alignment between predicted probabilities and actual outcomes. However, when score heterogeneity deviates from the underlying data probability distribution, traditional calibration metrics lose reliability, failing to align score distribution with actual probabilities. In this study, we highlight approaches that prioritize optimizing the alignment between predicted scores and true probability distributions over minimizing traditional performance or calibration metrics. When employing tree-based models such as Random Forest and XGBoost, our analysis emphasizes the flexibility these models offer in tuning hyperparameters to minimize the Kullback-Leibler (KL) divergence between predicted and true distributions. Through extensive empirical analysis across 10 UCI datasets and simulations, we demonstrate that optimizing tree-based models based on KL divergence yields superior alignment between predicted scores and actual probabilities without significant performance loss. In real-world scenarios, the reference probability is determined a priori as a Beta distribution estimated through maximum likelihood. Conversely, minimizing traditional calibration metrics may lead to suboptimal results, characterized by notable performance declines and inferior KL values. Our findings reveal limitations in traditional calibration metrics, which could undermine the reliability of predictive models for critical decision-making.

Conferences, Trip

Samuel が横浜に到着

Leave a comment

After defending his PhD last week, Samuel just arrived in 横浜市, at the International Joint Conference on Neural Networks (IJCNN’24), that will take place at the IEEE World Congress on Computational Intelligence (WCCI).

He will present our recent work on Boarding for ISS: Imbalanced Self-Supervised Discovery of a Scaled Autoencoder for Mixed Tabular Datasets,

The field of imbalanced self-supervised learning, especially in the context of tabular data, has not been extensively studied. Existing research has predominantly focused on image datasets. This paper aims to fill this gap by examining the specific challenges posed by data imbalance in self-supervised learning in the domain of tabular data, with a primary focus on autoencoders. Autoencoders are widely employed for learning and constructing a new representation of a dataset, particularly for dimensionality reduction. They are also often used for generative model learning, as seen in variational autoencoders. When dealing with mixed tabular data, qualitative variables are often encoded using a one-hot encoder with a standard loss function (MSE or Cross Entropy). In this paper, we analyze the drawbacks of this approach, especially when categorical variables are imbalanced. We propose a novel metric to balance learning: a Multi-Supervised Balanced MSE. This approach reduces the reconstruction error by balancing the influence of variables. Finally, we empirically demonstrate that this new metric, compared to the standard MSE: i) outperforms when the dataset is imbalanced, especially when the learning process is insufficient, and ii) provides similar results in the opposite case.

Trip

Incertitude(s)

Leave a comment

Cet été, le 7 août, je participerai au marathon des sciences, dans le cadre du 30ème Festival d’Astronomie de Fleurance, sur le thème “Incertitude(s)”. Malheureusement, comme nous ne rentrons pas en France cet été, mon intervention serra virtuelle, ou disons à distance. J’essayerai de mettre à la rentrée la vidéo de mon intervention, sur le thème assurance : des incertitudes individuelles aux certitudes collectives,

L’assurance repose sur l’idée de la mutualisation des risques, un grand nombre de personnes contribuant à alléger les malheurs d’un petit nombre d’entre eux. Ce transfert de risque se fait par l’intermédiaire d’un assureur qui, sur la base d’un contrat (dit aléatoire), vend la promesse d’une indemnité financière en cas de survenance d’un évènement incertain à la signature. L’assureur se doit d’imaginer les scénarios futurs, et de quantifier du mieux possible leurs probabilités afin de pouvoir satisfaire ses engagements, de telle sorte que les contributions des assurés permettent effectivement d’indemniser ceux qui ont eu un accident. Car le risque et l’incertitude ne disparaissent jamais, les actuaires se contentent d’essayer de les maîtriser autant que possible.

Trip

Sharing pictures from holidays in the Canadian Rockies (with R)

Leave a comment

My kids have a very popular blog (at least among their grandmothers) where they frequently post pictures from everyday’s life (since they live 5000km from them), as well as pictures taken from holidays. This afternoon, I tried to used the popupImage function from the leaflet package to post pictures, on a map (to explain where we spent our holiday this summer). This post is just to keep tracks of that code.

First, we need to load the appropriate R packages

library(leaflet)
library(mapview)

Then, we take a picture, and we locate it, for instance Mirror Lake (on the trail to Lake Agnes). Since leaflet uses openstreetmap, I recommend to use it also for location (and not google maps… coordinates can be slightly different)

df=data.frame(lat =51.41603, long=-116.23946,
nom = "Miror Lake",photo="http://freakonometrics.free.fr/jaspeR/_DSC5967.jpg")

I guess you can also use the metadata if you take pictures with a cell phone, and you add the location… but I am (very) old fashioned, and still use a camera to take pictures. Then you can add a dozen pictures

df=rbind(df, data.frame(lat =51.4164, long=-116.2442,
nom = "Lake Agnes",photo="http://freakonometrics.free.fr/jaspeR/_DSC6003.jpg"))
df=rbind(df, data.frame(lat =51.3215642,long=-116.193718,
nom="Moraine Lake",photo="http://freakonometrics.free.fr/jaspeR/_DSC5957.jpg"))

From that dataframe, we need two kinds of information: the location, and the url of the picture,

data_df=df[,c("lat","long")]
images = as.character(df$photo)

Then we can create the leaflet map (sorry for typos, but wordpress converts the > symbol into some “>” characters… which makes R pipe operator hard to read)

m = leaflet(data_df) %>%
  addTiles() %>%
  addCircleMarkers(
    fillOpacity = 0.8, radius = 5,
    lng = ~long, lat =~lat, 
    popup = popupImage(images)
  )

and export it (in a nice html file)

library(htmlwidgets)
saveWidget(m, file="jaspR.html")

Conferences, Trip

Annual UCSB InsurTech Summit

Leave a comment

Just a quick post to mention that an Insurtech Summit will be organized in May 2019, on Friday 3rd, by Mike Ludkovski, and I will be there, with Francois Millard (Vitality Group), Adam Tashman (Carpe Data, Santa Barbara), Emil Valdez (University of Connecticut), and Howard Zail (Elucidor, LLC, New York City). That will be nice… I will actually also give a talk on the Monday before at the actuarial seminar !

Conferences, Research, Trip

Rencontres Mutualistes

1 Comment

Lundi et mardi, je serais a Beaune, en Bourgogne, pour les premières rencontres mutualistes. On m’a demande d’intervenir en ouverture de la seconde journée, sur le thème “segmentation et mutualisation”.

Les slides sont dès à  présent en ligne. Comme j’ai peu de temps, je reviendrais sur les grands principes de la tarification et du rôle de l’actuaire. J’ai ensuite pense qu’une discussion autour du graphique suivant pourrait être intéressante, en particulier sur les deux bornes, inférieure (‘average pricing‘) et supérieure (‘perfect pricing‘)

On finira avec un rapide retour sur les pricing games, pour conclure.