Faithfulness of GNN explanations isn’t one-size-fits-all🧢 Our last @iclr-conf.bsky.social paper breaks it down across: 1. Evaluation metrics 2. Model implementations 3. OOD generalisation w: Antonio L. @looselycorrect.bsky.social @andreapasserini.bsky.social 1/5

is this AGI?

Submission deadline is extended to March 20 for submitting your paper to our #CVPR2025 workshop on Uncertainty Quantification for Computer Vision. Looking forward to see your submissions on recognizing failure scenarios and enabling robust vision systems! More info: uncertainty-cv.github.io/2025/

Are you at AAAI in Philadelphia and interested about #tensor-factorizations or #circuits or even both? Then join us today at our tutorial: "From tensor factorizations to circuits (and back!)" Details and materials here april-tools.github.io/aaai25-tf-pc... Time 4:15pm - 6:00pm, Room 117

Finally, LOgically COnsistent (LoCo) LLaMas can outperform solver-based baselines and SFT! I thank @nolovedeeplearning.bsky.social and @looselycorrect.bsky.social for the guidance in realizing this project, get in touch or come to chat in Singapore! arxiv.org/abs/2409.13724

🔊 Super excited to announce the first ever Frontiers of Probabilistic Inference: Learning meets Sampling workshop at #ICLR2025 @iclr-conf.bsky.social! 🔗 website: sites.google.com/view/fpiwork... 🔥 Call for papers: sites.google.com/view/fpiwork... more details in thread below👇 🧵

An incredible opportunity for the right person: The Santa Fe Institute is seeking applications for full-time resident faculty positions at all academic levels For me, this has been the best possible job ever. More info here: santafe.edu/about/jobs/r...

Protein instability is the main driver of inherited genetic conditions, according to the largest map of protein variants to date − Human Domainome 1.0 🔎 This dataset can help predict how proteins behave, paving the way for new treatments 💊 www.sanger.ac.uk/news_item/hu...

🦕 Special track at NeSy dedicated to KGs and Ontologies!! Have a look 🙌

My book is (at last) out, just in time for Christmas! A blog post to celebrate and present it: francisbach.com/my-book-is-o...

When building human allied AI systems, it is imperative to think of who is in control. In our blue sky paper at #AAAI, we argue that a deeper understanding of whether human is in the loop or if AI is in the loop is essential! arxiv.org/pdf/2412.14232

Neural Concept Binder @WolfStammer

Today Google's Gemini asked me if I wanted help with my messages to be a more effective communicator. I clicked on "yes". A button appeared that read "tips for procrastination". Here is what I got... Apparently procrastinating is a great skill to have. Thanks Google!! @garymarcus.bsky.social

Interested in estimating posterior predictives in Bayesian inference? Really want to know if your approximate inference "is working"? Come to our poster at the NeurIPS BDU workshop on Saturday - see TL;DR below.

I think one important aspect missing from the current #NeSy papers & talks is that a (#probabilistic) #neuro #symbolic approach to #AI and #ML does not bring full interpretability and trustworthiness but can actually give a false sense of reliability. See all the works on #reasoning shortcuts 1/2

The #NeurIPS experience is about to start! ✈ Drop me a line if you want to chat about #neurosymbolic reasoning #shortcuts, human-interpretable machine #concepts, logically-consistent #LLMs, or human-in-the-➰ #XAI! See you in Vancouver!

For those attending #NeurIPS2024 go to UNIREPS @unireps.bsky.social workshop to know more about representations similarity. Nice work led by @beatrixmgn.bsky.social 🌟

I made a starter pack with the people doing something related to Neurosymbolic AI that I could find. Let me know if I missed you! go.bsky.app/RMJ8q3i

We're doing a deep dive into possession value models. While VAEP, g+, PV & OBV are conceptually identical, they make different design choices. We look at using (no) goal vs. xG as the target variable. w/Lorenzo Cascioli @probberechts.bsky.social Lode Van Tente dtai.cs.kuleuven.be/sports/blog/...

The best one could be to speed up progress in AI at this moment of history is to shape whatever is the community's reward function so that a fraction of the compute and brain power currently knob-turning GPTs to gain 0.25% on whatever unrealistic benchmark try SOMETHING NEW.

if you ever wondered how to concisely represent causal models, don’t miss my talk on causal abstraction tomorrow at @ellisamsterdam.bsky.social

If you review for a #ML conference like @iclr-conf.bsky.social or @neuripsconf.bsky.social, YOU HAVE A RESPONSIBILITY TO REPLY TO THE AUTHORS. If the rebuttal doesn't address your concerns explain why. But giving a score of 2-3 then ghosting the authors is super rude. I say this as an AC. #MLSky

Kush R. Varshney: The AI Alignment Problem, Limits, and Solutions - in case you missed it -

Is anyone familiar with efficient tools/libraries for over/under-approximating of the volume of convex polytopes (in H-representation)?

