Theoretical #Engineer and #Computational Innovator in #physics of #complex #systems, #network science and machine #intelligence. Chair Professor @Tsinghua_Uni

Joined June 2018
139 Photos and videos
Research doesn’t always want to stay in papers. @netscisociety I’m a co-founder of SparseMind AI, a startup born directly from my research @Tsinghua_Uni Accumulating theoretical evidence and practical validations, our work is ready to move from research to real-world impact.
“AI's future is not bigger, but smarter,” says Prof. Carlo Cannistraci. His project, SparseMind AI, won third at the HICOOL Global Entrepreneur Competition, algorithmically removing up to 99% of model bulk. This “sparse revolution” makes AI more accessible & sustainable for all.
1
1
3
227
Last week in Boston for the @netscisociety #netsci2026 I had the honour to offer a talk on my theory for link prediction and its application to neural networks. Thanks to @HirokiSayama to feature the work.
Carlo Vittorio Cannistraci develops link prediction methods using network automata #NetSci2026 @NetSciConf @netscisociety
1
1
3
352
#NSIA @NSIAsatellite Natasa Przulj,Professor of Computational Biology, MBZUAI; Professor, Computer Science UCL, Inventor of graphlets kick-starts the inaugural edition of the Network Science Informs AI satellite @netscisociety #netsci26
124
1:4 NVIDIA semi-structured sparse training and inference acceleration will come soon possible @elonmusk @nvidia @NVIDIAAI
3
46
Towards @nvidia 1:4 Semi-Structured 75% Sparsity via Cannistraci–Hebb N:M Dynamic Sparse-to-Sparse Training See the main results. @netscisociety
2
108
Do not miss the opportunity to submit an abstract to the first satellite that discusses how "Network science informs AI" @NetSciConf 2026.
📢 ABSTRACTS OPEN! Round 1 submissions are now open for the inaugural Network Science Informs AI (NSIA) satellite at @NetSciConf ! 🧠🕸️ We explore network topology, sparsity & dynamics as drivers for next-gen AI. 🗓️ Deadline (Round 1): March 9 (AoE) 📍 Organized by: @CCNI_lab 🔗 Submit here: openreview.net/group?id=NSIA… For more info: network-science-ai.github.io… Follow us on LinkedIn: linkedin.com/company/network… #NetworkScience #AI #DeepLearning #ComplexSystems #NSIA2026 #NetSci2026
4
91
I am happy to announce that this year @NetSciConf 2026, we organize a satellite on "Network Science informs AI".
🚀 Announcing the inaugural NSIA Satellite at #NetSci2026! We’re bridging the gap between Network Science and AI to build the next generation of intelligent systems. Join us in Boston for a deep dive into topology, brain-inspired connectivity, and emergent learning. 📅 June 1, 2026 📍 Boston, USA 📝 Call for Abstracts: Feb 12 – March 9 (Round 1) Ready to redefine AI? Submit your work and join the dialogue. 🔗network-science-ai.github.io… #NSIA2026 #NetworkScience #AI #MachineLearning #ComplexSystems
3
80
Research doesn’t always want to stay in papers. @netscisociety I’m a co-founder of SparseMind AI, a startup born directly from my research @Tsinghua_Uni Accumulating theoretical evidence and practical validations, our work is ready to move from research to real-world impact.
“AI's future is not bigger, but smarter,” says Prof. Carlo Cannistraci. His project, SparseMind AI, won third at the HICOOL Global Entrepreneur Competition, algorithmically removing up to 99% of model bulk. This “sparse revolution” makes AI more accessible & sustainable for all.
1
1
3
227
Epitopological Learning and Cannistraci-Hebb Network Shape Intelligence Brain-Inspired Theory for Ultra-Sparse Advantage in Deep Learning Y Zhang, J Zhao, W Wu, A Muscoloni, CV Cannistraci The Twelfth International Conference on Learning Representations (ICLR) 2024, 2024
1
98
Brain network science modelling of sparse neural networks enables Transformers and LLMs to perform as fully connected Y Zhang, D Cerretti, J Zhao, W Wu, Z Liao, U Michieli, CV Cannistraci Thirty-Ninth Conference on Neural Information Processing Systems (NeurIPS 2025), 2025
1
56
Curious to investigate interpretable motifs also in dynamic sparse regimes. Open to discuss!
1
60
Long time I am not on X. I thank the organisers of the @ICCVConference "Beyond Euclidean: Hyperbolic and Hyperspherical Learning for Computer Vision" for inviting me as Keynote speaker in their inspiring workshop. @adn_twitts @geleonti @IndroSpinelli @GalassoFab10 et al.
19 Oct 2025
Carlo's @NetworkAutomata keynote on networks and their geometry BEW @ICCVConference
3
6
609
Carlo Vittorio Cannistraci retweeted
📢 Call for Papers: NeurReps 2025 ‼️‼️‼️ 🧠 Submit your research on symmetry, geometry, and topology in artificial and biological neural networks. Two tracks: Proceedings (9 pages) and Extended Abstract (4 pages). Deadline: Aug 22, 2025. neurreps.org/call-for-papers
2
38
177
31,314
It is fascinating to study complex systems such as the brain and to inspire new manners to train neural networks. We do this with a group of students @Tsinghua_Uni youtu.be/pzkctDSZn8g?si=Sfb2…
#AI learns from the brain, and the #brain is revealed through AI,” says #TsinghuaRen Prof. Carlo Vittorio Cannistraci @NetworkAutomata (Italy), Chief Scientist at THU Laboratory of Brain and Intelligence, who envisions a new generation of AI that is energy-efficient & effective.
5
410
This seasonal variation appears to be influenced by factors such as sunlight intensity and circadian rhythms, which may alter the timing of heart attack onset.
1
1
97
Interestingly, this shift has been linked to environmental factors like solar irradiance and even vitamin D levels, which are regulated by sunlight exposure. However, the exact mechanisms behind this phenomenon remain an area of active investigation
1
88
This is the top 10 of disinformation spreaders in X according to Grok 😂. Who do you think will be at the position number 10?
147
A clear analysis of the advantages of #DeepSeek
5
256
#DeepSeek what are the architectural and algorithmic Innovations Behind DeepSeek’s Speed and Efficiency? Its Mixture-of-Experts (MoE) design dynamically activates only 37 billion parameters per token (vs. ChatGPT’s dense activation of all parameters), slashing computational waste
1
1
688