Author | Lightnews

Florian Ingels

@fingels.bsky.social

37 followers 44 following 24 posts

Ph.D. in Mathematics Currently postdoc at @bonsaiseqbioinfo.bsky.social, in Lille. Investigating patterns (substructures) in structured data (sequences, trees, graphs) of predominantly biological origin. More at https://fingels.github.io/

fingels.github.io

Posts Media Videos Starter Packs

Pinned

Florian Ingels @fingels.bsky.social · Jun 6

NEW PREPRINT OUT
arxiv.org/abs/2506.04926

The Burrows-Wheeler transform of a string is, in general, more compressible than the original string, especially when it contains lots of repeated subtrings.

The extended BWT allows to deal with multisets of strings.

1/7

Decomposing Words for Enhanced Compression: Exploring the Number of Runs in the Extended Burrows-Wheeler Transform

The Burrows-Wheeler Transform (BWT) is a fundamental component in many data structures for text indexing and compression, widely used in areas such as bioinformatics and information retrieval. The ext...

arxiv.org

1 6 8

Reposted by Florian Ingels

Authueil @authueil.bsky.social · 16d

Un article très intéressant, qui montre que la très grande majorité des salariés trouvent que l'IA générative n'a pas d'utilité pour leur travail (voire que ça produit de la merde), alors que les chefs ont des attentes souvent irréalistes www.blogdumoderateur.com/ia-generativ...

IA générative : 97 % des entreprises peinent à prouver sa valeur business

Toutes les entreprises, ou presque, peinent à prouver la valeur business de l'IA générative. Malgré cela, 87 % des organisations annoncent une hausse de leurs investissements.

www.blogdumoderateur.com

14 75 140

Florian Ingels @fingels.bsky.social · 21d

Ma thèse consistait (en partie) à développer des méthodes formelles mathématiques pour améliorer la véracité des plantes virtuelles, en gros.

Florian Ingels @fingels.bsky.social · 21d

Il y a une autre partie du projet (non montrée sur le poster) qui consistait à développer des modèles informatiques des plantes, pour "simuler" des plantes virtuelles, et entrainer les algorithmes de reconstructions utilisés sur les vraies plantes.

2 2

Florian Ingels @fingels.bsky.social · 21d

C'est en effet le projet au sein duquel j'ai fait ma thèse ! Fabrice et Jonathan sont des anciens collègues!

1 2

Reposted by Florian Ingels

Nicolas Hervieu @nhervieu.bsky.social · Sep 19

Brillantissime Gorafi.

Sous couvert d’un très ordinaire professeur d’archéologie et d’un passionné pour les artefacts rares, il aurait été ainsi au centre d’une cellule d’ultra gauche radicale, dite “Antifa” et aurait opéré en Europe, en Asie et aux Etats-unis entre les années trente et les fin des années soixante. “Il s’est semble-t-il radicalisé tout seul mais il est possible qu’il y ait des complices” a estimé un porte-parole de l’université qui a annoncé qu’elle allait collaborer avec les autorités et le gouvernement de Donald Trump. “On sait qu’il a déclaré devant témoins, “les nazis, je hais ces gars-là” c’est un discours de haine radicale” a affirmé Laura Loomer, une chroniqueuse conservatrice.

6 140 470

Reposted by Florian Ingels

minimaliste13 @minimaliste13.cpesr.fr · Sep 19

C'est le minimum pour un·e candidat·e LR à la présidentielle.

Valérie Pécresse auditionnée dans une affaire de détournement de fonds publics

La région Île-de-France est soupçonnée d’avoir contourné le plafond du nombre de collaborateurs d’élus. Après la perquisition de son bureau en 2020, Valérie Pécresse est entendue par les enquêteurs d…

www.mediapart.fr

2 7 50

Florian Ingels @fingels.bsky.social · Sep 19

Chic alors!

ENQUETE. Aides aux entreprises : le géant Capgemini utilise-t-il le crédit d'impôt recherche pour payer certains de ses salariés inoccupés ?

A qui profitent le crédit d'impôt recherche et ses milliards d'euros d'argent public ? En 2023, Capgemini a perçu, au titre du CIR, 71 millions d’euros. Selon des témoignages et des documents exclusif...

www.franceinfo.fr

Florian Ingels @fingels.bsky.social · Sep 17

Stay tuned 👀

More combinatorics is coming

Florian Ingels @fingels.bsky.social · Feb 10

🎶 Last Christmas,i gave you m̶y̶ ̶h̶e̶a̶r̶t̶
40 pages of delicious combinatorics 🎶

Choose any word W of size m. How many words of size k>m admit W as their smallest lexicographical subword of size m ?

Find out in my latest preprint!

Bonsai Sequence Bioinformatics @bonsaiseqbioinfo.bsky.social · Feb 10

Preprint alert! Can we enumerate how many k-mers lexicographic minimizers have? Yes! (implicitly). Looking forward to applications of this 👌🏻 work by Florian Ingels, @mikael-salson.univ-lille.fr @camillemrcht.bsky.social arxiv.org/abs/2412.17492

1 3

Florian Ingels @fingels.bsky.social · Sep 8

Absolutely phenomenal video about the links between artistic creativity and mathematics (also, combinatorics and enumeration, which are kind of my brand)

This video really captures what I like doing when doing combinatorics

www.youtube.com/watch?v=_BrF...