🚨 Interpretable AI often means sacrificing accuracy—but what if we could have both? Most interpretable AI models, like Concept Bottleneck Models, force us to trade accuracy for interpretability. But not anymore, due to Concept-Based Memory Reasoner (CMR)! #NeurIPS2024 (1/7)

The Uncertainty in Artificial Intelligence conference (UAI) is on 🦋! I'm a bit biased, but this is a fantastic conference for #probabilisticML, #causality, #causalML, #tractable #probabilistic #models, #imprecise probabilities, #reasoning, #neurosymbolic approaches and more. #causalSky #statSky

The Human-Centered Explainable AI (HCXAI) Workshop will be back for the 5th time in 2025!

@ropeharz.bsky.social forced me to do this starter pack on #tractable #probabilistic modeling and #reasoning in #AI and #ML please write below if you want to be added (and sorry if I did not find you from the beginning). go.bsky.app/DhVNyz5

I think we saw this in the AI/ML sphere, not just in politics …

useful also to give you motivation to submit your #ICML papers

I'm looking for PhD applicants who have expertise in Gaussian processes and/or Transformers for an exciting PhD project If this sounds interesting, application deadline for funding is 3/12 Please share with people you think this might be relevant to! oatml.cs.ox.ac.uk/apply.html

Very proud to be a "paranymph" for my longtime friend Jaap (we're friends since we were 6)! He has done excellent work on Language Model interpretability before it was cool. Go check it out!

🔥🔥** cough cough **🔥🔥 (being summoned in a circle of fire) here is a tutorial for you slideslive.com/38992775/pro... and here is a modern take from the perspective of tensor factorisations arxiv.org/abs/2409.07953 tensor factorisations and circuits give you exact gradients and **any** marginals

Yeah, many people miss this: Gumbel softmax is not applicable when you want to branch! You can hack your way around it, but bias compounds very rapidly

Amazing opportunity for #Neurosymbolic folks! 🚨🚨🚨 We are looking for a Tenure Track Prof for the 🇦🇹 #FWF Cluster of Excellence Bilateral AI (think #NeSy ++) www.bilateral-ai.net A nice starting pack for fully funded PhDs is included. jobs.tugraz.at/en/jobs/226f...

What is common knowledge in your field but shocks outsiders? The rigorous definition of probability does not define probabilities as relative frequencies. The frequency nature of probability is hence a theorem (law of large numbers), but not the definition of probability.

Newsweek asked me for thoughts on the safety of chatbots. They published all of them except for one sentence: "I’d rather have kids talking to chatbots than listening to Donald Trump." ...they excluded that one for some reason. Article below. Beware spammy ads. www.newsweek.com/are-ai-chatb...

Medically adapted foundation models (think Med-*) turn out to be more hot air than hot stuff. Correcting for fatal flaws in evaluation, the current crop are no better on balance than generic foundation models, even on the very tasks for which benefits are claimed. arxiv.org/abs/2411.04118

Luc De Raedt: How to Make Logics Neurosymbolic - in case you missed it -

Neat ICML #XAI paper, with actual user studies! 🎩 "Our theory posits that absent explanation humans expect the AI to make similar decisions to themselves, and that they interpret an explanation by comparison to the explanations they themselves would give." proceedings.mlr.press/v162/yang22c...

I've seen things you people wouldn't believe. Attacks from reviewers on fire off the shoulders of #OpenReview. I watched logic fallacies glitter in the dark near @iclr-conf.bsky.social All those moments will be lost in time, like tears in the next resubmission.  Time to die. #ML #Ai #PhDlife

Writing a rebuttal right now. I'm tempted to reply with this:

It used to be just reviewer 2. Now it's reviewers 2, 3, and 4. And we gave them LLMs, for extra giggles. Thank goodness for reviewer 1. Please keep up the good work, we need you!

NeSy is on Bluesky! bsky.app/profile/nesy... Check out the 2025 edition of the NeSy conference: 2025.nesyconf.org General Chair: @lgilpin.bsky.social Programme Chairs: @e-giunchiglia.bsky.social @emilevankrieken.com @pascalhitzler.bsky.social @tarekbesold.bsky.social @arturgarcez.bsky.social

🔥 Follow this page 👉 bsky.app/profile/nesy... to stay updated with the 2025 Neuro-symbolic AI conference 🦕🦕🦕

It's tough to gain visibility as a young researcher, and it's job market season! Are you a theoretical computer science PhD/postdoc on the job market? I don't have a crazy juge audience but I'll try to help a bit: fill this form, and I'll tweet your pitch and info! docs.google.com/forms/d/e/1F...

Just realized BlueSky allows sharing valuable stuff cause it doesn't punish links. 🤩 Let's start with "What are embeddings" by @vickiboykis.com The book is a great summary of embeddings, from history to modern approaches. The best part: it's free. Link: vickiboykis.com/what_are_emb...

My amazing collaborators @sbadredd.bsky.social and @e-giunchiglia.bsky.social have landed! We're working on a Python library for accessible Neurosymbolic Learning called ULLER, that we plan to release soon. White paper: arxiv.org/abs/2405.00532