Silvia Lameri

Our paper "Mathematical Foundations of Geometric Deep Learning" (co-authored with Michael Bronstein) has now been officially added to the Geometric Deep Learning Book website as the recommended preparation material! (special thanks to Petar Veličković) https://lnkd.in/ec44bJeT The GDL Book does not assume extensive prior mathematical preparation, yet its foundations rest on several key areas of mathematics — Vector Calculus, Differential Geometry, Topology, Functional Analysis, Spectral Theory, and Graph Theory — which many students often find challenging. These topics are addressed in detail in the "Mathematical Foundations of Geometric Deep Learning" paper. In addition, the manuscript contains numerous historical annotations and margin notes that contextualize the mathematical concepts under discussion, while also highlighting direct connections to modern GDL architectures. #GeometricDeepLearning #MathematicalFoundations #MachineLearning #AI #Mathematics #Research #GDL

757

13 commenti

Our latest ICML paper on "Generative AI without training generative models" is on MIT News! We find out tokenizers that reduce data to use a few tokens compress information in unique ways. Tokens take on unique meaning. Once data is tokenized, optimization algorithms (instead of training new generative models) take a few iterations to carefully modify the tokens to take on typical generative AI tasks, like this in-painting task. Check out the MIT News article here: https://lnkd.in/etBbbQaa, and the paper here: https://lnkd.in/eJzsFiyM Led by Lukas Lao Beyer (MIT AeroAstro & LIDS). Collaboration with Kaming He (MIT EECS & CSAIL), Tianhong Li (MIT CSAIL), Xinlei Chen (Meta FAIR).

878

12 commenti

Looking forward to Antonello Longo presenting our latest work titled "Pixels-to-Graph: Real-time Integration of Building Information Models and Scene Graphs for Semantic-Geometric Human-Robot Understanding", at the IEEE CASE 2025 this week! As autonomous robots are increasingly deployed to support human operators in high-risk applications, the need for efficient human-robot cooperation and understanding is greater than ever. While robotic planning usually uses 3D geometric information, human operators are accustomed to a compact, high-level representation of the environment, such as top-down 2D maps. To bridge the gap, we introduce Pixels-to-Graph (Pix2G), a lightweight, CPU-only framework that generates structured scene graphs from image pixels and LiDAR maps in real time on the edge. The outputs of the method are a denoised 2D top-down environment map and a structured, segmented 3D point cloud. These outputs are seamlessly connected using a multi-layer graph that abstracts information from the object level up to the building level. This improves human-robot understanding and situational awareness for high-risk, time-critical missions, such as disaster response and infrastructure inspection. This research work was led by Antonello Longo in close collaboration with, Calvin Chung, Matteo Palieri, Sung Kyun Kim, Ali Agha, Prof. Cataldo Guaragnella and myself. Representing a successful collaboration between teams at the NASA Jet Propulsion Laboratory, Politecnico di Bari, and FieldAI. If you are attending IEEE CASE 2025, please attend the presentation on August 20, 2025. Check out the paper below: Paper: https://lnkd.in/gs4rd-vQ #Autonomous #Robotics #HumanRobotInteraction #HRI #SceneGraphs #SemanticMapping #BIM #PhysicalAI #Agents

305

6 commenti

Delighted that our working paper “Catching Bid-rigging Cartels with Graph Attention Neural Networks”, joint work with D. Imhof and E. Viklund, is out! We propose a novel #DeepLearning algorithm based on Graph Attention Neural Networks to detect collusive behavior in markets and tenders. Trained on data from 7 countries across 13 markets, our method achieves impressive accuracy rates of up to 91% in classifying cartels - significantly outperforming classical #MachineLearning approaches: https://lnkd.in/dmnFAt-3 #DataScience