Exploration & Epiphany

YouTube video by 3Blue1Brown

www.youtube.com

Reposted by Florian Ingels

Camille Marchet ⚡ @camillemrcht.bsky.social · Sep 5

Achievement unlocked: defend your habilitation thesis on the same day than your partner. That was quite a science + celebration day, thanks to all involved 💙✨

6 7 45

Reposted by Florian Ingels

rax ‘levon honkers’ king @raxkingisdead.bsky.social · Aug 26

can’t stop imagining Good Will Hunting but instead of solving graph theory problems on the blackboard he’s writing stuff like 4+3=7

18 14 260

Florian Ingels @fingels.bsky.social · Aug 9

What's the point of doing graph theory if not for pretty pictures ?!

1 1 2

Florian Ingels @fingels.bsky.social · Jul 30

Hi, I am a mathematician in a movie. I have written some non-sense equations on my (black)-board, will use fancy terms that are supposed to make me look smart, and will probably serve as a caution that whatever magical thingy happens in the movie is actually science.

Denis Colombi @uneheuredepeine.bsky.social · Jul 29

Hi, I’m an highschool teacher in a movie. I start my class with a random question on a topic I’ve never discussed before so the new girl can show off. Then if I say something relevant about the main conflict, the bell rang and the students leave even if the class started 5 minutes ago.

Gramsci did something spooky @tontonpogo.bsky.social · Jul 29

Hi, I'm a publicist in a movie. You'll see me in a meeting presenting some passé bullshit to a cadre of executives who will be impressed but have one little remark. I spend a lot of time in my office even if it's not really clear what I actually do. Chances are? I'm also a huge misogynist.

2 23

Florian Ingels @fingels.bsky.social · Jul 27

It seems that Youtube understood a little bit too well the mood of my holidays when suggesting this
www.youtube.com/watch?v=-WY5...

∫ A playlist when math starts to get serious ∫ Dark academia

YouTube video by Dark Dreamers

www.youtube.com

Florian Ingels @fingels.bsky.social · Jul 8

Both are true but waking up this morning was a real challenge haha!

somehow I live by this motto from an old Absinthe commercial

1 3

Florian Ingels @fingels.bsky.social · Jul 8

I do not recommend staying up late until 3 a.m. because your brain can't accept that you can postpone solving that one collegue's problem he sent you via email during the evening and that you read around midnight

but it's solved!

1 4

Reposted by Florian Ingels

Bonsai Sequence Bioinformatics @bonsaiseqbioinfo.bsky.social · Jul 3

Bonsai goes camping by the shore ☀️

Members enter the sea as the sun goes down

A picture of the group with pretty dusk colors and the beach in the background

1 5 10

Reposted by Florian Ingels

Camille Marchet ⚡ @camillemrcht.bsky.social · Jun 30

Thomas Baudeau defended his thesis on Studying the properties of viral long reads mapping methods - congrats docteur Baudeau you'll be deeply missed in the team. I'm very glad I got the chance to work with you. Thomas is also on the lookout for a postdoc 👀

3 7

Reposted by Florian Ingels

bioRxiv Bioinfo @biorxiv-bioinfo.bsky.social · Jun 17

REINDEER2: practical abundance index at scale https://www.biorxiv.org/content/10.1101/2025.06.16.659990v1

7 8

Florian Ingels @fingels.bsky.social · Jun 6

That's it ! We hope our results will motivate further research into this question of finding an optimal decomposition.

The preprint is here: arxiv.org/abs/2506.04926

7/7

Decomposing Words for Enhanced Compression: Exploring the Number of Runs in the Extended Burrows-Wheeler Transform

arxiv.org

Florian Ingels @fingels.bsky.social · Jun 6

C) We exhibit an infinite family of strings for which the ratio of the numbers of runs between the worst and the best decompositions is unbounded, i.e. can be arbitrarily large.

So, potentially lot to lose !

6/7

Florian Ingels @fingels.bsky.social · Jun 6

B) For any string, the best possible decompositions achieve a number of runs that is bounded by something that DOES NOT depend on the size of the string, but only on the size of the alphabet and the minimal size we allows strings in the decompositions to be.

So, potentially lot to gain.

5/7

1 1

Florian Ingels @fingels.bsky.social · Jun 6

A) The number of possible decompositions, even with our minimal constraints, is exponential, so brute force is doomed to failure.

Not very surprising, but still.

4/7

Florian Ingels @fingels.bsky.social · Jun 6

We add minimal constraints on the decomposition to avoid trivial answers, not very useful in practice.

In the article, we do not solve the problem nor prove its possible NP-completeness. However, we prove 3 results that, in our eyes, justify the interest of the problem.

So, here there are :

3/7

Florian Ingels @fingels.bsky.social · Jun 6

We consider a very simple, natural question, that somehow we couldn't find in the literature :

is there an optimal way to decompose a string so that the eBWT of this decomposition is more compressible that the BWT of the original string ?

2/7