Gabriele D'Acunto @officiallydac.bsky.social · 30/04/2026New algorithm (mixture-calsep) learns the structure from data. Tested on synthetic benchmarks and real financial data. 000
Gabriele D'Acunto @officiallydac.bsky.social · 30/04/2026Multiple agents. Heterogeneous causal models. No shared ground truth. We formalize their integration and alignment using network sheaves. Global coherence emerges from local consistency conditions on causal embedding and abstraction maps between agents. 100
Gabriele D'Acunto @officiallydac.bsky.social · 30/04/2026New preprint out! Networks of Causal Abstractions: A Sheaf-theoretic Framework arxiv.org/abs/2509.25236 100
Gabriele D'Acunto @officiallydac.bsky.social · 05/02/20263 x ✅ causal abstraction: arxiv.org/abs/2602.026... functorial signal processing: arxiv.org/abs/2510.112... probabilistic topological modeling: arxiv.org/abs/2510.129... #ICASSP2026 000
Gabriele D'Acunto @officiallydac.bsky.social · 18/12/2025📢New publication in the Proc. of the 59th Asilomar Conference on semantic alignment and communication of AI agents using network sheaves! 🤖 arxiv.org/pdf/2512.03248 030
Gabriele D'Acunto @officiallydac.bsky.social · 16/12/2025New publication in the Proc. of the 59th Asilomar Conference from the SPAICOM Lab for probabilistic modeling over higher-order topological domains arxiv.org/pdf/2512.03727 000
Gabriele D'Acunto @officiallydac.bsky.social · 13/07/2025Ready to fly to Vancouver 🛸 Eager to talk about causal abstraction at @icmlconf.bsky.social #icml2025 #causal #ML 010
Gabriele D'Acunto @officiallydac.bsky.social · 08/05/2025preprint: arxiv.org/pdf/2503.11718arxiv.org 010
Gabriele D'Acunto @officiallydac.bsky.social · 08/05/2025📢"The Relativity of #Causal Knowledge" has been accepted at the 41st edition of the Conference on #Uncertainty in #AI. Thanks to @clabat9.bsky.social for his essential contribution and commitment. Camera-ready coming soon 👽 #uai2025 #ML 100
Gabriele D'Acunto @officiallydac.bsky.social · 03/05/2025#ICML 2025 here we come! **Causal Abstraction Learning based on the Semantic Embedding Principle** Stay tuned for the camera-ready version (spoiler: new results are on the way!) Fabio Massimo Zennaro, @yfelekis.bsky.social and Paolo Di Lorenzo #causal #ML 020
Gabriele D'Acunto @officiallydac.bsky.social · 22/03/2025📢📢📢"The Relativity of Causal Knowledge" THANK YOU to @clabat9.bsky.social for working side by side with me on this exciting project and making this article possible. Big thanks also to @hansmriess.bsky.social and Fabio Massimo Zennaro for their valuable feedback on an earlier version of the paper. 072
Reposted by Gabriele D'AcuntoYorgos @yfelekis.bsky.social · 04/02/2025Hot off the press!💥 arxiv.org/abs/2502.00407 173