Alberto Acuto

A collective effort from @liverpooluni.bsky.social signal processing team

A collective effort from @liverpooluni.bsky.social signal processing team

In a few words: we find that accounting for real world costs can change significantly the decision making in network defence, forcing the autonomous agent to make some network concessions in order to preserve real-world assets.

Accounting for the Impact of Real-World Data and Costs in Autonomous Cyber Defence In cyber operations, a robust and versatile cyber defence system is vital. The introduction of machine learning techniques has reduced the cognitive load on human analysts who protect these networks, ...

🧪 at IEEE CSR instead we won the best paper award with a paper on autonomous cyber defence.

We investigated how much including real world costs and data can impact standard network defence.

Check here: ieeexplore.ieee.org/document/111...

GitHub - UoL-SignalProcessingGroup/StoneSoupNUTS: Tutorial repository for the paper Stone Soup goes NUTS Tutorial repository for the paper Stone Soup goes NUTS - UoL-SignalProcessingGroup/StoneSoupNUTS

You can also check the tutorial on Github!
github.com/UoL-SignalPr...

#engineering #particlefilter #bayesian

Stone Soup Goes NUTS: Adding Proposals and the No-U-Turn Sampler to Stone Soup Particle filters are essential for state estimation in non-linear and non-Gaussian systems, with performance hinging on effective proposal distributions. This paper presents the implementation of Kalm...

🧪 A couple of news:
At ISIF FUSION 2025 conference we presented our latest involvement with Stone Soup integration: particle filter proposals.

We have presented a paper highlighting the benefit of particle filtering solution in highly non-linear systems
ieeexplore.ieee.org/document/111...

I think there is no easy answer for this. If you limit the comments then these people would go where they can express themselves and those might be echo chamber for further "radicalisation".
I think as well it will get worse, and I am not sure how this would improve

I was blown out when yesterday Liverpool Echo was reporting a terrible crime in a city centre hotel and all the comments were about: "is this a migrant hotel?" "is it a normal hotel or migrant?"
Is that the first thing many people thinks and post.

Chiaro momento: "Better safe than sorry". Anche se ho la netta sensazione che qualche cittadino del North Dakota si sia impaurito del rischio Tsunami a Bismarck

a cartoon character says my goodness what an idea why didn 't i think of that ALT: a cartoon character says my goodness what an idea why didn 't i think of that

full crazy scorpio timeline

The website vs app split for Bluesky vs Threads captures the difference in user base and usage patterns beautifully.

Bluesky = knowledge sector workers who have it open all day in a tab next to their emails. “Wait what was I doing again?”

Threads = normies watching recipe videos on the commute.

Radio astronomer here! Fun fact of the day- the signals we detect are so faint that a cell phone ON THE MOON would be one of the brightest things in the radio sky! 🤯

This is why near radio telescopes you’ll see signs to turn off your cell phones (here’s the one by the VLA in New Mexico).

🧪🔭

indeed, the app store are good indicators of communities moving. But I reckon it will take time and opinion leaders to move across (or just cross-posting) to populate the various bubbles. Interesting to live it for the first time "in real time" (or until it lasts)

a naso direi al primo marotta league. Anche se non so se questi server resisteranno ad un normale weekend di calcio

cool that'd be great!

any chance of reposting here some paper highlights or discussions? It would be extremely interesting!

Il voto via posta aiuta, perché ti viene incontro se il giorno delle votazioni non puoi fisicamente andare. Qui in UK lo abbiamo usato per le votazioni locali e ci siamo trovati bene, 5 minuti, tutto in busta e spedito. Certo che per le europee noi in UK non potevamo votare se non tornando in Italia

yeah, sorry for being a data nerd

You know what would it be cool?
A platform/visualisation of where the users come from (geographically) and what are the topic they are talking about and how that evolves over time.

[I'm not going to lie, most of the liked posts will be about food and football]

Outside of work, I particularly enjoy (and moan) about politics, science, social causes (even if I might not interact much). I fell victim of the well known "academic syndrome" and I started #bouldering, re-aligning with my generation skillset.

A side from these projects, I collaborate on a number of projects involving tracking, sensor and data fusion in a variety of domains.
I am interested in #machinelearning #ML, #gaussianprocesses and #bayesianoptimisation (some papers might come around in the new year).

You can see the results here: scholar.google.com/citations?vi... #astrophysics #cosmology #lss

In my PhD, long long long time ago, I was working with cosmological simulations, the #BAHAMAS simulations, trying to make sense of baryons and their role in cosmological predictions. We published a #paper on the halo model investigating how much halo boundary assumptions and baryons were impacting

Or with more general, more advanced deep RL: scholar.google.com/citations?vi... Work done by CDT PhD student Wanrong Yang (under review), #paper

Some things I've been doing recently are about #ReinforcementLearning in #cybersecurity and network security. Like working with DSTL's #YawningTitan see here: scholar.google.com/citations?vi... (work presented at CAMLIS 2023) #paperday

Let's do some introductions: I am Alberto, I work at the University of Liverpool, in the Signal processing group.

Hello world! #Bluesky