Feb 2024· International Joint Conference on Artificial Intelligence· pp. 8234-8243· 13 citations· 78 references
Computer Science
TL;DR
A unified perspective on the inherent trade-offs between expressivity and tractability is provided, highlighting the design principles and algorithmic extensions that have enabled building expressive and efficient PCs, and a taxonomy of the field is provided.
Abstract
We present a comprehensive survey of the advancements and techniques in the field of tractable probabilistic generative modeling, primarily focusing on Probabilistic Circuits (PCs). We provide a unified perspective on the inherent trade-offs between expressivity and tractability, highlighting the design principles and algorithmic extensions that have enabled building expressive and efficient PCs, and provide a taxonomy of the field. We also discuss recent efforts to build deep and hybrid PCs by fusing notions from deep neural models, and outline the challenges and open questions that can guide future research in this evolving field.
Generative models produce long and high probability sequences, yet they often fail to satisfy explicit constraints set by users. Over the past two decades, Constraint Programming (CP) has provided a complementary paradigm: combining generative models with a constraint solver to guarantee feasibility. This survey review...
Alexandre Bonlarron, F. Pachet, Pierre Roy et al.· Proceedings of the Thirty-Fi...· 2 citations
This work introduces a new architectural component that embeds structured inductive bias into deep learning: an attention mechanism operating over tensor-product representations (TPRs) that outperforms existing architectural components in combinatorial generalization.
The intricacies and breadth of generative AI (GenAI) and large language models can sometimes eclipse their practical application. It is pivotal to understand the foundational concepts needed to implement generative AI. This guide explains the core concepts behind -of-the-art generative models by combining theory and ha...
Deep Neural Networks (DNNs) have proven to be extremely effective at learning a wide range of tasks. Due to their complexity and inexplicable internal state, DNNs are difficult to analyze. Several attempts to interpret their operation have been made, but analyzing them from the perspective of the knowledge encoded in t...
Ioanna Valsamara, Ioannis Mademlis, Ioannis Pitas· ACM Computing Surveys· 0 citations
Large language models (LLMs) are capable of completing a variety of tasks, but remain unpredictable and intractable. Representation Control (RepControl) seeks to resolve this problem through targeted interventions that modify high-level representations of concepts such as honesty, harmfulness or power-seeking. We forma...
This work presents a unified framework for probabilistic explainability based on sparse, anchored linear models, applicable to both binary classification and continuous regression, and addresses the resulting empirical problem using a Mixed Integer Programming formulation that yields provably optimal empirical solution...
F. Koriche, Jean-Marie Lagniez, Chi Tran· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduOct 13, 2026
Machine-learning networks can learn without detailed instructions from a designer. Those that do may offer insight into how the brain reorganizes itself during learning.
MIT News · Artificial Intelligence· news.mit.eduSep 29, 2026
Professor Sherry Turkle’s new book, “Artificial Intimacy,” offers a withering critique of chatbots and the antisocial dynamics she believes they encourage.
