Assoc. Prof. (SLecturer) @imperialcollege. Data Scientist. Topological data analyst. Algebraic statistician. Language enthusiast. Sourdough baker. Mom.
London, UK
Joined June 2020
- Tweets71
- Following81
- Followers537
- Likes223
🧠✨ The deformation of LLM latent spaces under adversarial influences can be characterized by persistent homology and you can find out all about it here 👉 arxiv.org/abs/2505.20435 and at @iclr_conf! 🚀 Joint with Aideen Fay, Inés García-Redondo, Qiquan Wang, Haim Dubossarsky
Oxford meetup of LoG: Had a great time presenting some of the work my fantastic group does on using tropical geometry to study NNs and using NNs for theorem discovery in algebraic geometry 🤓
We are hiring a postdoc at Imperial as a full-fledged affiliate of our new Erlangen Programme for Mathematical & Computational Foundations of AI Hub! Check it out, spread the word, and apply! 🤓👇
imperial.ac.uk/jobs/search-j…
Happy international women's day! I am proud to be featured in Imperial's Women in Science week on this occasion.
Dr Anthea Monod is a mathematician and Imperial's Node Lead for the new AI Hub, building the foundations for #MachineLearning models
She's worked in the best statistics depts in the world after a brief (but successful) career as a trader! 📈
@ImperialMaths
ALT A portrait of Anthea Monod, with the quote: "I recently came back to work from maternity leave, which is not only just about the challenges of being a woman, but also the challenges of carrying responsibilities."
Anthea Monod retweeted
@imperialcollege is at the forefront of AI innovation, co-leading a new £10m AI Hub with @UniofOxford 🤖
With Dr Anthea Monod (@AntheaMonod) from @ImperialMaths, the hub aims to unravel #ML models with a focus on pure mathematics 🧮
imperial.ac.uk/news/250854/n…
Thrilled to be leading the Imperial node of this new Hub! 👇 – at Barnet, London
A new Oxford AI research hub funded by the Engineering and Physical Sciences Research Council (@EPSRC) and led by Professor Michael Bronstein (@mmbronstein) to propel transformative AI innovations.
Read here: cs.ox.ac.uk/news/2286-full.h…
#compscioxford
ALT A blue image of a computer board with a dark blue speech bubble over it with white text reading "'This new research programme will break barriers between different fields and bring a diverse cohort of leading UK experts rarely seen together with the purpose of strong cross-fertilisation.' Professor Michael Bronstein". There is another smaller blue box at the bottom right reading "@compscioxford #compscioxford". On the bottom left there is a circular image of Professor Michael Bronstein and white text next to it reading "Image credit: Gabriella Sbordone".
I gave a talk in Geneva last week and almost nobody knew about my Swiss citizenship so voilà: for those who don’t know, I am Swiss. Ça fait du bien d’être rentrée chez soi ❤️🇨🇭🎿 – at Leukerbad, Switzerland
Did you know that sheaves can be used for community detection? My PhD student Arne Wolf did it in this paper 👉 arxiv.org/abs/2310.05767 which got accepted to this conference! 👇
🌟LAST CHANCE to submit your work to #COMPLEXNETWORKS 2023! 📊📈 Join us for networking, innovation, and FUN on the stunning French Riviera! 🏖️🇫🇷 Submit your abstract today and be part of something EPIC! 🚀
complexnetworks.org/
#NetSci #DataScience #DataNatives #Databeers
I was honored to lead a research project at WinCompTop @EPFL with the most brilliant young women researchers, Robin Belton, Ximena Fernández @pi_ene, Katharina Limbeck, Renata Turkeš, Qiquan Wang! TDA is headed in super exciting directions with this kind of talent!
Writing, staring, erasing, writing again.
Laughing, getting frustrated.
Conjecturing, proving, disproving.
Planning projects together.
Creating bonds.
This week at #WinCompTop has been one of the most beautiful research experiences I've ever had.
Hopefully not the last!
Rank functions do a great job at inference for both single and multiparameter persistence! Take a look at the super cool case studies on both simulated and real data, which we validated with stability theorems for rank functions 🤩
arxiv.org/abs/2307.02904
My paper with @YUEQI_IMPERIAL @vlontzos Luca Schmidtke and @BernhardKainz1 has been published at Information and Inference, check it out! 👇
academic.oup.com/imaiai/arti…
We propose a new dilation invariant bottleneck distance for PH to address metric distortion 🤓
Check out the full story for the paper I just tweeted about below! 👇🤓
New paper! A wonderful collaboration with the brilliant @AnnaSongMaths and @AntheaMonod @ImperialMaths on Morse Theory for Distance Functions to Surfaces. This project provides theoretical underpinnings for signed distance persistence homology SDPH arxiv.org/abs/2306.14716 1/10
👇I missed announcing the last couple of papers, which doesn't mean that I wasn't proud of those as well, but this one by @AnnaSongMaths and @AmbroseYim is awesome and awe-some 🤓
arxiv.org/abs/2306.14716
Anthea Monod retweeted
TDA centre meeting @OxUniMaths. Current research talks by the incredible @pi_ene @DrKateTurner @AntheaMonod @Francesco_dgv
You can now read about this online at Discrete & Computational Geometry!
link.springer.com/article/10…
Ever wanted to understand how tree topologies behave in the tropical geometric version of phylogenetic tree space?
Well, arxiv.org/abs/2010.06158
Keywords: BHV geodesic; clades; tree topology; tropical geometry; tropical line segment
Did you know that, although geodesics are very much not unique on the space of persistence diagrams, you can still get unique Fréchet means? Check out the paper on the arXiv by my PhD student @YUEQI_IMPERIAL 👇
arxiv.org/abs/2207.03943
However, to construct these unique Fréchet means, you need to be away from the diagonal by a certain amount. There is of course lots of previous work on doing this kind of "signal processing" but the approach is pretty different from what we need.