the Data Processing Inequality
No mentions found
This entity hasn't been tracked yet, or Iris is still building its knowledge base.
Related Articles from SNS
Information-Theoretic Bounds for Sparse Covariance Estimation in the Vertical-Split Distributed Model
Announce Type: new Abstract: We study the minimax estimation error for distributed covariance matrix estimation in the vertical-split (feature-split) setting, where two agents each observe different coordinates of $m$ i.i.d. and communicate a limited number of bits to a central server. [2025] established nearly tight bounds for dense (unstructured) cross-covariance matrices, we investigate whether imposing elementwise $s$-sparsity on the cross-covariance $C_{21}$ can reduce the required...
On the Generalization in Topology Optimization via Sensitivity-Conditioned Bernoulli Flow Matching
arXiv:2606.02179v1 Announce Type: new Abstract: Surrogate models for topology optimization (TO) exhibit highly variable out-of-distribution (OOD) generalization under distribution shifts such as changing loads or boundary conditions, yet the source of this variability remains unclear. We hypothesize that OOD performance is governed by how much information the conditioning signal preserves about the adjoint sensitivity (reduced gradient) that drives classical TO. Modeling the TO pipeline as a...
Capability and Robustness Cannot Both Be Free: An Information-Theoretic Bound for Vision-Language-Action Models
arXiv:2605.25889v4 Announce Type: replace Abstract: Vision-Language-Action (VLA) models reach high success rates on clean inputs but collapse under small adversarial perturbations: a $16/255$ PGD attack drops OpenVLA-7B's LIBERO success from $95\%$ to under $5\%$. Whether this trade-off has a theoretical floor was open. We prove that it does. For any VLA policy, capability $I(\Astar;\Api)$ and robustness $I(\Api;\Atildepi)-I(\Api;\delta)$ sum to at most $H(\Astar)+I(X;\Xtilde)$, the task...
Pope Leo warns AI boom can give Big Tech and the people who run it too much power
Pope Leo XIV has issued an encyclical warning that the rapid advancement of artificial intelligence could grant excessive power to Big Tech companies and their leaders. He argues that because AI lacks human experience or moral conscience, concentrated control over these systems risks creating new inequalities and dependencies. Therefore, the Pope calls for strong political regulation to ensure AI serves the common good and respects human dignity.
AI offers promise for agriculture, but smallholder farmers risk being left behind
AI offers promise for agriculture, but smallholder farmers risk being left behind Robert Egan Associate Editor Globally, agriculture faces mounting pressures. These are driven by climate change, land degradation, labor shortages, supply chain disruptions and the demand for food from a growing population. At the same time, productivity is uneven.
Harpoon: Generalised Manifold Guidance for Conditional Tabular Diffusion
arXiv:2602.07875v3 Announce Type: replace Abstract: Generating tabular data under conditions is critical to applications requiring precise control over the generative process. Existing methods rely on training-time strategies that do not generalise to unseen constraints during inference, and struggle to handle conditional tasks beyond tabular imputation. While manifold theory offers a principled way to guide generation, current formulations are tied to specific inference-time objectives and...
Automated Proving of Shannon-Type Entropy Inequalities via Fine-Tuned Language Models and Guided Tree Search
Announce Type: new Abstract: Proving Shannon-type entropy inequalities is a fundamental task in information theory that often requires constructing non-trivial linear combinations of known constraints, which is a combinatorial search problem that scales poorly with the number of random variables. We investigate whether small-scale large language models (0.6B--1.7B parameters), fine-tuned on atomic proof steps and combined with guided beam search, can automate this process. On a held-out test...
Actually, the SAT Was Necessary After All
Zvezdelina Stankova has taught mathematics at UC Berkeley for nearly three decades. But in 2023, while teaching introductory calculus for the first time since the beginning of the coronavirus pandemic, she noticed that something was quite wrong. The bottom 25 percent of students were not just struggling with the coursework, Stankova told me; “people were in freefall.”
Towards Serverless Semi-Decentralized Federated Learning with Heterogeneous Optimizers
Announce Type: new Abstract: We investigate cluster formation, involving the number and composition of clusters, in decentralized federated learning (FL) with heterogeneous machine learning (ML) optimizers. While clustering in centralized FL has enabled scalability and resource savings, its value and development in fully decentralized environments have yet to be explored. Optimizing cluster formation in such environments is challenging, especially due to the complex coupling between network...
Discovering Thermodynamically Admissible Dissipation Potentials via Grammar-Based Symbolic Regression
Announce Type: cross Abstract: Constitutive laws for inelastic materials must satisfy strict thermodynamic admissibility requirements, yet current data-driven approaches sacrifice interpretability, even when formal guarantees are provided by physics-encoded architectures. We propose a symbolic regression framework for the data-driven discovery of dissipation potentials governing the evolution of internal variables within the Generalized Standard Materials (GSM) formalism. Starting from the...