70

3 commenti

#variational #QML #strikesback #discretely In the recent work (#link in the first comment) we show that adaptivity in variational quantum algorithms can give an exponential advantage over non-adaptive approaches. By framing circuit recompilation as a discrete optimization problem, we identify a setting where hybrid quantum–classical loops are not just helpful, but necessary. Big thanks to my collaborators — 𝗭𝗼ë 𝗛𝗼𝗹𝗺𝗲𝘀, for super-energising discussions and pushing this through, and Chukwudubem Umeano, for helping to compile this piece together. Careful, long read below ❗️ 𝗔𝗱𝘃𝗮𝗻𝘁𝗮𝗴𝗲 𝗳𝗼𝗿 𝗗𝗶𝘀𝗰𝗿𝗲𝘁𝗲 𝗩𝗮𝗿𝗶𝗮𝘁𝗶𝗼𝗻𝗮𝗹 𝗤𝘂𝗮𝗻𝘁𝘂𝗺 𝗔𝗹𝗴𝗼𝗿𝗶𝘁𝗵𝗺𝘀 𝗶𝗻 𝗖𝗶𝗿𝗰𝘂𝗶𝘁 𝗥𝗲𝗰𝗼𝗺𝗽𝗶𝗹𝗮𝘁𝗶𝗼𝗻 The question of whether variational quantum algorithms really need the hybrid quantum–classical loop has been a matter of debate in the last two years. The challenge comes from the ever-growing power of surrogating quantum models, and advances in classical simulation methods for problems with structure. This has been framed as a question to the community: 𝘋𝘰𝘦𝘴 𝘵𝘩𝘦 𝘱𝘳𝘰𝘷𝘢𝘣𝘭𝘦 𝘢𝘣𝘴𝘦𝘯𝘤𝘦 𝘰𝘧 𝘣𝘢𝘳𝘳𝘦𝘯 𝘱𝘭𝘢𝘵𝘦𝘢𝘶𝘴 𝘪𝘮𝘱𝘭𝘺 𝘴𝘪𝘮𝘶𝘭𝘢𝘣𝘪𝘭𝘪𝘵𝘺? Counterexamples exist, but most of them correspond to specific Boolean function learning or Shor-type algorithms in disguise. What we tried to do is to identify a setting where adaptivity appears more naturally, and relies on a different structure. The starting point is a simple observation: in classical discrete optimization, exponential separations between adaptive and non-adaptive search are known. If this logic carries over, then there should be quantum problems where adaptivity gives the same kind of exponential edge. In our preprint, we propose such a problem — quantum circuit recompilation. The task is to recover a “hidden” circuit structure, starting with quantum data (a state) prepared with randomly placed T-gates (“puzzles”) embedded in partially random unitaries. The ansatz has the reverse structure, but we don’t know where the gates are, and there are exponentially many options. We show that discrete quantum optimizers can efficiently unravel these puzzles, finding the right locations with a quadratic number of circuit evaluations. In contrast, non-adaptive searches face an exponentially hard landscape: non-separable, highly entangled, and full of magic. Importantly, we also find that optimization remains feasible under noise, using finite-shot samplers, even as the solution space grows exponentially. This points to a very particular (and largely overlooked) role for discrete optimization in variational quantum algorithms. It also raises a bigger question: what new problems can benefit from this kind of adaptivity? While we remain open to new discoveries, we already see clear connections to fault-tolerant protocols with partial structure, and to blind quantum computing scenarios. Grateful for the support from EPSRC and the QCi3 Hub!

93

7 commenti

Are you interested in hierarchical dimensionality analysis? Here's our new "Taxonomic Graph Analysis" used to model the IPIP-NEO Personality Hierarchy. The project is led by Andrew Samo and Alexander Christensen, with the collaboration of Luis Garrido, Paco Abad, Sam McAbee, and me! In our preprint (link in the comments) we introduce a new approach to understanding personality structure: Taxonomic Graph Analysis (TGA), is a comprehensive network psychometrics approach that identifies hierarchical personality structures from the bottom up rather than imposing existing theoretical models. We applied TGA to the open-source 300-item IPIP-NEO dataset with over 149,000 participants and reveal a three-level structure of personality: 28 first-level dimensions (facets) 6 second-level dimensions (traits) 3 third-level dimensions (meta-traits) What makes this approach very interesting is how it addresses longstanding methodological challenges in personality assessment: Local independence violations Wording effects Dimensionality assessment Structural robustness While some dimensions aligned with traditional IPIP-NEO structure, we found significant deviations. The emergence of novel dimensions like "Sociability," "Integrity," and "Impulsivity" at the second level and a "Disinhibition" meta-trait at the third level represents a major departure from the traditional five-factor model. Our research effectively integrates empirical findings scattered across personality literature into a coherent hierarchical structure. This data-driven framework demonstrates TGA's value for investigating complex psychological constructs and offers a rigorous new perspective on personality taxonomy. This work exemplifies how innovative psychometric methods can reshape our understanding of fundamental psychological constructs. Its implications extend beyond personality assessment to potentially revolutionize how we conceptualize and measure human traits across psychology.  Link to the Published Paper: https://lnkd.in/dZKzzwMi

595

20 commenti